CHAPTER SIX
CHAPTER SIX
THE FORMALIZATION OF LANGUAGE AND THE “CYBERNETIZATION” OF LINGUISTICS
—Your model is truly on the scale of the axioms of Riemann’s space!
—Oh, thank you. By the way, can you remind me what those
axioms are, for I don’t quite remember them…
—Well, as a matter of fact, neither do I.
One Soviet structural linguist of the 1960s to another[1]
In the context of the political “thaw” of the middle 1950s, structural linguistics, along with a number of other areas of study that had been suppressed in the late Stalinist period, began to gain ground in the Soviet Union. Structural linguists actively interacted with representatives of other disciplines—mathematicians, logicians, computer scientists, and communication engineers—and borrowed from them new concepts, approaches, problems, and institutional support. A series of borderline research fields emerged: mathematical linguistics, machine translation, applied linguistics, cybernetic linguistics, and semiotics. These fields often overlapped; the boundaries between them were blurred. Each of them, however, in one way or another, stood in opposition to the traditional discourse of Soviet linguistics and the humanities. The links between these new fields and cybernetics proved crucial for establishing their legitimacy.
Soviet structural and cybernetic linguistics and its generalization as a cultural theory—semiotics—have recently been reexamined from a variety of perspectives. The British scholar Ann Shukman has called the Soviet semiotics movement “modernist, technological, Western” and argued that it “represented a new enlightenment and a rationalist positivism.”[2] In her view, Soviet semiotics opposed a “quasi-Marxist, quasi-structuralist branch of literary officialdom.”[3] The Soviet semioticians, she has argued, “found in the terminology and concepts of information and communications theory the framework for an approach to vaster problems than were ever envisaged by Shannon or Wiener” and “claimed to extend positivist knowledge over all fields of human activity.”[4]
In the post-Soviet period several prominent Russian structural linguists and semioticians have published their memoirs, stressing the specificity of Soviet structuralism as a social rather than cognitive phenomenon.[5] Vladimir Uspenskii has maintained, for example, that “mathematical linguistics was not so much a scientific field (what kind of field was that, if it consisted in the application of methods of one science to another?) as it was a [social] movement. In essence, it was one of the first informal movements” in the Soviet Union.[6] Alexander Zholkovsky has agreed that Soviet structuralism was a “complex socio-cultural movement (rather than purely theoretical research).”[7] In contrast to Shukman’s view, he has argued that the Soviet semiotics movement was a distinctively Soviet phenomenon, which, despite its opposition to the establishment, “shared or mirrored some of its counterpart’s features: authoritarian discourse, sectarianism, territoriality, ritualistic formulae, etc. . . . to a significant degree what Saul Morson calls ‘semiotic totalitarianism’ was characteristic of the movement as a whole.”[8]
Boris Gasparov—another former participant of the semiotics movement —has applied semiotic analysis to the movement itself and found a “parallelism (or, in the semiotic language, ‘isomorphism’) between the semiotic principles of semiotic studies and the . . . social and psychological conditions in which the semiotic school had been formed and worked.”[9] He has found hermetic tendencies in both the semiotic theoretical approach (with its aspiration to transcend the empirical and look for underlying structures) and the semioticians’ “behavioral code” (with their tradition of closed meetings, isolated as much as possible from the professional, social, and political environment). In Gasparov’s view, the Soviet semioticians had “totalizing thinking” and strove for “absolute synthesis” of cultural universe in pursuit of the great “hermetic semiotic utopia.”[10]
In his provocative analysis of the recent “crisis” of the semiotic movement, the art historian Grigorii Revzin—the son of the late linguist and semiotician Isaak Revzin—has pointed to the immanent roots of this crisis in the very ideal of natural science, adopted by the movement. This ideal, Grigorii Revzin has argued, stressed order, logical consistency, and predictability at the expense of the individual, the contradictory, and the accidental in culture. He has claimed that semioticians viewed culture as a “dominion of regularity with strict rules of the game” and, while ignoring the individual, ignored the element of freedom in culture.[11]
The picture may look more complex, however, if one considers semiotics not as a fixed worldview, a grand theory, or a set of declarations, but as discursive practice. The basic principles the semioticians declared would thus be viewed as a product of discourse rather than its pre-existing driving force. Semiotic language would then be seen not as an instrument in service of some already accepted theory but as a crucial element of the practice that created that theory. Observers often confuse semiotic theory with its parent discourse; nevertheless, the two are not necessarily “isomorphic.”
Shukman’s epithets—“modernist, technological, Western”—apply fully to semiotic language rather than to the actual direction of semiotic studies. The Soviet semioticians employed and invented a great number of direct calques of foreign terms—information, code, isomorphism, homeomorphism, manifestation, generation, opposition, model, discrete, continuous, linear, etc.—which were highly uncommon in the Russian-language humanities tradition, and this supplied a “characteristic ‘Westernizing’ tinge” to their language.[12] This engineering vocabulary created a deceptive impression, however, for the Soviet structural linguists and semioticians used these terms in a non-technical, often metaphorical sense to solve rather traditional problems of linguistics and cultural theory. They were trying to change the world not by rebuilding it, like engineers, but rather by reinterpreting it, like humanists.
This chapter examines how cognitive and social aspects of Soviet structural linguistics shaped each other in the process of the formalization of language and the “cybernetization” of linguistics. I will explore the interrelations between linguistics and cybernetics, between linguists and cyberneticians, and between linguistic concepts and mathematical formalisms. Soviet structural linguists and cultural theorists learned cyberspeak, the language of cybernetics, and, despite its cognitive limitations, were able to overcome political constraints and speak more freely.
From Linguistics to Communication Science: The Role of Roman Jakobson
Cybernetic ideas became an integral part of mainstream linguistics in a large part due to the efforts of Roman Jakobson, one of the central figures in twentieth-century linguistics.[13] He acted as a crucial link between different national communities of linguists, between different schools in linguistics, and even between linguistics and other sciences. He brought the legacy of Russian linguistic traditions to the West and the gist of Western innovations back to Russia. Jakobson was no “system builder”; he rather built communities, or “circles,” of likely minded researchers.
Jakobson was born in Moscow in 1896, in the family of a chemical engineer and industrialist. In 1918, he graduated from the Historical-Philological Faculty of Moscow University; while still a student, he became a founding member and president (until 1920) of the Moscow Linguistic Circle, an informal group of researchers, who gathered regularly to discuss new approaches in linguistics, poetics, metrics, and folklore.[14] In this period, Jakobson was introduced to Ferdinand de Saussure’s work on structural analysis of language as a system of signs, to Edmund Husserl’s phenomenological studies and the concept of universal grammar, and to Gestalt psychology with its emphasis on the constitutive role of relations of elements to the whole in a system.
In 1920, Jakobson went to Prague as an interpreter for the Soviet diplomatic mission and, in the view of contemporary political developments in Soviet Russia, decided not to return. In Czechoslovakia, he was among the founders of the Prague Linguistic Circle and served as its vice-president until 1939.[15] The Prague school developed an integrated functional and structural view of language; one of its best known contribution was an innovative structural theory of phonology.
In the works of Jakobson and the prominent linguist Prince Nikolai Trubetskoi, phonemes were defined both functionally (as sound elements which differentiate between morphemes) and structurally (as elements in a phonological system). Thus, the sound side of language is presented as a structural whole rather than a simple agglomerate of disparate phonemes. The organization of phonemes into a system with a strictly defined set of relations led Jakobson to the idea of implicational laws (rules) of phonology, which made it possible to predict certain structural elements if most of the other elements were known. The sound system of a particular language provided a characteristic pattern, which could serve as a basis for comparison between different types of languages.
Jakobson did not limit himself to the phonological problems though. He used to say, paraphrasing the Latin playwright Terence: “Linguista sum; linguistici nihil a me alienum puto!”[16] Jakobson promoted a structural approach for the whole of linguistics, in fact, for the whole of science, and in 1929 coined the term structuralism to denote his new vision:[17]
Were we to comprise the leading idea of present-day science in its most various manifestations, we could hardly find a more appropriate designation than structuralism. Any set of phenomena examined by contemporary science is treated not as a mechanical agglomeration but as a structural whole, and the basic task is to reveal the inner, whether static or developmental, laws of this system.
In the late 1930s and early 1940s, Jakobson fled the Nazis first to Scandinavia and then to the United States. His structuralist ideas began to spread widely in the academic communities he visited; at the same time, he added two key components to his phonological theory: (1) the idea that phonemes could be viewed as “bundles,” or combinations, of “distinctive features” (minimal distinguishable characteristics); and (2) the idea that there must be a system of binary oppositions underlying the system of relations among phonemes. All distinctive features, Jakobson argued, could be arranged into pairs of opposites: voiced/voiceless, nasal/oral, grave/acute, tense/lax, etc. If one phoneme differs from another by just one distinctive feature, these two phonemes form a binary opposition. More complex phonemic relations, such as vocalic and consonantal triangles, could be represented as a juxtaposition of several underlying binary oppositions. The entire phonological system of a particular language could thus be viewed as a set of phonemic binary oppositions with respect to various distinctive features.[18]
In 1942-46, Jakobson taught at the Ecole Libre des Hautes Etudes in New York. Claude Lévi-Strauss attended his lectures, quickly picked up the principal structuralist ideas, and built on their basis a new discipline of structural anthropology, which closely followed the pattern of Jakobson’s phonology, including the reconstruction of an underlying system of signifying oppositions and the arrangement of interrelated oppositions into triangles.
In 1943, Jakobson co-founded yet another Linguistic Circle—this time in New York. From 1943 to 1949, he taught at Columbia University, where he eventually became an object of an anti-Communist witch hunt. Some of Jakobson’s colleagues resented his active contacts with scholars from Eastern Europe and accused him of Communist sympathies. In 1949, Jakobson took a chair at Harvard University; in 1957, he was also named Institute Professor at MIT. He played an active role in academia until his death in 1982.
In December 1948, while recovering in hospital after a car accident, Jakobson read Wiener’s Cybernetics and was thrilled by it. He soon wrote to Wiener:[19]
It is indeed a book which is epoch-making. At every step I was again and again surprised at the extreme parallelism between the problems of modern linguistic analysis and the fascinating problems which you discuss. The linguistic pattern fits excellently into the structures you analyze and it is becoming still clearer how great are the outlooks for a consistent cooperation between modern linguistics and the exact sciences.
Anticipating his impending move from Columbia to Harvard, Jakobson wrote: “I am very happy that soon I will be able to work in closer contact with you.”[20] Wiener sent Jakobson “many thanks for what you have to say about Cybernetics”, expressed the hope to be able to work with him in the future, and concluded: “Please remember that everybody exposed to the philological virus, and I am, is extremely interested in matters philological.”[21] A few years later, when Jakobson took the chair at Harvard University formerly held by Wiener’s father, Wiener wrote to Jakobson a very warm letter, which, despite the clumsy syntax, expressed his feelings clearly: “As a successor of my father’s at Harvard, I cannot but think of you as a sort of quasi-relative.”[22]
Wiener followed Jakobson’s research and cited it in his book, The Human Use of Human Beings, referring, in support of his argument, to Jakobson’s and Benoit Mandelbrot’s view of human communication as a “game played in partnership by the speaker and the listener against the forces of confusion.”[23] Jakobson, in turn, gave Wiener credit for explaining teleology as control via feedback[24] and, in an attempt to accommodate Wiener’s language, spoke of the “feedback process between speaking and hearing.”[25]
Claude Shannon’s information theory played a more direct, instrumental role in Jakobson’s work. In April 1951, Jakobson wrote to Shannon that he was fascinated with one of Shannon’s papers and informed him: “I and my collaborators are involved now in problems of the application of your mathematical theory of communication to the phonemic analysis.”[26] In practical terms, Jakobson extended Shannon’s method of calculating the entropy of printed English to the analysis of spoken Russian.[27] On a theoretical level, Jakobson also saw fundamental correspondence between information theory and structural linguistics. He linked Shannon’s choice of binary digits (bits) as minimal units of information with his own earlier choice of binary oppositions as the structural basis for organizing phonemic distinctive features into a system. In Jakobson’s view, Shannon’s theory effectively extended Jakobson’s insight about the binary nature of spoken language to communication in general:[28]
The dichotomous principle underlying the whole system of distinctive features in language . . . has found corroboration in the binary digits . . . employed as a unit of measurement by the communication engineers. When they define the selective information of a message as the minimum number of binary decisions which enable the receiver to reconstruct what he needs to elicit from the message on the basis of the data available to him, this realistic formula is perfectly applicable to the role of distinctive features in verbal communication.
Jakobson applied the conceptual framework of information theory not only to distinctive features (as carriers of “phonetic information”), but also to the binary oppositions underlying morphological categories. According to Jakobson, since “nine binary choices underlie over one hundred simple and compound conjugational forms of an English verb,” the amount of “grammatical information” carried by the English verb can be calculated and compared to those of the English noun or the verb and noun of another language.[29]
Jakobson saw in the language of cybernetics and information theory a “modern, less ambiguous terminology”[30] and employed cyberspeak in parallel with more traditional linguistic terminology, creating his own distinctive vocabulary. Instead of the Saussurean opposition langue vs. parole, he often spoke of “code” and “message”; “subcodes” often replaced “styles”; speaker and listener now became “encoder” and “decoder,” and so on. Jakobson clearly modeled his vision of human linguistic communication on the pattern of Shannon’s engineering notion of a “communication system.” Shannon himself presented his well-known scheme as consisting of five parts:[31]
1) the information source, which produces a message or sequence or messages to be communicated to the receiving terminal;
2) the transmitter, which operates on the message in some way to produce a signal suitable for transmission over the channel;
3) the channel is merely the medium used to transmit the signal from the transmitter to receiver;
4) the receiver ordinarily performs the inverse operation of that done by the transmitter, reconstructing the message from the signal;
5) the destination is the person (or thing) for whom the message is intended.
Other essential elements of the scheme were (6) the message itself (written or spoken words, pictures, etc.) selected out of a set of possible messages; (7) the information, which is “a measure of one’s freedom of choice when one selects a message”; it “must not be confused with meaning”;[32] (8) the signal, a sequence of discrete symbols or a continuous function which is sent over the communication channel; and (9) the coding process, which changes the message into the signal.
