Chapter Six—
Roman Jakobson, or How Logology and Mythology Were Exported

It's certainly fair to say that Roman Jakobson is the biggest name in modern linguistics. No other figure dominates the field quite the way he did. This is only partly because of his uncanny productivity. It is also because he was around for so long and in so many places. Since he spent the last forty-one years of his life, from 1941 to 1982, in the United States, he is regarded as almost an American academic fixture. But before he came here, he had spent a couple of years in Norway and Sweden and almost twenty years in Czechoslovakia. He spent the first twentyfour years of his life in Russia, where he was born in 1896.

This geographically and chronologically wide-ranging career allowed Jakobson to have an extraordinary international impact on both linguistics and literary theory. In his early days he was closely associated with the Formalists and Futurists in his native country. After he went to Czechoslovakia in 1920, he founded the Prague Linguistic Circle, which is mentioned in every account of the history of structuralism. And in the United States his long career in the academic establishment made him a continuing presence in various schools of literary theory.

No one doubts the importance of Jakobson in the history of structuralism. The facts can be stated briefly: Jakobson received his copy of Saussure's Cours de Imguistique générale in 1920, although he had been familiar with Saussurian doctrine since 1917. He was troubled by certain aspects of Saussure's system, above all its inability to accommodate the notion of ongoing change in language, so he elaborated a new view of language that would take account of this temporal factor. The document

in which this view appears is "Problems in the Study of Literature and Language," which Jakobson coauthored with Formalist critic Yury Tynianov and published in 1928. This programmatic list of principles and recommendations introduced the notion that language is a "system" that "necessarily exists as an evolution." But it also contained a proclamation that was to play a significant role in the development of structuralist thought:"An analysis of the structural laws of language and literature and their evolution inevitably leads to the establishment of a limited series of actually existing structural types (types of structural evolution)."^[1] The following year, at the First Congress of Slavists in Prague, Jakobson proposed new modes of poetic analysis and, as he tells the story later, "christened" them the "structural method."^[2] Thus, by Jakobson's account, the foundation was laid for the elaboration of structural principles into a broad methodology.

Jakobson was interested not only in linguistics but in poetics as well. In fact, many people today feel that his work in poetics is superior to his work in linguistics. Some of his most obvious contributions to modern structuralism were in this field. For instance, the famous study of Baudelaire's "Les Chats" that he coauthored with Claude Lévi-Strauss in 1962 reads almost like a parody of the structuralist method. It combines semantic, grammatical, syntactical, phonetic, and prosodic methods of analysis to provide a ridiculously complicated set of organizational schemas—all for the purpose, the authors say, of giving the poem "the character of an absolute object."^[3] The article has its own involved history in the annals of modern criticism, a history that has secured it a place in virtually every account of the rise of structuralism. But then Jakobson's entire association with Lévi-Strauss, which had begun in the first years Jakobson spent in the United States, was a formative influence in that movement.

Another thing for which Jakobson is particularly well remembered in literary circles and in the history of structuralist method is his distinction between the two "poles" of language: the metaphoric and the metonymic. The idea is that all language production occurs on two "axes": the vertical axis of selection and the horizontal axis of combination. In other words, as we speak we perform two different operations: we select units as if from a vertical column of choices and then we string them together as if on a horizontal track. When we select, we act on the basis of a principle of similarity (because the different choices in the vertical column will necessarily bear some resemblance to one another), and when we combine, we act on the basis of a principle of contiguity (be-

cause, in the simplest sense, the units we combine go next to each other). These two operations have to do with two basic figures of speech: metaphor and metonymy. When we use a metaphor, what we are really doing is substituting one thing for another, which is to say that we are performing a "vertical" operation of selection. Thus when I refer to someone as a weasel, I am selecting from a whole column of terms that suggest furtiveness or some other quality belonging to both weasels and the person I am comparing them with. When we use a metonymy, what we are really doing is finding something that is associated and "contiguous" with the first term. When I refer to the food I eat as my board, it is not because food and board are similar, but because the idea of food calls to mind something contiguous with it, namely the table, or "board," on which it is served. Contiguity here, of course, has a slightly different sense from the one it has when we talk about the horizontal axis of selection.