Jakobson transferred this scheme, with minimal corrections, to human communication. He eliminated all traces of human-machine distinction from Shannon’s scheme and combined the information source with the transmitter in the notion of an “addresser”; similarly, the receiver and the destination were combined in the notion of an “addressee.” Jakobson added a “psychological” dimension to the physical channel of communication and called this new concept a “contact”; the message and the signal were not distinguished and subsumed under the rubric of “message.” Instead of the coding process, Jakobson preferred to speak of the “code”; this code, however, established the correspondence not between the message and the signal, as in Shannon’s scheme, but between the message and its meaning. And finally, most importantly, Jakobson replaced the notion of information with the referential meaning of a message, or “context,” in his terms. Jakobson’s overall scheme of linguistic communication involved the total of six “constituent factors”:[33]
1) the addresser (a speaker and an encoder);
2) the addressee (a listener and a decoder);
3) the context (a “referent” in “another, somewhat ambiguous, nomenclature”);
4) the message sent by the addresser to the addressee;
5) the code that is “fully, or at least partially, common” to them; and
6) the contact, a “physical channel and psychological connection” between them.
Jakobson utilized this scheme to develop a new vision of human language. He postulated six “functions” of language, each of them related to the corresponding factor in the scheme above:
1) the emotive (expressive) function aims at a “direct expression of the speaker’s attitude toward what he is speaking about”;
2) the conative (imperative) function brings focus on the addressee and finds its “purest grammatical expression in the vocative and imperative”;
3) the referential (cognitive, denotative) function has an “orientation toward the context”;
4) the poetic (aesthetic) function is the dominant function of verbal art, for it sets “focus on the message for its own sake”;
5) the metalingual (metalinguistic) function is used “whenever the addresser or the addressee needs to check up whether they use the same code,” for example, when the meaning of words is discussed;
6) the phatic function (Bronislaw Malinowski’s term) serves “to establish, to prolong, or to discontinue communication, to check whether the channel works,” and so on.
In Jakobson’s view, this new conceptual apparatus made explicit certain things which traditional linguistics could not see. He argued, for example, that, unlike “the less ‘biased’ communication engineers,” many linguists overlooked the phenomenon of “code variability,” pertinent to the use of metaphors: “Metaphoric creations are not deviations but regular processes of certain stylistic varieties, which are subcodes of an overall code.”[34] Jakobson discussed these issues in the 1960 article, “Linguistics and Communication Theory,” which he dedicated to the memory of his father, an engineer.
In 1957, a new journal Information and Control was started in the US; Jakobson served on its editorial board along with Wiener and Shannon. Jakobson was also among the founders of the Center for Communications Sciences at MIT.[35] At this Center, affiliated with MIT's Research Laboratory for Electronics, he collaborated with electrical engineers, mathematicians, psychologists, logicians, and neurophysiologists to study communication between humans as well as in man-made systems. Jakobson’s research resulted in three chapters for a projected book on sound and meaning.[36] He also took part in the work of the Center for Advanced Study of Behavioral Science at Harvard University[37] and the Harvard Center for Cognitive Studies.[38] When the American Society for Cybernetics was organized in 1965, Jakobson became its member.[39]
In the application of his apparatus to solving various practical problems, Jakobson saw an opportunity to put linguistics in a larger interdisciplinary framework and to expand his approach to other fields. In particular, he viewed structural linguistics as the basis for a general theory of structurally organized sign systems: “Language as one of the sign systems, and linguistics as the science of verbal signs, is but part of semiotics, the general science of signs.”[40] Jakobson traced the origins of main semiotic ideas to John Locke’s “semeiotike,” Charles Pierce’s “semiotic,” and Saussure’s sémiologie. Jakobson considered his six-element scheme of linguistic communication applicable to all semiotic systems; he maintained that “the comparative study of poetry and other arts as a teamwork of linguists and experts in such fields as musicology, visual arts, and so forth stands on the agenda.”[41] At the same time, he argued that “confronting” natural language with other sign systems, in particular, with scientific or technical languages, might illuminate some of the universal and particular properties of semiotic systems. For example, he contended that “in the relation between context-free and context-sensitive structures, mathematics and customary language are the two polar systems, and each of them proves to be the most appropriate metalanguage for the structural analysis of the other.”[42]
In the end, Jakobson arrived at a certain hierarchy of sciences: poetics (the study of poetic verbal messages) is part of linguistics (the study of communication by verbal messages in general), which is part of semiotics (the study of communication by any type of messages), which, in turn, is included into the host of communication sciences, along with social anthropology and economics.
Following Lévi-Strauss, Jakobson argued that “in any society communication operates on three different levels: exchange of messages, exchange of commodities (namely goods and services), and exchange of women (or, perhaps, in a more generalizing formulation, exchange of mates).”[43] Jakobson maintained that linguistic exchange was subject to the same kind of analysis as economic exchange, and therefore linguistics had much in common with economics. Lecturing in Japan in 1967, he told his audience a story involving himself and the famous Harvard economist Vassily Leontieff, the pioneer of input-output analysis:[44]
[Leontieff] lectured and made the introduction to economics and I had to lecture in the same class immediately after him on linguistics. And when I entered, there was the big scheme of the economic exchange, students wanted to erase it and I said, “Stop, I will lecture with this scheme.” Because the problems of output and input [in linguistics and economics] are exactly the same.
In these lectures, Jakobson recalled that he had once been accused of being an advocate of “God-given truth.” He replied: “No. I am not for ‘God-given truth.’ I am for a ‘code-given truth.’ . . . a truth given by the real code of language.”[45]
Structuralism Condemned: Ideological Campaigns in Soviet Linguistics
The late Stalinist period was marked by a series of ideological campaigns in the sciences and the humanities, which affected virtually all disciplines. With Stalin’s personal approval, Trofim Lysenko delivered the speech, “On the Situation in Biological Science,” at the July-August 1948 session of the Lenin All-Union Academy of Agricultural Sciences (VASKhNIL).[46] At this session, Lysenko outlined the irreconcilable opposition of “two trends in biology”—the “idealist” genetics and the “materialist” “Michurinist biology”—and set an exemplar for other scholarly fields. Physicists and psychologists, engineers and geologists, historians and linguists—all had to discuss how to reorganize their work “in light of decisions of the VASKhNIL session.” Numerous academic institutions held meeting patterned on the VASKhNIL model, including the phrasing of the title of the principal report. At a meeting at the Institute of Language and Thought, for example, its director Ivan Meshchaninov gave a report entitled “On the Situation in Linguistic Science.”[47]
In accordance with the exemplary scenario of Lysenko’s attack on genetics, linguists identified their own Russian-born founding father (this lot fell upon the late Nikolai Marr), the foreign-born reactionary scientists (such roles were often assigned to Wilhelm von Humboldt and Ferdinand de Saussure), and the evil doctrines to be weeded out (the part of “formal genetics” was played by comparative historical linguistics and structuralism).[48] Imitating Lysenko, one of the most active Marrists Fedot Filin spoke in October 1948 of “two trends in linguistics” and labeled structuralism a “modern idealist [trend in] linguistics, which has reached a dead end.”[49]
The Lysenkoist attacks and the corresponding campaigns in other disciplines were part of a general Cold War struggle against foreign influence (“kow-towing before the West,” in the language of Soviet ideologues). The next phase of this crusade took the form of a virulent anti-Semitic campaign against “cosmopolitanism,” which labeled Jews as “rootless cosmopolitans.”[50] Structuralism became a convenient target in this campaign too; the foreign origins and the emphasis on universal structural features of language made structuralism a perfect propaganda case for a “cosmopolitan” doctrine and provided the basis for accusations of the malicious neglect of the specificity of national languages.[51]
Jakobson, whose emigration to the West and “cosmopolitan” (i.e., Jewish) origins made him particularly vulnerable, became a very popular target of criticism during the Marrist and the anti-cosmopolitanism campaigns. Jakobson’s political “faults” threw a dark shadow over his linguistic doctrine, which was now associated with “cosmopolitanism” and “bourgeois linguistics.” One critic combined these political labels with the critique of excessive “formalism” in structural linguistics to construct the image of an academic enemy in which political and cognitive flaws were closely integrated:[52]
Structuralism is one of the most reactionary trends in modern bourgeois linguistics. This militant formalist doctrine is mobilized completely for the struggle against Marxism in linguistic science and embodies the basic features of bourgeois cosmopolitanism. . . . The theoretical and organizational formation of structuralism in linguistics is inextricably connected with the names of the typical bourgeois cosmopolitan scholars Roman Jakobson and Prince Trubetskoi—people who relinquished their motherland.
This critic charged that the Moscow Linguistic Circle represented a “pillar of formalism, a reactionary bourgeois trend, which attempted to poison the mind of the Soviet intelligentsia . . . It is characteristic that Jakobson has currently retreated to the USA—a bastion of bourgeois reaction, the principal source of imperialist aggression.”[53] The critic objected to the “reduction of the essence of phenomena to a system of formal oppositions” and accused structuralists of idealism.[54] As a healthy alternative to this “relic of bourgeois ideology,” the critic cited Marr’s “new doctrine of language” with its “dialectical unity of language and thought,” untainted by any formal representation.[55]
In 1950, however, the situation in Soviet linguistics drastically changed. Stalin suddenly distanced himself from the “excesses” of the policy, which he himself had actively promoted—an ingenious strategy, which might be called “ten steps forward, one step back.”[56] In June 1950, he published in Pravda the article, “Concerning Marxism in Linguistics,” in which he condemned Marr’s “new doctrine of language,” censured the “Arakcheev-like tyranny” in science, condemned the monopoly of a single approach, and called for “freedom of criticism.” His article validated the traditional comparative historical linguistics of the nineteenth century, which stood in sharp contrast to the class-based rhetoric and fantastic speculations of Marr. Even the most trivial statements, however, could prove a menace if turned into a dogma, which all Stalin’s pronouncements surely were. “Stalin’s doctrine of language” was quickly canonized with all its accidental blunders, like the claim that the hitherto unknown “Kursko-Orlovskii dialect” served as the basis for the Russian language.[57] Endless praise of Stalin’s linguistic “genius” even became the object of sarcasm in the well-known song, which originated in the Stalinist GULAG:[58]
Comrade Stalin, you’re a truly great savant,
By your knowledge of linguistics I am stunned.
I’m a humble Soviet inmate with no roots
And no comrades but gray wolves in the woods.
For the right or wrong reason, the monopoly of Marr’s “new doctrine of language” was gone, and its place was taken by traditional comparative historical linguistics. Structural methods in phonology were no longer a taboo, but traditional linguists still preferred to study the features of each phoneme in isolation from other phonemes and depreciated Jakobson’s attempt to arrange phonemes into a unified phonological system and to measure its characteristics. In 1954 Jakobson’s estimate of the average number of distinctive features of the Russian phoneme (5.79) was thoroughly ridiculed in the leading Soviet linguistic journal:[59]
It is absolutely clear that the indicated method [of Jakobson] does not inquire into the very matter of language—in this case, the phoneme—and only registers the average number of phonemic features taken in an abstract form. These figures have no direct meaning for linguistics; for what can, say, a phoneticist obtain from the figure 5.79 for a real description of some phoneme of the Russian language? Nothing! . . . The structural approach is not a method of linguistic research but merely a technique of formal mathematical representation of the results of research, and this form of representation does not add any new knowledge of language.
In the politically charged context of academic discussions, the denunciation of superfluous mathematical formalisms led immediately to the accusation of “formalism,” from which it was a short road to the charge of “idealism,” a major deviation from the Soviet version of Marxism. Although not explicitly forbidden, structural methods remained suspicious.
Czech structural linguists who tried to protect their research by dissociating themselves from the notorious figure of the “bourgeois linguist” Jakobson added oil to the fire. In 1951 the Czech press called Jakobson “the genuine evil spirit of our linguistics.” “The role of Jakobson as the chief pillar of structuralism in linguistics,” wrote one former member of the Prague Circle, “is one of the refined ideological weapons used for the disorientation of the outstanding representative intellectuals of the left and for the struggle against the proletarian Weltanschauung.”[60] In 1953 Problems of Linguistics reprinted an article by another Prague Circle linguist, who castigated the “sworn enemies of Marxist philosophy and socialism, such as Roman Jakobson, who are deliberately preaching idealist nonsense.”[61] Ironically, at about the same time in the United States one Polish-American activist demanded that Jakobson “should be registered as a foreign agent because of the sympathies for the Marxian ideology in his writings.”[62] Jakobson’s role as a mediator between linguists in the West and in the East became equally unacceptable to both sides.
Jakobson in Moscow
Khrushchev’s political “thaw” in the middle 1950s opened possibilities for reestablishing international scientific contacts frozen during the Cold War. In early 1956, for the first time after he left the Soviet Union in 1920, Roman Jakobson was invited to visit Moscow to attend the first meeting of the International Committee of Slavists as the American representative. The Soviet Embassy in Washington, however, reported that, according to an unnamed “French correspondent,” Jakobson allegedly had made “hostile remarks” about the Soviet Union. The question whether Jakobson should receive an entry visa was debated at the level of the Party Central Committee. In April 1956, the head of the Department of Science and Institutions of Higher Education Kirillin reported the results of his investigation to the Central Committee:[63]
According to the Cadres Administration of the Ministry of Foreign Affairs of the USSR (R. Jakobson was on diplomatic service for a certain period of time), the following is known about R. Jakobson. He was born in Moscow in 1896, a Jew, received his education in Russia. After the October Revolution he worked at the People’s Commissariat of Enlightenment, at the All-Russia Council on the National Economy, and after 1920 at the Soviet embassies in Revel and Prague (the last office: the head of the press bureau of the Czech Soviet Republic); since 1928, he has been a nevozvrashchenets.[64] Before the Second World War he lived predominantly in Prague and after that in the USA.
R. Jakobson is currently professor at Harvard University and a Slavist well-known abroad; he is one of the founders of the Prague Linguistic Circle and a prominent member of the modern school of structural linguistics (structuralism, a trend in bourgeois linguistics, has spread rather widely in Western Europe and the USA). Many works of R. Jakobson are concerned with the study of the phonological system of the Russian language. He has also done research in literary studies (the study of poetic meters, including studies based on Russian literature). He has works on Maiakovskii, which were written in 1921-23, about this poet’s rhythmics and language; according to the opinion of Soviet specialists on Maiakovskii, these works, despite their general formalist flavor, contain a large number of valuable, correct observations. Unlike some of the bourgeois scholars who try to deny the authenticity of the famous monument of Russian literature, “The Tale of Prince Igor’,” R. Jakobson takes the right position on this issue.
Academician V.V. Vinogradov, who met with R. Jakobson personally last year at the meeting of Slavists in Belgrade, reports that Jakobson acted toward Soviet scholars with good will and allowed himself no hostile remarks.
A refusal to issue him an entry visa to the Soviet Union might provoke an undesirable reaction; this could become a pretext for hostile outbursts against the Soviet Union in foreign countries.
The Department of Science and Institutions of Higher Education of the Central Committee believes it is possible to give R. Jakobson the permission to visit Moscow for the meeting of the International Committee of Slavists.
The Party authorities eventually allowed Jakobson to come, but decided to keep him under surveillance.
During his visits to Moscow for a session of the International Committee of Slavists in May 1956 and the 4th International Congress of Slavists in September 1958, Jakobson met informally with a number of young scholars and discussed with them the innovations brought into linguistics by information theory. In particular, Viacheslav Ivanov—Assistant Professor of Linguistics at Moscow University and acting deputy editor-in-chief of the journal Problems of Linguistics—met with Jakobson in May 1956 and informed him about recent Soviet works in this field. The authorities, who closely monitored Jakobson’s contacts, undertook an investigation of his private meetings. As Ivanov later recalled, the leadership of the Philology Faculty of Moscow University had interrogated him about his meeting with Jakobson:[65]
[I was asked] whether I was present at Roman Jakobson’s hotel room at a secret meeting with scholars from socialist countries. It was assumed that he called this meeting with subversive purposes. This intelligence information . . . was right in one respect: I indeed visited Jakobson in his room (which was against the Soviet citizen’s rules of cautious behavior), because he had asked me to bring him the issue of Uspekhi matematicheskikh nauk with an article by his schooldays friend Khinchin on information theory, about which we were all very excited then. Certainly, I did not tell [them] that.
In October 1956, as a follow-up to their meeting, Ivanov sent Jakobson a copy of the famous mathematician Andrei Kolmogorov’s paper, “The Theory of the Transmission of Information”;[66] Jakobson quickly incorporated some of Kolmogorov’s ideas into his work.[67]
While in Moscow, Jakobson seized the opportunity to speak at several academic institutes before large audiences. Speaking in the Department of Translation at the Institute of Foreign Languages, he emphasized the applicability of his six-partite scheme of linguistic communication to translation.[68] Researchers at this Institute wholly embraced this view and later generalized it to machine translation.[69] Jakobson’s ideas gradually began to find their way into Soviet linguistics.[70] In particular, Soviet linguists further developed his controversial phonological theory of distinctive features and applied it to describe sound patterns of various national languages; in the debate between structuralism and traditional comparative historical linguistics, Jakobson’s phonology became the “cornerstone” of radical partitions within the linguistics community.[71] Jakobson’s emphasis on semantics, which distinguished his approach from most contemporary Western theories, played a particularly important role in Soviet structural linguists’ turn to formal studies of meaning. In the spirit of Jakobson’s overarching structuralist project, Soviet linguists extended the methods of structural analysis developed in phonology, morphology, and syntax, to semantics.
In a piecemeal fashion, it became possible to cite Jakobson in Soviet scholarly literature, to review his works, and even publish some of his papers in Russian. A volume of his collected works prepared for publication in 1963; with the end of Khrushchev’s thaw, however, the manuscript was shelved and later completely discarded.[72] Although in the 1970s Jakobson was no longer a politically controversial figure and his works were widely read and cited by Soviet linguists, a collection of his works came out in Russian only in 1985, three years after his death.[73]
Structural Linguistics Meets Mathematics: Mathematical Linguistics
In June 1956, soon after Jakobson’s first visit to Moscow, two young assistant professors of Moscow University—linguist Viacheslav Ivanov and mathematician Vladimir Uspenskii (a former student of Kolmogorov’s)—met to discuss the agenda for an interdepartmental research seminar on mathematical linguistics, which they intended to organize at the University. They drafted the following list of key topics:[74]
1. Statistics. 2. Machine translation. 3. The mathematization of language. Specificity. Mathematical definition of grammatical categories.
4. Mathematical logic (Syntax. Reichenbach.[75]) and information theory.
5. Perhaps, some other topics.
Uspenskii recalled that they were driven by the “irrational urge to find in language rigorous laws, similar to mathematical ones in their rigor.”[76] The agenda for the “mathematization” of linguistics was set. To give their seminar more legitimacy, Ivanov and Uspenskii persuaded Petr Savvich Kuznetsov—full professor of the Philology Faculty of Moscow University—to join them. Kuznetsov was one of the leading Soviet phonologists; as early as 1948-49, he took considerable political risks and introduced elements of Jakobson’s phonological theory in his courses at Moscow University.[77] Kuznetsov was also Kolmogorov’s childhood friend and was familiar with mathematics, unlike many other philologists on the faculty. In September 1956, the seminar was announced under the title, “Some applications of mathematical methods of research in the study of language”; according to Uspenskii, “the wordiness of the title was the result of the fear that the words ‘mathematical linguistics’ might provoke anger from the alarmed authorities.”[78] The seminar attracted a mixed audience of mathematicians and linguists, which was unusual for both disciplines. At the first session of the seminar, Uspenskii posed before its participants two problems formulated by Kolmogorov: to give formal definitions of the grammatical case and the iambic poetic meter.
It was no accident that the seminar started off with definitions; the mathematicians among its participants were not content with the intuitive notions, to which linguists were accustomed. Uspenskii expressed the opinion prevalent in Kolmogorov’s circle, when he argued that a mathematician in linguistics “plays the part of a litmus test: if a definition satisfies the mathematician, then it must satisfy anyone” and considered striving for unambiguous definitions one of the “specific traits of the mathematical style of thinking.”[79] Uspenskii even criticized Jakobson’s work, “Toward the Logical Description of Languages in Their Phonemic Aspect,” much admired by Soviet structural linguists, for not being sufficiently logical.[80] In 1957, Uspenskii offered an extracurricular mathematics course for linguists, where he taught linguistics students some of the principles of mathematical thinking, which, in his view, they specifically lacked—the principles of “clear explication of major abstract concepts, delineation between the definable and indefinable, between the deductive and the inductive, etc.”[81]
Along with the formulation of two problems for the seminar, Kolmogorov suggested possible leads to their solution. For the case definition problem, he proposed to associate each noun with a number of possible states that can be occupied by the object described by this noun. He then called two states “equivalent” if they were expressed by the same morpheme. Further, two states were called “absolutely equivalent” if they were equivalent in regard to any object that could occupy these states. Finally, Kolmogorov proposed to group all states into classes so that within each class all states would be absolutely equivalent, while any two states from different classes would not be absolutely equivalent. These classes, in Kolmogorov’s view, could then be called grammatical cases.[82] Jakobson considered this an “apt definition.”[83] If directly applied, however, Kolmogorov’s definition produced far too many cases for the Russian language. Soviet structural linguists translated Kolmogorov’s concepts “object” and “state” back into such linguistic notions as “word” and “context” and further developed his approach, trying to reduce the number of case-generating contexts.[84]
The significance of Kolmogorov’s interest in the iambic meter went far beyond a simple desire for clarity and rigor. He considered poetry to be a unique window into human mind amenable to mathematical analysis. In the fall of 1960, Kolmogorov gave at Moscow University a series of lectures on “The theory of probability and the rhythmic analysis of Russian poetry,” which was quite unusual for a professional, especially high-standing, mathematician. In these lectures, he inquired, “why not to choose the creation of poetry as an example of [mental] activity? Is it any worse than to choose salivation as such an example?”[85]
Kolmogorov devoted a number of articles to the study of various meters and rhythms in Russian poetry.[86] In traditional poetics, the iamb was defined as a metrical foot consisting of an unstressed syllable followed by a stressed syllable. Unfortunately, in practice this strict rule was so often violated in iambic verse that it could not serve as an acceptable definition. For example, in Pushkin’s Eugene Onegin, perhaps the most well-known classical iambic poem of the Russian language, almost three-fourths of its 5300 lines violated this rule and more than a fifth of all even syllables were not stressed.[87] Nevertheless, an intuitive perception of the Onegin lines as iambic suggested that this traditional definition might not capture the actual diversity of iambic poetry. Kolmogorov attempted to find a definition that would incorporate such cases and would be still consistent with the intuition of a poetry reader.
Kolmogorov incorporated into his work Roman Jakobson’s and Vasilii Trediakovskii’s observations of the specificity of the iamb, although they “did not think in terms of giving a definition.”[88] Eventually, he arrived at the following definition: a poetry line was a correct iambic line if it did not contain any words violating the iambic structure, i.e. the words that contained more than one syllable and were located in the line in such a way that the stress fell on an odd syllable. Again, Kolmogorov’s definition was further developed by professional philologists (the requirement that the last strong syllable in the line must be stressed was added) and incorporated into their work.[89]
Kolmogorov studied patterns of deviation of actual stress structures from classical meters in the poetry of various authors, and concluded that statistical characteristics of such deviations could serve as a valuable analytical tool. The most frequent sort of deviations typical of a particular poet would provide a unique “statistical portrait” of this poet; a less frequent kind characteristic of a particular poem would signify the “general coloration” of this poem; and the least frequent type would indicate a special device used by the poet to emphasize a particular passage.[90] In connection with Kolmogorov’s studies of stress deviations, structural linguist Isaak Revzin outlined his vision of the distinction between traditional philology, based on intuition, and mathematical linguistics, grounded in strict verification procedures:[91]
Literary theorists usually rush to one of the two extremes: they either directly link certain [literary] devices with particular content, or deny any connection between the two. Statistics, on the other hand, makes it possible to evaluate whether certain hypotheses, formulated originally on the basis of the researcher’s intuition, are correct.
Kolmogorov regularly published his findings in the journal Problems of Linguistics; this line of research, however, abruptly came to an end, when Pravda published the article, “Pushkin on a Diagonal,” which severely criticized one statistical study of Pushkin’s poetry.[92] This study had been published in the journal The Theory of Probability and Its Applications, edited by Kolmogorov.[93] After that, according to Ivanov’s memoirs, Kolmogorov “was scared to publish his works on poetry.”[94] Many works of Kolmogorov’s group remain unpublished; Uspenskii even expressed the concern that they might never be published.[95]
Measuring the “Information Content” of Poetry
Although the attempts to apply information theory to linguistics looked like an extension of this theory into a new field, in fact they effectively brought information theory back to its own origins. Claude Shannon’s model of a discrete source of information was based on the apparatus of Markoff processes, which described a stochastic system with a finite number of states, where the transition probabilities that governed the system’s passage from one state to another depended only on the current state. “To make this Markoff process into an information source we need only assume that a letter is produced for each transition from one state to another,” wrote Shannon.[96] Interestingly, Markoff processes had been originally introduced and analyzed in the early 20th century by the Russian mathematician Andrei Markov, Sr., precisely for the purpose of stochastic description of letter sequences. In 1913, Markov published in the Bulletin de l’Academie Imperiale des Sciences de St. Petersbourg a statistical study of Pushkin’s Eugene Onegin, which illustrated his theory.[97]
Shannon’s 1951 article, which so fascinated Jakobson, described a method of measuring the entropy of printed texts by conducting experiments, in which subjects would be shown a portion of text and asked to guess the next letter, then this letter would be revealed and the next letter guessed, and so on. Shannon assumed that native speakers of English would possess an intuitive knowledge of letter frequencies in English texts, and would use their knowledge in this guesswork. By the results of their guesses, he evaluated such frequencies and then determined the entropy of the text as defined by the information theory.[98] Entropy here served as a direct measure of information: the more difficult it was to predict a sequence of letters, the greater was the entropy of the text, and the more information this text could be said to provide.
A group of Soviet researchers modified Shannon’s method by eliminating “zero information” choices from the final score (i.e., those situations where the letter could be determined unambiguously, and there was no “true” guessing). They conducted a series of experiments with different styles of texts: business style (scientific and political texts), belles-lettres style (literary texts), and the dialogue. They concluded that the highest redundancy (i.e., lowest entropy) was characteristic of the business style, for the following reasons: (1) the use of a large number of stable word combinations; (2) limited vocabulary, a large part of which consists of special terminology; and (3) logical and strictly normalized construction of sentences. The use of metaphors and the avoidance of stock phrases made the redundancy of literary texts lower, and the lowest redundancy was found in the dialogue, where the frequent use of stereotypical expressions was said to be fully compensated by the lack of strict normalization and a greater freedom of choice.[99]
Kolmogorov’s group also modified Shannon’s method; in this case, experimental subjects could express the degree of their confidence in the guess (high confidence, low confidence, or no confidence—i.e., refusing to guess) and also suggest a set of two or three possible letters instead of one.[100] This group, too, conducted experiments with various types of texts and measured their entropy. Kolmogorov was particularly pleased to remark (in private) that from the viewpoint of information theory newspapers were less informative than poetry since political discourse employed a large number of stock phrases and was highly predictable in its content.[101] The verses of great poets, on the other hand, carried much more information, despite the strict limitations imposed on them by the poetic form, for poetic expressions were much more difficult to predict. Banal poems, however, resembled newspapers more than artistic texts. “I remember Kolmogorov’s excitement,” recalled Ivanov, “when he correctly guessed the entire content of some poem dedicated to an official holiday and printed in a calendar, after reading only the first few lines. Such texts contain zero information.”[102] Kolmogorov’s students greatly enjoyed this type of research, occasionally making playful and ideologically dubious remarks even in serious scientific publications. Roland Dobrushin, for example, wrote that “the probability of the word ‘socialism’ appearing in today’s newspaper is greater than zero, while for the works of Pushkin, the same probability is nil.”[103]
According to Kolmogorov, the total entropy of poetic language has two components: “information capacity” and “language flexibility.” The first component reflects the capacity of a text of a given length to express various meanings, or “extralinguistic” content (in simplified terms, the “total number of ideas that can be expressed in a message” of a given length). The second component pertains to “linguistics proper,” for it describes the diversity of means—through synonymy, metaphor, and so on—to express a given meaning (in other words, the “total number of ways to express the same content in a given language”).[104] Kolmogorov also introduced another entropy-like parameter—the complexity of a text—which described the formal constraints imposed upon a text (for example, meter and rhyme in poetry). The complexity of a text cannot exceed the author’s language flexibility, otherwise the author’s vocabulary would not be diverse enough to contain, say, an appropriate synonym or metaphor for a rhyme—a constraint similar to the limitation on channel’s information capacity in Shannon’s theory. Therefore, the more complex is poetic form, the more flexible must be the poet’s language; otherwise “not the poet would control the poem, but the poem would control the poet.”[105] According to Kolmogorov, only in artistic texts is language flexibility used consciously to create an emotional impact, and in this specific sense artistic works could be said to contain more information than ordinary texts.[106]
Entropy as Complexity
In the Soviet version of cyberspeak, entropy—one of the central notions of cybernetics—obtained a new interpretation, which went beyond narrow parallels with thermodynamics and opened up new possibilities for a formal description of complex systems. While Shannon interpreted entropy as a measure of uncertainty, and Wiener as a measure of disorder, Kolmogorov viewed it as a measure of complexity. Kolmogorov put forward two alternatives to the probabilistic (Shannon’s) approach to the definition of information: the combinatorial approach (drawing on the work of Ralph Hartley) and the algorithmic approach. “I feel it is an important problem to free oneself, whenever possible, of unnecessary probabilistic assumptions,” he wrote.[107] The main problem Kolmogorov associated with the probabilistic approach was that it precluded the possibility to calculate the amount of information in the case of a unique message:[108]
The probabilistic approach is natural in the theory of communication along transmission channels of “mass” information consisting of a large number of disconnected or weakly connected messages satisfying certain specific random laws.