In an article on two types of aphasia, Jakobson actually used neurological research to support his findings. He found that one type of aphasic disturbance consists in a similarity disorder, that is, where the metaphoric pole of language has been disturbed and the subject has trouble choosing individual words, whereas the other type consists in a contiguity disorder, that is, where the metonymic pole has been disturbed and the subject has trouble stringing words together. These findings confirmed for him that the two axes of language actually represent two basic mental operations involved in speech formation.^[4]

What is important in Jakobson's observations on metaphor and metonymy is the conclusion that he draws from them. If there are two basic figures of speech, two principles of language production, it is because there are two basic types of language: our old friends, poetry and prose. Poetic language, since it is rich in metaphor, relies more on the axis of selection. This makes sense because poetry is less a temporal and sequential form of language than is prose. Prosaic language, by contrast, is sequential and linear and thus relies more on the axis of combination. In fact, the principle of similarity, or equivalence, is so dominant in poetic language that it fairly overwhelms any instances of combination that we might find in language of this sort. Thus we arrive at Jakobson's celebrated statement, set in italics in his article "Linguistics and Poetics": "The poetic function projects the principle of equivalence from the axis of selection into the axis of combination" (SW, 3:27). What this means is that in poetry even the combination, the linear sequence of words, is dominated by principles of similarity.

But here, too, the main thing is how poetic language calls attention to itself "as such." In "Linguistics and Poetics" Jakobson gives his wellknown schema of the six "functions" of language. Among the functions he lists are the poetic function and the "referential" function. The referential function is also called the "denotative" and the "cognitive" function, and it has to do with what we normally think of as the transmission of information in simple communication. Each function is distinguished by its orientation toward one of six corresponding factors. Referential language is oriented toward the context . Poetic language is oriented toward the "message as such, . . . the message for its own sake" (SW, 3:25). Once again, the idea is that poetic language is different from ordinary language because it calls attention to itself.

What is a man of science like Jakobson doing concerning himself with the difference between poetic and prosaic language? Hasn't this problem been strangely resistant to any kind of scientific solution? Isn't it more and more tempting to conclude, in the face of so many failed attempts to solve it, that there is no such scientifically describable difference? As it happens, Jakobson from the very beginning was wrestling with all the old myths, and his later attempts to come up with a scientific formulation for some of them make a lot more sense if we look at what he was doing early in his career.

Jakobson's own account of his linguistic genealogy is puzzling. In the "Retrospect" to the first volume of his Selected Writings he mentions a few linguists that helped inspire him to take a new direction in language studies. But he also says that the strongest impulse came from the "great men of art born in the 1880s," and he lists Picasso, Joyce, Braque, Stravinsky, Khlebnikov, and Le Corbusier (SW, 1:631–32). He goes on to say that his first topic for the analysis of language "in its means and functions" was the poetry of Khlebnikov (SW, 1:633). In the "Retrospect" to another volume he credits Bely with inspiring him to take on the analytic study of verse (SW, 5:569). I mentioned earlier that Jakobson was closely associated with Formalists and Futurists. What he says about himself and much of the evidence about his early career suggests that the fundamental aspects of his thinking about language came not from professional linguists but from other sources—chiefly poets, artists, and literary critics.

One of Jakobson's earliest published writings is a long essay on Khlebnikov called "Modern Russian Poetry."^[5] If ever there was evidence that Jakobson was a man of his age and a product of indigenous tradition, this is it. The only thing that distinguishes it from the work of, say,

Shklovsky is that Jakobson pays slightly more attention to the grammatical aspects of his subject. Otherwise the essay is pure Formalism, complete with all the mythical and Platonic baggage we saw in Shklovsky's work.