But what is the real life significance, for example, of saying what the “amount of information” contained in the text of “War and Peace” is? Is it possible to include this novel in a reasonable way into the set of “all possible novels” and further to postulate the existence of a certain probability distribution in this set? Or should we consider separate scenes from “War and Peace” as forming a random sequence with “stochastic relations” disappearing quickly enough within the distance of several pages?
Kolmogorov proposed an algorithmic approach for measuring the amount of information in one individual object relatively to another individual object. “Real objects which are to be studied are very (unboundedly?) complex,” he argued, “but the relationship between two really existing objects are fully assessed in a simpler schematic description. If a geographical map gives us important information about a section of the Earth’s surface, the microstructure of paper and paint appearing on the paper has no relationship to the microstructure of the section of the Earth’s surface shown on this paper.”[109] Kolmogorov’s algorithmic approach was “founded on the application of the theory of algorithms to the definition of the notion of entropy or complexity of a finite object and of the notion of information in one finite object about another.”[110]
Kolmogorov defined the “relative complexity” of an object (depending on the “method of programming”) as the minimal length of the “program” that can produce this object.[111] “If some object has a ‘simple’ structure,” he explained, “then for its description it suffices to have a small amount of information; but if it is ‘complex,’ then its description must contain a lot of information.”[112] Accordingly, if the method of programming is elaborate, the length of the program can be small even for a very complex object; a simple method, on the other hand, would produce a much longer program.
Kolmogorov concluded that, within the algorithmic approach, “such expressions as the ‘complexity’ of the text of the novel ‘War and Peace’ must be viewed as being practically uniquely determined,” if one could evaluate the amount of a priori information about the language, style, and content of the text.[113] Kolmogorov’s complexity theory can be seen as a very telling example of a new mathematical theory emerging partly in response to the problem of measuring the “information content” of a text—a problem brought into linguistics by cybernetics.
Structural Linguistics Meets Machine Translation: Applied Linguistics
On January 7, 1954, the first public demonstration of machine translation from Russian into English took place in the New York office of the IBM Corporation and soon an article reporting the event appeared in the journal Computers and Automation.[114] In September 1954, director of the Institute of Scientific Information[115] Dmitrii Panov published a brief summary of this article in a Soviet mathematics review journal.[116] Panov soon became deputy director of Institute of Precise Mechanics and Computer Technology (ITMVT)[117] and organized a machine translation group there. In the end of 1955, the group tested on BESM computer the first version of their program, which included an English-Russian translation algorithm and a dictionary of 952 English and 1073 Russian words. Panov professed a pragmatic approach and preferred to develop a large set of ad hoc rules instead of building a general formal model of language. “A study of linguistic structures may be interesting from many different viewpoints,” he reasoned, “but from the point of view of translation, such a study seems to be directed at issues of secondary importance and actually distracts from solving main problems.”[118]
Soon after Panov’s review came out, mathematician Aleksei Liapunov set up a similar group at the Division of Applied Mathematics.[119] His graduate student Olga Kulagina and linguist Igor’ Mel’chuk from Moscow University developed a program for the translation of French mathematical texts into Russian; soon linguist Tatiana Moloshnaia joined the group and started working on an English-Russian algorithm. The French-Russian translation program used a dictionary of 1100 words and was implemented on STRELA computer. The initial French-Russian algorithm already included a few rules of analysis and synthesis, and the English-Russian machine translation project aimed at performing a comprehensive syntactic analysis. Liapunov’s group elaborated the concept of “elementary grammatical configurations,” which became the starting point for a number of formal linguistic models. Liapunov considered machine translation a “beyond-the-clouds” problem, which cannot be ultimately solved in its complete form, but can provide a general direction for specific projects in computational mathematics and linguistics. “The development of machine translation,” he wrote, “requires a productive use of both linguistic and mathematical-cybernetic methods and often raises substantial questions in both fields.”[120]
Liapunov believed machine translation was part of a great cybernetic revolution, the advent of which he prophesied in innumerable public lectures in assorted scientific institutions of Moscow and Leningrad throughout the 1950s. In the fall of 1954, he spoke at the Philology Faculty of Moscow University on computers and machine translation; the lecture reportedly ended with a scandal and Liapunov was deported from the room for spreading cybernetic ideas; at that time, the Soviet press still labeled cybernetics a “pseudo-science.”[121] Soon Liapunov gave a similar lecture at the Moscow State Pedagogical Institute of Foreign Languages.[122] Linguistics professor of this Institute Isaak Revzin recalled:[123]
He did not talk much about machine translation, rather said a few words in the end. [Instead,] he spoke of Cybernetics (I heard this word for the first time!) in a broad philosophical aspect, and the romantics of the first stories of fantastic prospects of cybernetics . . . surrounded the figure of Liapunov, who already looked like a preacher, with an aura of a pioneer.
He linked the question that had been haunting me since my student years—that of translation—with the entire host of complex human problems, from genetics and medicine to the problems of control, and this convinced me right away. I instantly decided that this was the way to go.
As reports on machine translation experiments in the West began to reach the ear of the authorities, they started actively to encourage this type of research. Capitalizing on the interest in machine translation by his superiors at the Institute of Foreign Languages and the Ministry of Higher Education, linguist Viktor Rozentsveig organized at this Institute the Department of Translation—of both human and machine type—and became its chair.[124] He also created the Association for Machine Translation,[125] which gathered for its first session in December 1956. In his memoirs, Uspenskii recalled:[126]
The status and the boundaries of the Association were intentionally left totally amorphous. No formal document constituting this Association ever existed. The term “association” was a very fortunate find: neither “institute,” nor “laboratory,” nor “society,” but an “association,” which associated nobody knew whom or what. It was far from clear—and this was the strength of the whole idea—who or what comprised this Association, and whether it was comprised of anything at all.
The Association had no formal membership; nevertheless, under the aegis the Institute of Foreign Languages the Association obtained the right to hold sessions and publish a bulletin.[127]
The wave of interest in machine translation finally reached the leadership of Moscow University. Rector Ivan Petrovskii, a prominent mathematician, strongly encouraged the spread of mathematical methods into new fields. In January 1957, he raised the issue of machine translation at a meeting where teaching and research plans of the humanities departments were discussed:[128]
I would like to draw your attention to [news] bulletins of TASS [the Telegraph Agency of the Soviet Union] . . . Here is what the latest issue says: according to an American newspaper, in the United States [computers] translate texts from Russian into English, mostly scientific texts. They have been studying problems of machine translation for ten years. In England, this problem has also been studied for a long time. As the result of work done at Georgetown University, they will be making experiments in January. . . .
Some work is being done in our quarters too. Apparently, we must engage in this business more intensively. And I think that this work would also aid the development of your philological science.
Dean of the Philology Faculty Samarin instantly recognized which way the wind was blowing and recalled the semi-legal existence of the research seminar on mathematical linguistics led by Ivanov, Uspenskii, and Co., which he had barely tolerated before but was most eager to claim to his credit now. Samarin confidently assured the rector that the Philology Faculty was keeping up with the newest developments, Assistant Professor Ivanov was working on this problem, and Samarin himself was just about to propose creating a special research group on machine translation. “This is a vast area [of research], and we must study it,” Samarin told the rector, “Give us a laboratory.” “We’ll help you in all possible ways,” said Petrovskii, “But you must take into account that machines cannot give an artistic translation.” Professor of the History Department and University provost Galkin immediately jumped in: “[The machines] will not give us much. Theirs is a primitive translation.” “At first, I was against this machine,” confessed Petrovskii, “but later realized that it was difficult to pass a judgment on such matter with assurance. At the beginning of this century, the Parisian Academy of Sciences refused to consider any design of aircraft heavier than air; they said it was nonsense.” “Such talks were heard about cybernetics too,” remarked dean of the Economics Faculty Sokolov. “Well, take railroads,” continued Petrovskii, “How many scholars claimed that it was a folly, that railroads could not possibly be useful! I do not think that this [machine translation] could be harmful.” Galkin again displayed Party vigilance and warned against possible ideological deviations: “This is not harmful, but one should not think that robots can replace everything. As a result, among the faculty spreads a formal approach, which dispenses with meaning in translation.”[129]
In this discussion, one could discern the overtones of the familiar ideological charges of “formalism” and the apprehension the authorities had about committing a possible ideological mistake. In the end, Moscow university leadership took little action in support of machine translation research. Uspenskii was allowed to teach mathematics to linguistics students off curriculum, but a laboratory for machine translation did not materialize. Traditional philologists of Moscow University realized that the development of research in machine translation would elevate the status of structural linguistics and mathematical linguistics and had strong reservations about that.
Structural linguists had to find support outside the linguistics establishment. In January 1958, they organized the Committee on Applied Linguistics under the Section on Speech Analysis of the Commission on Acoustics of the USSR Academy of Sciences.[130] This complex administrative arrangement provided a temporary protection from the watchful eye of the leadership of the Academy Division of Literature and Language, hostile to structural linguistics. The Committee served to coordinate the work of disparate groups of structural linguists scattered over various institutions and, together with the Institute of Foreign Languages, helped organize the first national conference on machine translation.
This conference took place in Moscow in May 1958 and attracted 340 participants, among them linguists, mathematicians, engineers, and physiologists.[131] The conference adopted a resolution calling for greater institutional support for research in this field. As a result of these efforts, in November 1958, the Ministry of Higher Education issued a resolution, “On the development of scientific research in the field of machine translation.” This resolution prescribed rectors of all Soviet universities to “provide strong support to those professors and researchers who conduct studies in the field of machine translation and mathematical linguistics” and to introduce extracurricular courses in machine translation and mathematical linguistics for students specializing in mathematics and philology in several major universities.[132]
In accordance with this resolution, the Institute of Foreign Languages established the Laboratory of Machine Translation headed by Rozentsveig. This “laboratory” had no equipment and only four “engineers” on staff (in fact, they were linguists), including Ivanov’s former students Alexander Zholkovsky and Yuri Shcheglov.[133]
Another “ghost” laboratory, whose equipment existed mostly in the form of blueprints, was the Laboratory of Electromodeling[134] (headed by Lev Gutenmakher), which split off ITMVT, wandered a few years on its own, and later joined the Institute of Scientific Information. Ivanov and Uspenskii worked there part-time: Ivanov as the head of the mathematical linguistics group, Uspenskii as the head of the mathematical logic group. The Laboratory studied the possibility of building an “information machine” and supported theoretical inquiries into logical analysis of natural language and the principles of construction of artificial “information languages” for various areas of science and technology.[135] The Laboratory regularly published its working papers; in May 1957, it organized a conference on information machines with over 500 participants.[136]
In addition to the conceptual disagreement with the linguistics establishment, some of the Soviet structural linguists soon ran into political trouble, which made their search for an institutional niche outside the linguistics community even more urgent. In December 1958, Ivanov was fired from Moscow University; soon linguists Mel’chuk, Vladimir Dybo, and Andrei Zalizniak, who protested against Ivanov’s dismissal, were forced to leave the University too.[137] The main charges against Ivanov were political: he maintained close links with Jakobson and with the Nobel prize winning poet Boris Pasternak. The resolution of the academic council of the Philology Faculty on the Ivanov affair read:[138]
In his explanations, Ivanov declared his disagreement with the judgment of the Soviet public and Party activists concerning the anti-Soviet novel Doctor Zhivago by Pasternak.
Besides, the commission [of the faculty] attempted to find out the nature of Ivanov’s relations with Roman Jakobson—a traitor to his motherland, a nevozvrashchenets, who is currently a US citizen. At a meeting of the commission, Ivanov denied any personal contacts or unofficial meetings with Jakobson. A few days later, however, speaking to the Party bureau secretary Iushin, Ivanov had to admit that he had played a mediating role in organizing a meeting between Jakobson and Pasternak at the latter’s dacha. Besides, in the same conversation, Ivanov spoke about another private meeting with Jakobson in Moscow.
Members of the commission have familiarized themselves with Ivanov’s papers presented at the International Linguistic Congress in Oslo and the International Congress of Slavists in Moscow, and also with his article published in the journal Problems of Linguistics, no. 5 (1958). From these materials it becomes clear that Ivanov by all means supports and popularizes Jakobson’s works at international conferences and in the Soviet press.
Such personal contacts with and apologetic attitude toward the ideas, conceptions, and the persona of Jakobson, an enemy of Marxism, are incompatible with the dignity of a Soviet patriot scholar.
Ivanov’s political troubles, as we can see, were closely linked with his scientific activity: it was his interest in Jakobson’s linguistics that brought Ivanov into conflict with the authorities. Science and politics went hand in hand; disagreement with dominant scientific views sowed political discord. Conversely, political complications engendered professional marginalization. Ivanov could not find a job in any linguistics institution and had to accept a position as the head of the machine translation group at ITMVT.[139] The meetings of Ivanov and Uspenskii’s seminar on mathematical linguistics at Moscow University were terminated; the center of activity in this field shifted to the machine translation laboratory at the Institute of Foreign Languages. Ivanov was also dismissed from the editorial board of Problems of Linguistics, and structural linguists now had to publish their work mostly as working papers of various machine translation groups. The only way to legitimize research in structural linguistics at the time seemed to be to emphasize its inextricable connections with machine translation, portraying these two fields as the theoretical and applied sides of the same coin.