The effort is always to locate the essence of poetic language, to find the quality in poetry that makes it different from ordinary discourse. When Jakobson proposes solutions to the poetry-prose puzzle, we find the same phrases and concepts turning up as we had seen in the writings of Futurist poets like Khlebnikov and Kruchenykh and Formalist critics like Shklovsky. Poetic language is different from ordinary language because in it "linguistic representations (both phonetic and semantic) focus more attention on themselves" (SW, 5:304). He quotes Khlebnikov approvingly and says that "poetry is the formation of a self-valued, 'selfspun' word" (SW, 5:305). In passages reminiscent of Shklovsky he describes how words undergo a gradual process of petrification until they cease to be felt as words. When words reach this point, we read, a poet needs to come along and revive them so they can be felt as such. Or better yet, a poet can come along, as Khlebnikov has done, and invent whole new words, which can't avoid being felt as new because they are. "The form of words in practical language easily ceases to be consciously felt, it dies away, becomes petrified, whereas the perception of the form of a poetic neologism, a form given in statu nascendi, is absolutely compulsory" (SW, 5:333). Khlebnikov seems to have accomplished the resurrection of the word that Shklovsky had called for five years earlier. When the poet exploits consonant similarities in adjacent words, Jakobson says, "the word acquires a sort of new sound characteristic, the meaning bestirs itself, the word is perceived like a friend with a suddenly unfamiliar face, or like a person we don't know, but in whom something familiar can be divined" (SW, 5:342).

When Jakobson had spoken of the self-valued word and the tendency of poetic language to call attention to itself, he was really just answering the question of what poetic language does, not what it is . But he returns again and again to the question of what there is about poetic or literary language that makes it poetic or literary. "Poetry is language in its aesthetic function" (SW, 5:305). What quality is there that makes poetic language have this aesthetic function? Jakobson started out by considering literary language in general, without for the moment distinguishing literary from poetic language. He came up with a name for the distinctive quality in literary language: literaturnost ', or "literarity." In a famous sentence that gave the term currency in the Formalist movement, he says

that "the object of a science of literature is not literature, but literarity, that is, what makes a given work a literary work" (SW, 5:305).

Naturally, neither Jakobson nor any other Formalist thinker ever came up with a truly satisfying definition of literaturnost '. This ultimate object of the "science of literature" always seemed to be on the other side of the horizon. But that makes perfect sense, because literatumost ' is intrinsically an ideal, essentialist concept; and if it is ideal, then it is not accessible to the kind of knowledge implicitly understood in the phrase science of literature . No, unfortunately, Formalism was no science at all. It was founded in mystical and mythical notions. Now it appears that it was an essentialist and idealist doctrine as well, and Jakobson's quest for that pure aesthetic quality of literary language is a perfect example.

In the first part of this book I talked about inner form and showed that it was one of the most pervasive concepts in Russian language theory. It was also one of the most unscientific concepts in theories that masqueraded as scientific. Not surprisingly, the term springs up in Jakobson just as naturally as it had in Potebnia, Bely, and Shklovsky. At one point Jakobson discusses Khlebnikov's word inventions, saying that words of the sort that he creates "seem almost to be seeking out a meaning for themselves. In this case," he continues, "one cannot, say, speak of the absence of semantics. These are, to be more precise, words with negative inner form" (SW, 5:353). A few paragraphs later he says that the word in Khlebnikov's poetry "loses its object-quality [predmetnost '], then its inner and finally even its outer form" (SW, 5:354). It's hard to extract from these statements a precise sense of inner form as Jakobson understands it. Still, the mere fact that he doesn't pause to explain the term suggests that he is using it in what at least he regards as its customary sense. Inner form seems to have to do with the word's meaning and, more than this, with the mysterious way in which the word means that meaning. Thus a word with "negative inner form" is one that appears to retain the connectedness with its meaning even though the meaning—in the sense of a true, "objective" meaning—is lacking.