Applied Linguistics Applies Itself to Theory
Structural linguists did not merely find a refuge in the field of machine translation; the compelled alliance with programmers, engineers, physiologists, and psychologists brought Soviet structural linguists into productive contact with a host of new ideas and approaches. Ironically, instead of adopting a more pragmatic approach that would help solve practical problems of machine translation, structural linguists focused their efforts on theoretical problems of linguistics. They employed the conceptual and mathematical apparatus of cybernetics to address some of the traditional linguistic problems and to create new formal models of language. They emphasized both the practical importance of structural linguistics for machine translation and the significance of machine translation research for theoretical linguistics. By the middle 1960s, they proposed “to view machine translation at current stage as a theoretical rather than practical problem.”[140] Referring to machine translation, Ivanov commented: “This area of applied linguistics so far has been applying itself to linguistics proper by fostering its transformation into an exact science and providing criteria for testing and selecting models of language.”[141]
All Soviet specialists in machine translation could be roughly divided in two groups. The first group, consisting largely of computer programmers, had access to computers and pursued a pragmatic approach based on creating efficient computer algorithms and increasing the size of machine vocabulary. The second group, to which most structural linguists belonged, had little or no access to computers and very limited opportunities to test machine translation algorithms in practice. This situation resembled a popular Soviet-era joke about someone’s order to fill a swimming pool with water only after everybody would learn how to swim. Liapunov’s group at the Division of Applied Mathematics, for example, could use only 5 minutes of computer time per week; whereas the translation of one sentence on the computer STRELA took about 3 minutes.[142] The position of Ivanov’s group at ITMVT was even worse. In the late 1950s, ITMVT, being the leading computer design organization in the country, did not have a computer of its own; all ITMVT researchers could use only five hours of computer time per week at the Computing Center of the Academy of Sciences. Because of the shortage of computer time, the machine translation group, which was considered of secondary importance within ITMVT, did not have access to computers at all. Their work, consequently, did not bring much practical outcome. Director of ITMVT Sergei Lebedev planned to transfer this group to the proposed Institute of Cybernetics.[143] The plans to organize this Institute never materialized; in 1961, ITMVT was transferred from the Academy of Sciences to the defense sector, Ivanov left, and the machine translation group dispersed. Another specialist in machine translation, linguist Isaak Revzin—one of the co-authors of The Fundamentals of General and Machine Translation—saw a computer from a distance only once in his lifetime.[144]
In this situation, structural linguists, whose own interests concerned theoretical linguistics much more than practical machine translation, seized the opportunity to shift the emphasis from building working systems of machine translation to elaborating general linguistic models. At the 1958 conference on machine translation in Moscow, the Leningrad linguist M.M. Steblin-Kamenskii outlined a broad agenda for the contribution of machine translation research into theoretical linguistics:[145]
1) the critique of all traditional grammatical concepts such as “sentence,” “parts of the sentence,” and “parts of speech”;
2) the demonstration that the same linguistic fact can be described in different ways, depending on definitions, and therefore all linguistic dogmas should be reconsidered;
3) the critique of linguistic “semantism,” or the inclination to study isolated meanings outside the context of the structure of language. Until the study of meaning produces the discovery of a finite number of elements interconnected through a system of regular relations—something similar to a system of phonemes with respect to the sound pattern of language—the notion of meaning would remain amorphous and unbounded.
This agenda partly reflected works already under way. At the same conference, Isaak Revzin argued that “the existence of more than two hundred different definitions of the term ‘sentence’ makes it impossible to develop [a theory of] syntax on rigorous deductive principles and shows that the definition of basic linguistic units must be approached differently.”[146] Revzin particularly attacked such definitions, typical of Soviet traditional linguistics, as “a sentence is a more or less completed thought” and proposed to formulate a new definition in terms of set theory, which he (after Liapunov and Kulagina) considered an appropriate meta-language for linguistics.
Other basic linguistic concepts were revisited by Igor’ Mel’chuk, a former student of Viacheslav Ivanov and an active participant in machine translation research in Liapunov’s group. Despite his more than modest administrative status, Mel’chuk’s colleagues regarded him as an “informal leader” of Soviet structural linguistics.[147] In particular, he criticized the vagueness of such terms as “a stable word combination” and “an idiom.” Mel’chuk proposed to measure the degree of stability of word combinations (by comparison with the number of other combinations a given word could enter) and to call “stable” only those word combinations whose degree of stability exceeded a certain threshold. Similarly, he spoke of the possibility to evaluate the degree of “idiomaticity,” depending on whether a given word has a unique or rare meaning in a given idiom. He emphasized that his new definitions would not only contribute to theory, but also facilitate practical work on machine translation:[148]
This definition [of idioms] makes it possible to construct translation dictionaries automatically. In order to do that, one has to store in a computer parallel texts in two languages and rules for establishing correspondences between elements of these texts. Guided by this definition, the machine would be able to distinguish free combinations from idioms and to create lists of the latter.
In another instance, he proposed to move away from speculations about the “metaphysical essence” of the notion of linguistic system toward the study of the property of “systematicity” and proposed to evaluate the complexity of linguistic description as a quantitative measure of the degree of systematicity.[149] Mel’chuk deliberately based his definitions on procedures that could in principle be performed automatically by computer and generally declared the possibility of computer implementation an essential condition for constructing a linguistic theory. He wrote:[150]
Language is first of all the means of communication of meaning (or content); it is a mechanism (i.e., a system of rules), which transforms a given meaning into a text and also extracts meaning from a given text. . . . The only possible way to investigate language and to understand what it is should be the construction of its functional (i.e., cybernetic) models, that is, the creation of systems that would “act” like language itself.
Machine translation research was thus becoming a means for developing theoretical linguistics; the “fundamental” and the “applied” reversed their roles. Applied linguistics now applied itself to solving theoretical problems. Mel’chuk considered the construction of automatic models of language the “central task of modern synchronic linguistics.”[151]
When criticizing the pessimistic views on machine translation often expressed in the West, Soviet structural linguists characteristically did not refer to any practical successes, but presented a theoretical argument; they maintained that, in order to prove that a particular problem cannot be solved by an algorithm, “it is necessary first of all to formulate this problem clearly in some strict terms, for example, to pose it as a mathematical problem.”[152] An attempt to formulate the task of machine translation as a rigorous mathematical problem, however, encountered difficulties in case of natural languages; a strict formulation was elaborated only for the translation from an artificial logical language into Russian.[153] Soviet structural linguists also argued that machine translation could serve as an experimental test of their theories:[154]
Machine translation and, generally, automatic analysis and synthesis of texts acquire a particular significance for linguistics: they provide experimental confirmation of linguistic statements and data. If previously in linguistics we had only experimental phonetics, now one can observe the emergence of experimental morphology, experimental syntax, and, what is particularly important and promising, experimental semantics.
Speaking of the “experiments,” however, Soviet linguists often meant no more than the theoretical possibility of constructing a computer algorithm, rather than actual computer runs; to obtain scarce computer time for testing linguistic theories was usually out of the question. Soviet machine translation was more of a thought experiment, in which the implications of an imaginary computer implementation were explored. Research on machine translation often provided models and metaphors for theoretical studies, rather than new experimental data. Ivanov, for example, conceptualized (after Warren Weaver) machine translation as a process of decoding a text (a text in English is interpreted as an encoded Russian text) and pointed to its similarity with the reconstruction of the Ursprache in traditional linguistics: “In historical comparative linguistics, the linguist acts as a decoder, who reconstructs the original message based on its several forms encoded in different ways.”[155] Ivanov maintained that the construction of an interlingua for machine translation from language A to language B was similar in method to the reconstruction of the parent language of A and B in historical comparative linguistics. In his view, this opened the possibility of a “unification of different departments of linguistics into one general theory of relations between language systems.”[156]
Research groups pursuing machine translation as a practical goal and those that viewed it as a tool for advancing theory gradually drifted apart. At a conference on information processing, machine translation, and automatic reading in Moscow in January 1961, Ivanov and Mel’chuk remarked on the “noted separation between the linguistic side of research (machine translation as applied typology of languages) and the mathematical task of selecting the optimal algorithm for a given type of language.”[157]
In Search of Meaning
By the middle 1960s, the enthusiasm for machine translation in the Soviet Union, as well as in the West, began to wane. Liapunov’s former student and his foremost hope in this field Olga Kulagina began to pull out of this area of study.[158] Liapunov expressed his disappointment to Mel’chuk, and the latter replied: “On the fate of MT [machine translation]. Not everybody is tired [of it]; I will never give up. (Perhaps, there is a fraction of your fanaticism in me too; probably, that’s why I adore you).”[159] Mel’chuk’s fanaticism was aimed, however, not so much at practical implementation of machine translation but at a new approach in theoretical linguistics, which modeled linguistic competence on the process of machine translation. This approach was embodied in the so-called “meaning—text” model, on which Mel’chuk and his colleague Alexander Zholkovsky began working in the middle 1960s.[160]
The authors initially viewed the “meaning—text” model as a prototype for a machine translation program. They suggested to express the meaning of the original text in a special “basic” semantic language. In their view, machine translation would involve three steps: “a transition from English to Basic English (independent meaning-oriented translation); a transition from Basic English to Basic Russian (translation proper); and a transition from Basic Russian to the idiomatic Russian (independent meaning-oriented synthesis).”[161] Basic languages would be built on the same syntactic structures, but the lexical component would have a national specificity: Basic Russian would include real Russian words with their “Russian” meanings, as well as some fictitious (not used in practice but grammatically correct) words which would communicate meanings ordinarily expressed by word combinations.[162] Viktor Rozentsveig—the head of the Laboratory of Machine Translation at the Institute of Foreign Languages—argued that translation via basic languages was similar to the use of a simplified literal translation (podstrochnik) by human translators dealing with vastly different types of languages.[163]
The lack of computer facilities made it impossible to implement the “meaning—text” model as a machine translation program. Instead, Mel’chuk and Zholkovsky came to view this model in a more theoretical light—as a model for linguistic competence in general.[164] They realized that the transition from text to meaning occurs not only in translation from one language to another, but also in the production and understanding of texts in the same language. They came to see formal rules of morphological, syntactic, and semantic analysis and synthesis as not merely convenient devices for solving the practical problem of machine translation but as components of a new linguistic theory.
For Mel’chuk and Zholkovsky, the potential possibility of computer implementation was a theoretical condition rather than practical requirement. It served, in Mel’chuk’s words, as a “guiding criterion” for defining their model “in a purely formal fashion, that is, with unambiguous and logically consistent formulations that do not require any additional information.”[165] “’Formal’ equals ‘scientific,’” he maintained and emphasized the necessity to strive for ever higher degree of formalization.[166] The imperative that if a linguistic theory is to be precise, it must be formulated as a computer algorithm, led to the identification of linguistic theory with a set of formal rules, and language itself with a computer-like mechanism:[167]
It seems natural to consider the central task of linguistics to be the creation of a working model of language—a logical device which, operating on a purely automatic basis, would be capable of imitating human speech activity. This device should be thought of as a system of data and rules, which comprise, so to speak, the grammar or the “handbook” of language, its “working” description, which in principle can be implemented in a computer program. . . . the speaker has a certain meaning in mind and constructs a corresponding text, while the listener receives a certain text and extracts meaning from it. Language here functions as a mechanism in the full meaning of the word, namely, as a device for the transformation, “meaning—text—meaning.”
The “meaning—text” model consists of two components: one (“semantic analysis”) translates a given text into a set of alternative “semantic representations” (which reflected possible ambiguities in the text); the other (“semantic synthesis”) translates a given semantic representation into a set of synonymous texts. Semantic representation of a given text expresses its meaning in a special “semantic language”; this language is based on a set of elementary semantic units—“atoms of meaning”—which comprises its basic alphabet. Semantic representations are build and transformed into one another by a set of formal rules and so-called “semantic axioms,” or “universal laws of reality” (e.g., “Any action entails the undesirable loss of part of the actor’s resources” or “People usually want more than they are entitled to”).[168] In Mel’chuk’s view, this semantic language could be compared with Leibnitz’s Lingua mentalis, a universal language of thought.[169]
The notion of meaning—pivotal in the “meaning-text” model—was defined by Mel’chuk and Zholkovsky operationally: meaning is “what is common in all texts that are intuitively perceived as equivalent to the original text.”[170] This idea became popular with Soviet linguists thanks to Roman Jakobson’s propagation of Shannon’s proposal to define information as “that which is invariant under all reversible encoding or translating operations,” briefly, as “the equivalence class of all such translations.”[171] Indeed, Mel’chuk maintained that meaning in his model played the role of information in Jakobson’s six-partite scheme of linguistic communication (a modification of Shannon’s scheme).[172] Later Mel’chuk described meaning as the “invariant of the set of all synonymous expressions of the language(s) in question.”[173] He took synonymy as a “basic, intuitive notion”[174] and based the notion of meaning on the concept of synonymy, not vice versa. “Here we have a perfect analogy with ‘weight,’ ‘length,’ ‘speed,’ etc.,” reasoned Mel’chuk, “the notion of ‘identical weight’ (‘identical length,’ etc.) logically precedes” the notion of ‘weight’ itself.[175]
The main strength of the “meaning—text” model lay in its ability to express the same meaning in different ways by utilizing not only lexical synonyms, but also other linguistic tools of synonymy—an idea originally developed by Zholkovsky and his colleagues in machine translation research.[176] Zholkovsky and Mel’chuk elaborated a procedure of “multiple” synthesis, which combined lexical and syntactic variations and produced a large number of synonymous variants of translation (ideally, all possible ones), and then “filtered out” those that were unacceptable. The procedures of analysis and synthesis of natural language texts, elaborated originally within the framework of machine translation, thus became components of a model of human communication. The dual meaning of the word “model” here proved crucial: what started as an attempt to build a working model (imitation) of linguistic communication, later became a tool for understanding language, then came to be seen as the only objective representation of a linguistic theory, and eventually was considered to be the theory itself.
Structural Linguistics Meets Cybernetics: Cybernetic Linguistics
Structural linguists working in machine translation groups were scattered over various teaching and engineering institutions without much support for serious theoretical or experimental research; they remained on the margins of both linguistics and computer science communities. In June 1960, Mel’chuk described the situation as follows:[177]
The research agenda of linguistics groups substantially deviates from the general profile of the institutes with which they are affiliated and these groups usually do not have the support of the directorship. These groups feel a severe shortage of skilled mathematicians, who would work as computer programmers on machine translation. It would make sense for such groups to unite if they could obtain good mathematicians.