But isn't inner form just another essentialist doctrine when it comes right down to it? Doesn't it make perfect sense that the Russians should have seized hold of this concept from a romantic thinker like Humboldt, then redefined it in such a way that it would now satisfy their indigenous, age-old religious need for essences, ideals, eide, iconic prototypes, and such Platonic otherworldly notions? Potebnia had taken what for

Humboldt had been a dynamic concept and, in characteristic Russian fashion, replaced it with something static, fixed, and eternal.

This search for essences is utterly typical of Jakobson's thought. A curious chapter in the Jakobson story that is seldom mentioned helps account for Jakobson's essentialism (or his essentialism helps account for it . It is his association with phenomenology. Anyone familiar with this movement in early twentieth-century philosophy knows how difficult it is to define phenomenology, above all since Edmund Husserl seemed to want to redefine it every time he wrote a new book about it. But one can say at least that, in the eyes of Husserl's contemporaries, it was a philosophy of pure thought. The phenomenologist sought, through the "phenomenological reduction," to abstract or "bracket" from thought all prior experience, all presuppositions about the world, in order to isolate "thought" itself. At another stage, the phenomenologist, through the application of "eidetic intuition" and "eidetic reduction," brackets away the qualities in a phenomenon (that is, anything that appears to consciousness) that are peculiar to that phenomenon in order to arrive at the essence or eidos common to all similar phenomena. Phenomenology is thus, among other things, a science of essences. This is undoubtedly what made it appealing to the Russian mind in the early twentieth century. A number of Russian intellectuals came to champion Husserl's cause in their own country, largely at the instigation of one Gustav Shpet, who studied with Husserl. Interestingly enough, Shpet was later to write an entire book with the Humboldtian-Potebnian title The Inner form of the Word .^[6]

Elmar Holenstein has investigated the Jakobson-Husserl connection and shown the degree to which Jakobson's thought was formed by his exposure to phenomenology. He feels so strongly about the impact of phenomenology on Jakobson, in fact, that the subtitle to his book on Jakobson's theory of language describes that entire theory as "phenomenological structuralism."^[7] One fundamental feature that phenomenology and Jakobsonian linguistics have in common is precisely the notion of essences, Holenstein thinks. In an article devoted exclusively to Jakobson and Husserl, he shows how the notion makes its appearance in both fields. Holenstein describes the search for essences in phenomenology as "the search for invariants in all the variations, for the general in everything particular."^[8] This is because in phenomenology essences have to do with those qualities in a thing that remain unchanged despite variations in the way the thing is perceived. For example, the quality of

extension is invariant because an object cannot exist without it. The two chief concepts in Jakobson that Holenstein sees as analogous are literarity (Holenstein translates it as "literaricity") and poeticity (poèticnost^[*] '), a term that Potebnia had used and that Jakobson reintroduced about a decade after the Khlebnikov article to describe the quality that makes a specifically poetic work poetic.^[9] Holenstein then draws an important conclusion from his observation. "Jakobson," he says, "knows just as well as Husserl that in the search for the essentialities success is not to be had with induction and statistics, but only with phenomenological analysis and insight into the object of investigation itself."^[10] The standard tools of empirical science thus have no place in this central task of linguistics. In other words, to the extent that it seeks essences in the objects of its own investigation, Jakobsonian linguistics, that important precursor to the scientific methods of structuralism, is not a science in the sense many of us understand when we speak of a science of language.

I don't know whether phenomenology was the reason for the essentialist impulse in Jakobson, or if the essentialist impulse was already there and was the reason for Jakobson's interest in phenomenology. The second hypothesis makes more sense, for essentialism was a native organism of long standing in Jakobson's mother country. But it doesn't really matter much in the end. The fact is that Jakobson's thought is fundamentally essentialist and that if phenomenology either fed or caused his essentialism, if it played a role in the development of his thought as he himself acknowledges, then the essentialist impulse is a strong one in the formation of structuralism. Holenstein is not the only person to have asserted that phenomenology in some sense leads to structuralism.^[11]

Jakobson's essentialism was not merely a feature of the youthful thinking associated with his early Formalist-Futurist period; it was to remain with him for his entire career. For instance, the concept of poeticity that Holenstein mentions comes from an article that Jakobson published (in Czech) in 1933 and 1934 called "What Is Poetry?"^[12] Even at that date, well after the heyday of Russian Formalism and presumably well after the era of romantic and mystical language philosophy, Jakobson is trotting out the same old notions about the specificity of poetic language. "Poeticity," he says, "is present when the word is felt as a word and not a mere representation of the object being named or an outburst of emotion, when words and their composition, their meaning,

their external and inner form acquire a weight and value of their own instead of referring indifferently to reality" (SW, 3:750).