Soviet linguistics establishment effectively turned structural linguists out of doors. Soon, however, they came back through the cybernetics window.
In April 1959, Leningrad University and the Committee on Applied Linguistics organized an international conference on mathematical linguistics with 486 participants, one third of which were mathematicians, over half were linguists, and others were communication engineers, physicists, logicians, and cyberneticians.[178] At this conference, the corresponding member of the Academy of Sciences mathematician Andrei Markov, Jr. (the son of the author of “Markoff processes”), stated that “the time has come to move from amateur studies in one’s spare time to serious goal-oriented work planned from a single center” and proposed to set up such a center to coordinate research in mathematical linguistics and machine translation.[179]
Also in April 1959, the Academy of Sciences set up the Council on Cybernetics under academician Aksel’ Berg’s chairmanship; soon Ivanov organized within the Council the Linguistics Section. In this section, various groups of structural linguists, including Ivanov’s group from ITMVT, Rozentsveig’s group from the Institute of Foreign Languages, and Sebastian Shaumian’s group from the Institute of Slavic Studies, could coordinate their activities. Linguists, mathematicians, logicians, ethnologists, and other specialists participating in Section’s meetings elaborated detailed plans for institutional development of structural linguistics, machine translation, and mathematical linguistics in the Soviet Union. Under Berg’s pressure, in February 1960, the presidium of the Academy of Sciences appointed a commission to examine the situation with structural linguistics; the prominent traditional linguist Viktor Vinogradov was appointed chairman and Berg was among the members. The Linguistics Section of the Council on Cybernetics prepared a draft, and on May 6, 1960, the presidium adopted the resolution, “On the Development of Structural and Mathematical Methods in the Study of Language.” The resolution called for the organization of sectors of structural linguistics in a number of institutes of the Academy, the founding of the journal Structural and Mathematical Linguistics, and the establishment of the Institute of Semiotics. As the result, several such groups were created: a sector of structural and applied linguistics, including a machine translation group, at the Institute of Linguistics;[180] a sector of structural linguistics at the Institute of the Russian Language;[181] a group for the study of language with mathematical methods at the Leningrad branch of the Institute of the Russian Language; the Sector of Structural Typology at the Institute of Slavic Studies[182] (Ivanov was appointed its head); and a group at the Institute of Oriental Studies.[183] Moscow University finally yielded to the pressure and, in September 1960, opened the Division of Theoretical and Applied Linguistics at the Philology Faculty (since 1962—the Division of Structural and Applied Linguistics); the Division was headed by the prominent linguist Vladimir Zvegintsev. Soon similar divisions appeared at universities in Khar’kov, Kiev, Leningrad, Novosibirsk, Riga, and Tbilisi, and at the Institute of Foreign Languages in Moscow.[184]
Structural linguistics began to fashion themselves as cyberneticians. In their contribution to the volume Cybernetics Must Serve Communism, edited by Berg, Ivanov and Shaumian stressed the connection between structural linguistics and information theory, a key component of cybernetics:[185]
The basis of cybernetics is the study of the laws of transmission and processing of information. . . . In the transmission of information, it is often necessary to convert information from one sign system into another, i.e., from one code into another. This type of conversion is called coding. Any code is a language, and coding is nothing else but translation from one language into another. Therefore, studying codes and coding is a linguistic problem, and the theory of codes and coding is a linguistic theory. . . . it is precisely the concepts of code and coding that serve as a bridge between structural linguistics and cybernetics.
Ivanov and Shaumian ironically inverted Jakobson’s view of language as a code and claimed any code to be a language and therefore an object of linguistics.
Structural linguists defined their field as an “abstract theoretical discipline studying the construction of formal models of language ”and maintained that it belonged to the “complex of sciences united by cybernetics into an integrated ensemble.”[186] The rubric of structural/cybernetic linguistics, therefore, subsumed all formal approaches, including descriptive linguistics (what is usually called “structural linguistics” in the US), transformational-generative grammar, and Jakobson’s phonology. In short, cybernetic linguistics encompassed all those trends in linguistics that did not bear the mark of approval by Soviet linguistics establishment. A study of transformational grammar could thus be justified before the authorities as a “cybernetic,” rather than purely linguistic, enterprise. For example, one proposal addressed to the Party Central Committee for organizing an international conference on transformational analysis read:[187]
At the current stage of the development of structural linguistics, to the forefront comes a new method of mathematical modeling of language, transformational analysis. It emerged at the intersection of linguistics, on one hand, and mathematical logic, semiotics (a general theory of sign systems), and cybernetics, on the other. Transformational analysis provides a necessary basis for the closest interaction of linguistics with mathematical logic, semiotics, and cybernetics. Transformational analysis has a primary theoretical importance and at the same time opens new remarkable possibilities for solving applied linguistic problems of cybernetics, such as automatic machine translation of scientific and technical literature, writing programs for automatic information machines, and a number of other applied fields in linguistics and cybernetics.
The construction of transformational logical-mathematical models of concrete languages, in particular, of the Russian language, is one of the most urgent tasks of modern structural linguistics.
The Central Committee Department of Science promptly approved this proposal.[188]
Structural linguists made efforts to translate their problematic into the cybernetic language; they emphasized, for example, that “from the current cybernetic standpoint, a human being can be seen as a device that performs operations on various sign systems and texts.”[189] The use of this language allowed them to find similarities between specific linguistic phenomena, such as aphasia, and computer processing of natural language texts:[190]
One can compare deviations from the linguistic norm that occur in machine-produced texts . . . to non-normalized human discourse (like dialects, children’s talk, or poetic speech) or to pathological deviations from the linguistic norm (aphasia). In the case of aphasia, one finds errors linked to the limitations on the capacity of short-term memory, which is typical of machines. . . . Literal interpretation of idioms, often observed in aphasia, is comparable to similar mistakes in machine translation, when each word in an idiom is related to the corresponding dictionary entry, which does not mention that this word might be part of an idiom.
In their attempts to fit structural linguistics into the agenda of cybernetics, Soviet linguists went farther than Wiener himself, who was very cautious in his assessment of the prospects of cybernetic analysis of language. In his memoirs, Ivanov recalled that during Wiener’s visit to Moscow in 1960, the latter tried unsuccessfully to convince Ivanov to share in his pessimistic outlook on the possibility of applying mathematical methods to linguistics and the humanities.[191]
The close alliance between Soviet structural linguistics and cybernetics, almost a corporate merger, did not always prove to the institutional advantage of both fields. In accordance with the May 1960 resolution of the presidium of the Academy of Sciences stipulated, preparations were under way for the organization of the Institute of Semiotics. Markov agreed to become its director; Ivanov was to be his deputy. In parallel, however, cyberneticians were making efforts to establish the Institute of Cybernetics and also wanted Markov as director. In June 1960, leading cyberneticians and structural linguists gathered to discuss the situation. Liapunov suggested to create the Institute of Semiotics and Cybernetics and to model it after the structure of the Council on Cybernetics with mathematical, linguistic, economic, biological, and computer departments. Uspenskii and Shaumian argued that semiotics was not part of cybernetics and deserved a separate institute, perhaps with the inclusion of some cybernetic departments. Markov himself, on the contrary, considered semiotics part of cybernetics—in fact, its central part—and rejected the title, “Institute of Semiotics and Cybernetics,” as “not only wrong but harmful, in the sense that it pointed to the potential possibility of partition.”[192] Markov decided to name the would-be institution “the Institute of Cybernetics,” assuming that cybernetics encompassed semiotics and structural linguistics. Structural linguists had to stake their future on cybernetics. In the end, Markov’s views on the overall agenda of cybernetic research clashed with those of Liapunov’s; Liapunov moved from Moscow to Novosibirsk; and the Institute of Cybernetics never materialized. Because of the skirmishes within the cybernetics community, structural linguists lost their chance to unite in a single institution. The short-term advantages of the close alliance with cybernetics turned into the long-term perpetuation of the marginal status of structural linguistics.
Structuralism Plus Cybernetics: The Language of Soviet Semiotics
While remaining on the margins of Soviet linguistic studies, structural linguists made forays into neighboring disciplines—not only cybernetics, but also the humanities. Work on information retrieval systems at the Laboratory of Electromodeling entailed research on logical analysis of natural language, formalized languages for various scientific fields (such as geometry and chemistry), and the construction of thesauruses, or specialized vocabularies for various fields;[193] it was in this context that the notion of semiotics was first brought in.[194] Structural linguists found “deep parallels” between structural grammar and logical syntax and semantics and, in general, between natural language and artificial languages (including scientific languages). They described “the whole field of study of natural and artificial languages (including the sociology of language and psycholinguistics) as part of semiotics, the science of sign systems and codes.”[195] The blending of structuralist (“sign systems”) and cybernetic (“codes”) terminology lay a foundation for the unique language of Soviet semiotics—a new methodology for the humanities.[196] In 1964, the Linguistics Section of the Council on Cybernetics was renamed the Semiotics Section.
To achieve an official recognition of semiotics was no easy task. In the same 51st (additional) volume of The Great Soviet Encyclopedia, in which for the first time appeared the entry on cybernetics, Uspenskii attempted to include a separate entry on semiotics. This did not work out, however, since the term “semiotics” had already been defined earlier in this encyclopedia as a study of symptoms in medicine. Nevertheless, structural linguists managed to write “their” semiotics into the entry, “Semantics in Logic”; logical semantics, they explained, was “closely connected with semiotics—a general theory of sign systems (not to be confused with semiotics in medicine).”[197]
In their new role as “semioticians,” structural linguists extended structuralist and cybernetic methods of research from the study of natural language to the exploration of artificial (logical, programming) languages, mythology, religion, folklore, literature, film, painting, and architecture—all viewed as systems of signs. Linguists Zalizniak, Ivanov, and Toporov argued that “different sign systems model the world in different ways” distinguished by their “degree of detachment” from the modeled objects and their “modeling power.”[198] They proposed a semiotic hierarchy of sign systems, in which natural language occupied an intermediate place between mathematical systems (highly abstract, or “detached,” with minimal modeling power) and religion (the least “detached,” with maximal modeling power). The three authors maintained that both language and religion functioned as “automated formal programs imposed on all members of a group” and put forward an agenda for the study of religion with the methods of structural linguistics and semiotics. They set out to develop a “structural typology” of religions similar to the structural typology of languages.
The structural-typological approach in semiotics assumed that the distinctive features that distinguished signs from one another within one system could also be used to differentiate between different sign systems (in the same way that the distinctive features that distinguished phonemes from one another comprised a general sound pattern, which distinguishes a given language from others). The three authors proposed to construct the most complete unified system of elements present in all known religions (similarly to a universal set of speech patterns in phonology) and then elicit from this system a set of distinctive features that would allow one to describe each religion by its distinctive features. Complementing their structuralist language with the cybernetic one, the authors compared the task of constructing this universal set of distinctive features with the construction of an interlingua in machine translation. Furthermore, they suggested to separate elements carrying “religious information” from the rest of religious texts by constructing an “information language” for a given religion in a similar way that information languages were then created for various scientific fields. In the end, they envisioned the “possibility of translation—represented as an algorithm—from one religious system into another.”[199]
Ivanov also suggested to “speak of literary evolution in terms of the transmission of information” and pointed to the possibility of a literary theory based on “models similar to the modern cybernetic models of biological evolution.”[200] He argued that “stochastic poetics, as much as modern biology, draws upon the role of chance in the evolution of complex structures.”[201] Ivanov compared frequent breaks with an established literary tradition to mutations; the latter produced a large pool of experimental texts, out of which, through the process of (natural?) selection, the “stabilized species” of literary genres emerged.
In December 1962, the Linguistics Section of the Council on Cybernetics and the Sector of Structural Typology of the Institute of Slavic Studies organized the Symposium on the Structural Study of Sign Systems, the first conference on semiotics in the Soviet Union. At this meeting, an attempt was made to explore the prospects of applying structuralist and cybernetic methods far beyond linguistics to a wide range of subjects—literature, painting, religion, advertising, fortune-telling, etiquette, and music—interpreted as cultural sign systems.[202]
In a representative paper, delivered at the Symposium, linguists Zholkovsky and Shcheglov introduced a new approach they called “structural poetics” (to stress the analogy with structural linguistics); they proposed to analyze an artistic (e.g., literary) work as a device, “whose operation is directed toward one goal: to bring the recipient (the reader) into a desired mental state and to cause reactions required by the author.”[203] Elsewhere Zholkovsky referred to the recently published Russian translation of William Ross Ashby’s Introduction to Cybernetics to compare an artistic work with Ashby’s concept of a cybernetic “amplifier”:[204]
An artistic work is constructed of pieces of reality as a complex multitiered amplifier, which operates in the reader’s mind. . . . The artist utilizes the automatic features of our imagination as a great reservoir of mental energy, which allows him to drive our thoughts and feelings to the pre-determined destination at the end of the road called plot. . . . One speaks of the effect of amplification in the plot when a certain series of facts or events is constructed in such a way that it, so to speak, drives itself.
Zholkovsky further suggested to view “any artistic text as a kind of machine, which acts upon the reader’s mind as a transformer—something, which at the first glance can be called a machine in the figurative sense but in reality can apparently be called such also in the serious, cybernetic sense.”[205] Zholkovsky illustrated this point by identifying “social machines” and “linguistic machines” in a literary text. Speaking of the “social machines,” he referred to the textual allusions to laws, rules, and customs that automatically launched a chain of actions in a plot. The “linguistic machines,” in his view, were “proverbs, quotations, well-known aphorisms, myths, etc., . . . that is, such linguistic situations in which after saying A, one has to say B.”[206] “Power machines” (e.g., vehicles or weapons) and “information machines” (e.g., labyrinths) often played a similar function by evoking a chain of predictable events and associations. In Zholkovsky’s analysis, all these elements functioned as literary devices (or, “devices” in the literal sense), which propelled the complex mechanism of a plot.