Even better evidence of the tenacity of Jakobson's essentialism is his 1965 article titled "Quest for the Essence of Language" (SW, 2:345–59). Early in this article Jakobson refers to Saussure's idea of l'arbitraire du signe (pointing out, by the way, that the idea is twenty-two hundred years old) and then embarks on his own quest for instances where the signe is not so arbitraire . In some cases, he says, borrowing terms from Charles Sanders Peirce, there is a "diagrammatic resemblance" between the sign and the signified. For example, the phrase veni, vidi, vici bears a diagrammatic resemblance to the signified because the order of the words is the same as the order of the actions they designate. Another place where the diagrammatic resemblance shows through is in the length of words. With few exceptions the length of words corresponds to the notions of degree and number (superlatives are almost always longer than positives; plurals are almost always longer than singulars).

So strong was Jakobson's faith in the "essence" in language that he was led to adopt some strangely paradoxical positions later in his career. In his 1965 article "On Linguistic Aspects of Translation" Jakobson introduces another one of his favorite ideas, namely the interdependence of signatum and signum . "There is no signatum without a signum, " he says (SW, 2:260). This is a striking sentence, because it means there is no signified thing (signatum ) that exists independently of the sign (signum ) standing for it. The signified exists only when and because there is a sign for it. In other words, language is a kind of enclosed universe unto itself that is ultimately indifferent to the existence or nonexistence of a "real" world without. This sounds much like the lonely, postmodern language philosophies of the 1970s, when it was fashionable to see language as a big subterfuge, forever undermining itself and always refusing to point to any kind of concrete reality. I don't mean to suggest that Jakobson is taking such an extreme position, but what he's saying certainly seems out of step with any theory that cherishes the idea of an essence in language. In the interdependence theory the relation between sign and signified, like the relation between sign and outside world, is bound to be arbitrary: sign and signified depend on each other; the signified exists as a function of the sign. But the signified cannot be said to necessitate the sign, and if the converse is true, namely that the sign necessitates the signified (which exists, remember, only as a function of the sign), then it will have to do so by means that are mysterious

indeed. Essentialist theories, by contrast, always want to see a necessary connection between sign and signified, sign and thing, or sign and concept.

Naturally, poetic language was always there to restore Jakobson's faith. In 1965, only seven years after writing "On Linguistic Aspects of Translation," Jakobson writes about poetic language, saying that in it there is "an essential change in the relation between the signifier and the signified, just as there is between the sign and the concept."^[13] In other words, you may speak all you like about the interdependence of sign and signified, provided that you're speaking only of everyday language. When it comes to poetry, however, there is an "essential change." Jakobson doesn't say here just what this change is, but it's easy enough to fill in his meaning. In "Quest for the Essence of Language," written in the same year, he borrows a term from Peirce and refers to certain "iconic" properties of language, by which he means properties that show a factual similarity between signifier and signified. One highly significant iconic property in language may be found in the selection of phonemes. That is, certain sounds in language bear a factual similarity to the things they signify. Jakobson even cites Mallarmé's "Crise de vers" and the paradoxical example of jour, a dark-sounding signifier with a light signified, and nuit, a light-sounding signifier with a dark signified. Amazingly, Jakobson's point is not, as it was for Mallarmé, that this example proves the arbitrary nature of the sign-signified connection but instead that Mallarmé was "deceived" (Jakobson's translation of Mallarmé's deçu, "disappointed") by the perversity of this one case. Jakobson has been busy showing how the connection is nonarbitrary when it comes to the phonemic qualities in words, and he now goes on to say that the connection is particularly nonarbitrary in poetic language. The trouble with this view is that it brings all kinds of factors from the external world into the picture and thus wrecks the serene isolation of the interdependence model. Sign and signified are no longer purely interdependent; now they depend on something outside, like the lightness of the day or the darkness of the night.