Keeping up with innovations in linguistics, Zholkovsky and Shcheglov soon reworked their “structural poetics” into “generative poetics.” In the same way as Noam Chomsky’s generative grammar demonstrated how complex grammatical structures were generated from kernel sentences, the authors argued, a structural description of the work of art should provide “the demonstration of [this work’s] generation from a certain theme and material according to constant rules.”[207] At the next stage, the authors transformed their “generative poetics” into “poetics of expressiveness”; this time drawing upon the “meaning—text” model, elaborated earlier by Zholkovsky together with Mel’chuk. Zholkovsky and Shcheglov developed the “theme—expressive devices—text” model of literary competence, which was “a functional model of the ideal logic underlying the correspondences between theme and text, and not an imitation of the actual history of the text’s creation by this author.”[208] The text was “derived” from its theme, or semantic kernel, with the help of “expressive devices,” or various types of transformations of the plot.[209] In the same way as the “atoms of meaning” comprised the basic alphabet of the “semantic language” in the “meaning—text” model, a set of expressive devices in the “theme—expressive devices—text” model was to constitute an “alphabet of expression.”[210] The authors illustrated their literary reconstructions with the help of complicated diagrams, which resembled, according to Uspenskii, “flow-charts of plumbing that adorn restrooms on Soviet trains.”[211]
Zholkovsky retrospectively described his model of poetics as “strongly linguistic, structuralist, and even ‘cybernetic’ in its foundations.”[212] He marvelously explicated the subconscious progressive scale of scienticity, which underlay many semiotic works: a study could be merely linguistic, “strongly linguistic,” structuralist (that is, “more strongly” linguistic), and “cybernetic” (presumably, the ultimate degree of scienticity). To Zholkovsky and Shcheglov, the term “cybernetic” evidently meant something different from what it meant for Wiener, for their models involved neither control via feedback, nor any measurements of the amount of information. For them, “cybernetic” seemed to be identical with “automatic”; “automatic,” in turn, was identified (in case of linguistics and the humanities) with “formal”; and “formal”—with “scientific.” Cybernetics—in this sense—was thus seen as a paragon of scientific perfection in the humanities.
The works of the late professor of Russian literature at Tartu University Yury Lotman provided an example of a very different—more substantial and more critical—use of the cybernetic language in Soviet semiotics.[213] For several decades, Lotman was an informal leader of the so-called “Moscow-Tartu school of semiotics.” In 1964-73, he organized a series of biennial summer schools in Kääriku (Estonia) on semiotics under the umbrella name of “schools on secondary modeling systems.”[214] These schools attracted the leading Soviet semioticians and set the agenda for semiotic research in the Soviet Union for many years. The 1966 summer school was attended by Roman Jakobson, who thereafter maintained close contacts with Soviet semioticians, particularly, with Lotman.[215]
Rejecting the attempts to reduce the essence of an artistic work to either its form of expression (e.g., in Russian formalism of the 1930s and in the works of Zholkovsky and Shcheglov) or to its political message (as in the dominant Soviet literary theory), Lotman spoke against both “the formal analysis of disparate ‘devices’ and the dissolution of the history of art in the history of political doctrines.”[216] A devotee of the “relationary mode of thinking,”[217] Lotman consistently modeled his works after structuralist phonology, exploring the relationships of artistic elements within a work, the relationship of one artistic work to another, and the relationships of various arts within a culture. In each case, he attempted to reconstruct the underlying relational system of signs—similar to the system of distinctive features in phonology—which would reduce the diversity of cultural phenomena to a limited set of basic principles and rules.
Soviet semioticians viewed culture as communication, i.e., as a constant exchange of cultural texts (or “messages” in Shannon’s and Jakobson’s terms) written in a particular language, or “code,” common for all participants of such exchanges. Art works, religious rituals, fashions, and patterns of everyday behavior were all regarded as texts; accepted symbolism in artistic works, common signs of social status, and customary norms of behavior were viewed as “codes.” Studying a culture meant trying to reconstruct its “grammar,” or underlying code, based on the bulk of available cultural texts.
In a more specific attempt to incorporate the language of information theory into semiotics, Lotman scrutinized Kolmogorov’s analyses of poetic texts with regard to their “information content.” In an uncharacteristic move for a literary theorist speaking of a mathematician, Lotman found Kolmogorov’s conclusions “subjective and vague.”[218] Assuming for a moment Kolmogorov’s position, Lotman pondered:
We want to explain the reasons for the unexpectedly high information content of a poetic text and we declare rhythm and rhyme—features pertaining to poetry only—to be a force that reduces information load; while tropes and alike—those stylistic figures that in principle can be featured in any artistic text (including prose)—we declare to be a source of increasing diversity. It turns out that the information content of a poetic text diminishes on the account of specifically poetic structural elements and grows on the account of extrapoetic ones.
In other words, Lotman challenged Kolmogorov’s assumption that the complexity of poetic form was reducing the information content of a poem and draining the reservoir of the author’s language flexibility. In Lotman’s view this assumption neglected the receptive end of poetic communication. While the author could be said to choose among many possible variants expressing a given content in a poem, the reader saw only one final version and identified the content of the poem with it. Other variants simply did not exist for the reader; another variant would be perceived as expressing different content. According to Kolmogorov, much of the flexibility of the author’s language was spent on satisfying the requirements of the form and did not contribute to information content; in Lotman’s view, on the other hand, this flexibility was incorporated (via form) into the content of the text. For the reader, Lotman wrote, “the entropy of a language’s flexibility turns into the entropy of the variety of poetic content.”[219] The author and the reader deal with different sources of diversity and as a result they evaluate the entropy of a text differently. The same text, therefore, would carry different information for them.
Formal Language and a New Ideal of Objectivity
For structural linguists and cyberneticians, formal linguistic models were associated with scientific objectivity, and imprecise terminology with “subjectivism.” In his 1962 book Models of Language, Revzin asserted that meaning in linguistics could be defined only through the formalization of this concept and warned: “Where there are no means of formalization, there is scope for any subjective structure. Only with the presence of clear formal rules for the establishment of identity of meaning is there a guarantee against subjectivism.”[220] Logic and mathematics were seen as the ideal that linguistics should strive to achieve. At a 1959 conference on mathematical linguistics, Shaumian proposed to construct linguistics as an abstract, purely deductive science, the concepts (e.g., the phoneme) and laws of which were to be postulated a priori rather than discovered empirically.[221] Revzin put forward a similar agenda:[222]
Philologists need not only inductive, but also deductive methods of research in order to obtain a system of general concepts to help them to make sense of the data obtained by the analysis of actual languages. . . . In its deductive part, Linguistics, it seems, can be constructed just as Logic or Mathematics are constructed; a certain minimal quantity of primary indefinable terms is established, and all the rest of the terms are defined by means of the primary ones. At the same time certain primary statements as to the connections between these terms (axioms) should be clearly formulated and all other statements should be proved, i.e. reduced to certain other statements.
Indeed, at one point Zholkovsky and Mel’chuk declared that “meaning should be considered an undefinable concept,” similar to “such undefinable concepts as set, point, information, energy, or elementary particle.”[223] Their ideal encompassed logic, mathematics, and the natural sciences in a single uniform image of the “exact sciences.”
Structural linguists saw their mission in bringing the light of the exact sciences into linguistics. Ivanov and Shaumian wrote in 1961:[224]
The emergence of structural linguistics signifies a revolution in the study of language, resulting in its transformation from an empirical and descriptive field of knowledge into an exact one. Through structural linguistics, the study of language enters the family of exact sciences, such as physics, chemistry, and biology.
Revzin later described his early efforts to approach language as a “natural phenomenon” and develop formal mathematical models analogous to those in “other natural sciences, for example, in physics.”[225] The reference to “other natural sciences” tacitly assumed that structural linguistics already was (or, at least, strove to become) a natural science.
Structural linguists themselves explained their striving for “exact language” by the desire to evade the fuzziness and malleability of the ordinary Soviet linguistics discourse, which had to adapt to each turn in the official policy towards linguistic issues, most notably the Marrist and later anti-Marrist campaigns. In his memoirs, Ivanov recalled: “We were tired of the phraseology of the official philosophy. We wanted to deal with precisely defined concepts and with terms that were defined through rigorously described operations.”[226] Revzin also linked the “demand for objective, exact methods” with the “distrust in all kinds of sociological phraseology and linguistic journalism.”[227] Similarly, Lotman maintained that “one of the psychological motivations that drive scholars in the humanities to the exact methods is their weariness of the . . . verbiage that is sometimes introduced under the name of science.”[228] The same attitude was expressed by mathematicians; Kolmogorov, for example, addressed the October 1956 session of Academy of Sciences with the following remark:[229]
work in computer translations yields interesting results, in particular for people designing computers and is especially interesting to linguists who, in the process of this work, are forced to specify their formulations of the laws of language. . . . Now it is impossible to use vague phrases and present them as being “laws,” something that unfortunately people working in the humanities tend to do.
Jakobson expressed similar sentiments as early as 1919, speaking at a gathering of the Moscow Linguistic Circle: “Until recently, the theory of art—in particular, the history of literature—had been not science but causerie. . . . Causerie knows no precise terminology; on the contrary, [it knows only] a variety of denominations, an equivocation, which provides a ground for puns; all this often makes [merely] a very pleasant conversation.”[230] From the very beginning, structuralists were very concerned with the imprecision of the language of linguistics, whether it was related to salon talk or officially proclaimed canons.
Another source of inspiration for Soviet structuralists was the orderliness and the predictive power of natural scientific theories; the chemical periodic table served here as an archetype. In the late 1960s, Ivanov wrote to Jakobson and his wife:[231]
I am trying to construct something like a general system of distinctive features for various ethnological descriptions. Something resembling a “periodic system” should come out of it: if certain polarities are known from the description of an ancient or primitive society, then one could make conjectures about other polarities and later prove their existence, if other sufficiently detailed descriptions allow for it. The universality of such ethnological (as well as phonological and morphological) polarities is striking; curiously, binarism here is obvious and it had been discovered by ethnologists long before linguistic binarism. Durkheim . . . and Mauss uncovered the presence of such systems as early as 1903, when linguists did not even think about such matters. Nevertheless, the essential unity of linguistics and ethnology was hardly evident prior to Roman’s [Jakobson’s] talk in Bucharest last year. So slowly move the humanities.
Soviet semioticians wanted the humanities to move faster and worked for it. In another letter, Ivanov wrote about his vision of a “synthesis of the humanities and the natural sciences . . . an essential reconfiguration of various scientific disciplines, as the result of which linguistics and semiotics will assume a prominent position.”[232]
In his 1967 article “The Study of Literature Must Be a Science!” Lotman argued that “at present, linguistics has got ‘ahead’ of the humanities, and it is in linguistics that new methods of universal scientific significance are being developed.”[233] The opponents of structuralism, in Lotman’s view, were “declaring war on Science itself, for what they are condemning is not specific for structuralism but is pertaining to scientific thinking in general.”[234]
Shaumian also maintained that structural linguistics was based on the methodological principles “imperative for any scientific theory as a logical system.”[235] He listed three such principles: homogeneity, consequentiality, and uniformity. The principle of homogeneity, in particular, dictated that “scientific explanation within a certain theory cannot be built on facts lying outside the subject of this theory.”[236] This implicitly precluded pseudo-Marxist speculations characteristic of traditional Soviet linguistic discourse. The principle of consequentiality, in Shaumian’s view, prescribed that if one accepted a premise, one had to accept all its logical implications, “no matter, are they confirmed by empirical facts or not.”[237] This principle took care of the accusations of the “detachment from reality,” often mounted at structural linguistics. Finally, the principle of uniformity required diverse subjects to be treated with the same uniform methods.[238] Zholkovsky and Shcheglov later called the same principle “systematicity” and postulated it as the “ideal of scientific poetics.” They strove to represent various artistic works “in a uniform conceptual framework,” i.e., to create a “formal meta-language” and develop formal representations of these works “composed entirely of elements of this meta-language.”[239]
On the one hand, structuralists viewed formal language as a powerful research tool and a guarantor of objectivity; on the other, their attitude toward cyberspeak was sometimes ambivalent. Some of Lotman’s references to the categories of information theory were filled with irony. Instead of testing his theories with cybernetic apparatus, he rather “tested” the cybernetic language itself and explored the limits of its applicability. In the early 1960s, he wrote, for example:[240]
The information content of an artistic text is considerably larger that that of an ordinary text. The redundancy of the artistic message [of a text] approaches zero; the redundancy is preserved [only] at the level of language. When we learn how to measure this redundancy precisely, we will obtain an objective criterion of artistic merit.
One could read this passage as a call for persistent efforts to measure the redundancy of artistic texts. On the other hand, Lotman may be implying here that since there is no way to measure this redundancy precisely, we will never have an “objective” criterion of artistic merit.
The Critique of “Formalism” as an Ideology
Semiotics gained notoriety in the eyes of the authorities right after the very first 1962 symposium on sign systems in Moscow. Unfortunately, the symposium took place shortly after Khrushchev’s fateful visit to the Manezh art gallery in Moscow on December 1, 1962. The visit was arranged by his conservative entourage and provoked Khrushchev’s predictable violent reaction against nonrepresentational artists.[241] Conservatives used the occasion to launch a sweeping onslaught on liberalization in culture under the banner of the campaign against “abstractionism” and “formalism” in art. Among other “formalist” trends, they censured semiotics. In a draft of his speech before the Academy of Sciences, Secretary of the Party Central Committee, chairman of the Ideological Commission Leonid Ilichev criticized the semiotics symposium for the attempts to “emasculate the ideational content of art . . . and reduce everything to the formal methods of analysis.”[242] Another critic mused: “Can a work of art or literature be regarded as a system of signs? No, it cannot; for that would lead to abstractionism.”[243] “Looking at us, the authorities smelled the enemy,” recalled Ivanov, one of the symposium organizers, “It gradually became impossible to publish [our works] in Moscow, at least on semiotic topics.”[244]
The charges of “formalism” and “abstractionism” were brought against both structural linguistics and semiotics. One critic charged that structural linguistics relied on a “subjective idealist world view” and accused structural linguists of the “detachment from real linguistic facts and transition into the realm of pure abstractionism.”[245] Fedot Filin—a leading spokesman for the linguistics establishment—condemned the attempts to represent language as a “closed system of signs, a cipher, a cybernetic device” and “make an absolute out of formal methods.”[246] “All natural languages develop not by rigid logical-mathematical programs, but spontaneously,”[247] he argued, citing the ideologically validated thesis that the languages of small ethnic groups in the Soviet Union tended to wither away and be replaced by the Russian language.[248] “The application of mathematical and cybernetic methods can play only a subsidiary role in the study of social functions of language,” he concluded.[249] The formalisms of structural linguistics, indeed, bracketed pseudo-Marxism rhetoric, and the critics perceived this neglect of the dogma as a direct challenge and labeled it accordingly.