Essentialism was similarly at the root of Jakobson's theories of poetic language. In Jakobson's immediate cultural context the idea of the "word itself," poetic language that calls attention to itself, goes back to the zaum ' theorists, who believed that poetic language should be ideally signifying. Poetic language that created its own signified world signified that world through an ideal and necessary connection between sign and signified. Poetic language that did not create its own world simply sig-

nified through an ideal, direct connection. In both cases the word had to exist "as such" and call attention to itself, because, owing to its ideal signifying powers, by calling attention to itself it was also calling attention to its signified.

When Jakobson refers to the iconic properties of language, he is using iconic in a sense borrowed from Peirce. And yet isn't his whole notion of poetic language iconic in the Orthodox sense, too? All Jakobson's essentialist notions—literarity, poeticity, inner form—point to the presence of an immaterial, invisible core in words, a core that is ultimately neither accessible to ordinary perception nor susceptible of scientific description; it is incarnate, like the prototype in an icon, like the Word in the Son of God, like God in creation. What this means is that despite his efforts to make the study of language a respectable, scientific discipline, Jakobson is really proposing the same Orthodox model of the aesthetic epistemological attitude as we have found in Bely and in the overtly religious thinkers I have spoken of. The way Jakobson sees things, the reader of poetic works of art performs the same operation as the worshiper of icons. Reading means reaching beyond the material dimension of language back to that irretrievable essence, whose existence is known only by faith, just as the existence of the prototype of the icon is known only by faith.

Jakobson was no stranger to religious doctrine. This dimension of his work is often overlooked. Though he was Jewish by birth, he converted to Orthodox Christianity in the late 1930s, and though he was not an overtly religious man, he devoted a considerable amount of scholarship directly or indirectly to religious matters. Over a period of at least sixty years he produced a staggering amount of scholarship on early Slavic texts, almost all of them religious. During his years in Czechoslovakia he devoted a tremendous amount of time to the study of early Czech writing, most of it religious. One of his chief scholarly interests for many years was Saint Cyril, the ninth-century Macedonian missionary who, together with his brother, Methodius, is credited with inventing a Slavic alphabet and is responsible for spreading the use of the Slavic vernacular in the liturgy. Jakobson's interest in these subjects is generally linguistic and historicolinguistic, but a glance at his writings shows that he was intimately familiar with Christian theological doctrine in the Czech and the Russian traditions. Any linguist studying liturgical texts is certain to run up against the theological notion of the Word and is thus likely to raise many of the same issues I have been discussing. This is true for Jakobson.^[14]

The volume of Jakobson's Selected Writings in which the early Slavic material is collected was published only in 1985, and this may partly explain why the standard books on Jakobson contain almost no mention of it. A more reasonable explanation, however, is that the commentators on Jakobson have not quite known what to make of it. They like to think that the Jakobsonian canon is restricted to the basic set of principles usually associated with him and officially sanctioned by American literary academia. But this view is just as narrow as the view that accepts Jakobson's more widely known theories as scientific.

One could describe even the method of poetic analysis that Jakobson uses in one of his most scientific, analytic texts, his article on Baudelaire's "Les Chats," as iconic. The effort there is to pierce through the material veil of word-sounds so as to discover underneath an invisible abstraction of grammatical relations. This method, of course, has a more familiar name: structuralism. But it is clearly iconic in its conception. This means that the iconic principle is likely at the heart of Jakobson's contribution to the entire structuralist enterprise. What does one find once one has pierced the material veil? What is the hidden god? A prototype. Or an abstract structure of relations. And this, conveniently enough, is the subject of the next part of this book, where Jakobson will play a significant role once again.