Traditional philologists accused structural linguists of “textocentrism,” which one critic described as the “part of modern formal linguistics . . . that limits the notion of context to those relations that can be found in a printed text with purely formal means. Thus, no attention is paid to the extralinguistic context.”[250] Here the term “formal” was used in a technical sense, without ideological connotations; other critics, however, made formalism an ideologically loaded concept.
The label of “formalism” had long been used in Soviet discourse to denote political errors in scholarly and artistic work; the campaigns against “formal genetics” and “formalism in music” could serve as vivid illustrations.[251] As in other fields, the anti-formalist argument was often advanced from within the scientific community rather than merely imposed from above by Party ideologues. The prominent linguist Vasilii Abaev—a staunch opponent of structuralism, who had himself suffered earlier from various ideological campaigns—chose the label of “formalism” as a weapon in his polemics with structural linguists. He granted formalism the status of ideology and thus portrayed structural linguistics as an ideological sinner:[252]
Formalism becomes unacceptable when it functions as an ideology, that is, when it attempts to substitute the form of phenomena for their essence . . . Hereafter we will speak of formalism precisely in this sense: not as a method or tool in science or art, but as an ideology. . . . the essence of structuralism is . . . in dehumanizing linguistics through its ultimate formalization.
Abaev alleged that the “dehumanization of linguistics is just one step in the general dehumanization of culture,” by which he referred to the process of “degeneration,” represented by abstract art and dodecaphonic music.[253]
Abaev saw the root of formalist perversions in mixing mathematics with the humanities. “The coupling of pseudo-linguistics and pseudo-mathematics is the essence of mathematical linguistics,” he wrote;[254] he probably thought that “true” mathematics would not touch this subject. “While in the physical sciences mathematics is the way to get closer to reality,” he argued, “in the humanities, mathematics becomes a way to avoid reality.”[255] What was most important in the context of structural linguists’ efforts to legitimize their work through machine translation, Abaev insisted that the significance of machine translation was limited to purely technical problems and denied any value of such works for linguistics in general. “There is no need to rebuild the whole edifice of linguistics to accommodate it to the needs of machine translation,” he argued.[256]
With all his penchant for political labels, Abaev very perceptively identified the source of strength of structural linguistics and realized the general threat it posed:[257]
the more formalized linguistics would be, the less of it would belong to the humanities. . . . [There is] a common tendency to push the humanities to the background with respect to the physical and engineering disciplines or . . . to make the humanities imitate “exact” sciences and thereby create an illusion of a “synthesis” of sciences, while in fact draining blood from the humanities.
There could be no doubt what “synthesis” Abaev referred to: in the same issue of Problems of Linguistics, right after his article, Isaak Revzin wrote: “It is commonly known how important a role in this synthesis is played by cybernetics, the science of general laws of control by means of processing and ordering information. Linguistics enters the group of sciences united by cybernetics owing to the development of structural linguistics.”[258]
In March 1957, the Institute of Linguistics in Moscow held a conference on the relations between the synchronic analysis and the historical study of language with some 400 participants.[259] Speaking at this meeting, Abaev compared the structuralist to a person who tried to jump on a moving train while holding a heavy suitcase. Structural linguist Vladimir Toporov sharply replied that Abaev would probably never catch this train, for he was held back by the heavy load of worn-out concepts, false truths, and ballooned authorities. Structural linguistics posed a challenge not only in the field of theory but also—perhaps more importantly—in the very style of its discourse, openly critical and unbowed before the academic and administrative authority. Recalling this debate, most unusual in its boldness, Revzin wrote: “When all this was happening before my eyes, which were accustomed to the ‘discussions’ of the Stalin period, I could not think of scholarly arguments; exultation and ire were rising in my soul.”[260]
Playing the Cybernetic Game
In 1968 Mel’chuk completed a manuscript of his book on the “meaning—text” model; six years had passed, however, before the first part came out in 1974; the second part was never published in Soviet years. According to one of his former colleagues, Mel’chuk never bothered to formulate his ideas “according to the rules of the game accepted by the greater part of the linguistic community.”[261] Mel’chuk ignored the official hierarchy and deliberately dropped out of the rat race for titles and degrees. His independent style, which reflected the marginality of his research program within Soviet linguistics and his opposition to the scientific establishment, eventually led to his conflict with the political authorities. In July 1974, Mel’chuk wrote to Jakobson that his life in the Soviet Union became “totally unbearable” and asked him to find a position in one of American universities.[262] On January 26, 1976, The New York Times published Mel’chuk’s open letter protesting against Soviet persecution of the prominent dissident scientists Andrei Sakharov and Sergei Kovalev. This was the last straw. In March, despite Mel’chuk’s highest reputation as a scholar and the author of numerous first-rate publications, his superiors at the Institute of Linguistics gave him negative evaluation and soon dismissed him from his job. In a new open letter to his Western colleagues, he wrote:[263]
For several years I have had practically no possibility to publish my papers in leading Soviet linguistic journals. . . . The second volume of my monograph Toward a Theory of Linguistic Models of the Meaning—Text Type (the product of many years of painstaking work) waited for publication more than 8 years and failed to be, after all, published. I was and am forbidden to teach, to take part in many scientific conventions, to go abroad for contracts with Western colleagues. Immediately after the appearance of my letter in The New York Times Soviet linguistic periodicals and publishing houses began suppressing references to my works, acknowledgements by other authors mentioning my name ... and even my name as the editor or translator.
Under such conditions and having no permanent job I am left without the least possibility for normal continuation of my linguistic research.
Soon the police informed Mel’chuk, whom for obvious reasons nobody wanted to hire, that, as a “parasite” and a “sponger,” he was to be deported from Moscow and possibly imprisoned.[264] References to his works disappeared from print; the circulation of his ideas was thus severely limited. In his memoirs, Uspenskii stated that Mel’chuk’s dismissal from the Institute of Linguistics marked the end of the “silver age” of Soviet structural linguistics.[265] In May 1977, Mel’chuk emigrated from the Soviet Union and accepted a position at the University of Montréal.
As a result of a general crackdown on independently minded intelligentsia in the late 1960s and early 1970s, Mel’chuk’s wife linguist Lidiia Iordanskaia and a number of his colleagues—Iurii Apresian, Alexander Zholkovsky and others—also lost their jobs. Some of them emigrated, some repeated the earlier move of structural linguists into the field of machine translation and found positions at the “Informelektro” Institute of the Ministry of Electrical Engineering Industry, where they started working again on a machine translation program.[266]
Theoretically inclined structural linguists perceived the alliance with cybernetics and the emphasis on the usefulness of their theories for machine translation as an unavoidable compromise. In this sense, the situation in the 1970s resembled the late 1950s, when, under the cover of machine translation, structural linguists could work on theoretical problems. Ivanov’s recollections of those early days captured well the feelings of Soviet “cybernetic linguists”:[267]
I cannot say that we intentionally deceived anyone, but it is now impossible to overlook the fact that in those past discussions the practical utility of new methods was if not strongly exaggerated then at least strongly emphasized. Society oriented itself toward pragmatic tasks, and the authorities were willing to allow anything that advanced those tasks. Everybody knew the rules of the game. And we yielded to [the spirit of] the time.
For Mel’chuk, the necessity to deal with the rules of the Soviet game ended when he crossed the Soviet border as an émigré. The first thing he did when he arrived in Vienna, however, was to write, in his words, “a showy and shallow article, ‘Linguistics and Cybernetics,’—partly for money.”[268] The cybernetic game had to be played on this side of the border too.
Formal Language: Limitations and Freedom
Soviet structuralists borrowed cybernetic language but often skipped cybernetic theory; their inspiration, perhaps, was not so much mathematics but mathematicians themselves. Structural linguists and semioticians did not live in complete hermetic isolation; on the contrary, they directly faced a completely different professional world with very distinct social standards of behavior. They discovered that mathematicians’ discursive style—the meticulous observation of form in conducting seminars, leading discussions, writing papers, and so on—was very different from their own. Linguist Isaak Revzin, for example, recalled how mathematician Uspenskii led his seminars on mathematical linguistics at Moscow University:[269]
The pedantic and formal manner that was characteristic of Uspenskii’s every statement and every action was not at all boring; . . . it was jovial and jocular and was accompanied by exceptional indulgences: during the seminar, it was permitted to interrupt the speaker, ask questions, go to the blackboard, and start an argument; it was even permitted not to understand the speaker and be proud of this. All this was unheard-of among linguists; it looked especially striking at the Philology Faculty [of Moscow University] and was totally unthinkable at our Institute of Foreign Languages.
Speaking at the Moscow Institute of Foreign Languages in September 1958, Jakobson particularly impressed his listeners, when he “did not merely present a monologue, but also had a dialogue with the audience.”[270] Accustomed to the monologue-type discourse of the Stalin era, structural linguists perceived this new style—which they viewed as characteristic of the mathematicians, natural scientists, and foreign scholars—as a liberation.
In his talk, among other things, Jakobson furnished a quotation from Catherine the Great, “Freedom is the right to do whatever is permitted by law.”[271] Mathematicians’ formal style and Jakobson’s reflections of Russian history juxtaposed in Soviet semiotic discourse in a remarkable way. Structural linguists realized that formal norms had a dual nature—both restrictive and liberating—and in this sense were similar to social norms.
A reflection on social norms is implicit, for example, in Isaak Revzin’s treatment of the linguistic norm. In his view, any norm consists of two components: a set of permissions and a set of prohibitions. Two attitudes toward the norm are then possible: (1) all that is not explicitly permitted (e.g., specified in a normative dictionary or grammar), is prohibited; (2) all that is not explicitly prohibited, is permitted. The first attitude, Revzin argues, is characteristic of normative stylistics, and this role is assumed by a “group of writers, who become a kind of ‘priests’ or guardians of linguistic norm, and their books are universally recognized as collections of correctly constructed, exemplary sentences.”[272] One can easily recognize in this description not only the “guardians” of linguistic orthodoxy, but also the notorious “classics of Marxism-Leninism,” whose exemplary quotations were employed to set norms in every academic field. The second attitude, Revzin writes, is “often taken in poetry; although poets frequently violate even explicit prohibitions.”[273]
The perception of poetry as an archetypal realm of freedom, where “all that is not explicitly prohibited, is permitted,” has permeated semiotic studies. For both Lotman and Revzin, poetry opens up room for freedom precisely because of the rigidity of constraints imposed on poetic form. In overcoming the limitations of form, poets create their own worlds, in which certain “expressions, which had no meaning in [ordinary] language before, now become justified.”[274] Revzin takes as an example Chomsky’s famous sentence, “Colorless green ideas sleep furiously,” which is syntactically correct but semantically meaningless, and attaches meaning to it by turning it into an elegant poetic line:[275]
The idea furiously sleeps;
It tosses and turns in bed,
It pounds, it screams, it weeps,
And whispers in my head.
In a similar way, Revzin argues, are justified the “unusual statements that formulate new truths in science.”[276] In a similar way, it seems, structural linguistics and semioticians have “justified” their own “new truths”; they employed formal discursive rules to free themselves from dogmas rather than to make their reasoning “exact.” In opposition to the traditional Soviet linguistics and literary theory with their imprecise language and fixed content emerged structural linguistics and semiotics with “precise” language and very flexible content.
On the one hand, formal language imposes severe constraints on discourse: not everything can be expressed in a given formalism, and many things have to be left out. On the other hand, formal language may paradoxically prove capable of describing things that cannot be talked about in ordinary language because of various social and political taboos. If one glanced at a typical volume of semiotic works, one would find a very unusual set of subjects, many of which were seldom or never discussed in traditional Soviet literary and cultural studies. “Small” topics like fortune-telling, card games, rules of etiquette, detective stories, oriental myths, Buddhist cosmology, limericks, Kamasutra, and the oeuvres of such officially censured poets as Akhmatova, Pasternak, and Mandelshtam stood in sharp contrast to such traditional “big” subjects of study as the language of Lenin, the literature of socialist realism in the fight against bourgeois decadence, and the image of a young progressive worker in the writings of members of the Union of Soviet Writers.
The dogmatic class-based interpretation of the social functions of language in the dominant Soviet discourse left researchers little room for originality. If one wanted to place a literary work in a social and political context, one had to follow primitive sociological schemes in the spirit of Soviet pseudo-Marxism. The only safe way not to give a pseudo-Marxist interpretation of the content of an artistic work was not to give any interpretation of its content at all and focus on its form. The only safe way to avoid political clichés was to ignore political context and focus on the literary context.
Soviet semioticians developed a unique analytical apparatus for the analysis of content through the form. They viewed structural relations between elements of an art work and between different art works as crucial components of the content of art. The language of semiotics hardly “masked” any hidden political message, it rather described (and by this description, created) the image of a cultural universe free of political messages.
By adopting the language of structuralism and cybernetics, Soviet semiotics effectively bracketed a whole set of questions characteristic of traditional Soviet linguistics and cultural theory. As with the Ockham razor, they cut off the “unnecessary” concepts. A simplistic model of communication as information exchange, canonized by semioticians, supplanted yet more simplistic models of culture as an arena of class struggle advanced by Soviet establishment.
Zholkovsky keenly observed that semiotics “mirrored” some of the features of the dominant discourse; this “mirror,” however, was a distorted one. Semiotic discourse may be seen as a carnival-type reflection of the dominant Soviet discourse: a parody on the scientistic vocabulary of Soviet pseudo-Marxism, a mocking imitation of Soviet obsession of technology, and an ironic inversion of the role of formalism in mathematics—from the tool of logical rigor to the instrument of freedom. For example, Zholkovsky and Mel’chuk’s declaration that meaning was an undefinable concept, like “such undefinable concepts as set, point, information, energy, or elementary particle,” hardly looked like an attempt to imitate the rigor of the exact sciences. Instead, one could interpret their argument as an ironic overturn of common stereotypes. Their implicit message could go like this: “We want to make our theory as rigorous as mathematics, physics, or information theory. Since these ‘exact’ sciences see no need to define their basic concepts, neither do we.” Interestingly, the notion of carnival attracted the attention of Soviet semioticians, who further developed the conception of carnival as an inversion of binary oppositions, first offered by the Russian thinker Mikhail Bakhtin in his study of Rabelais.[277] Characteristically, they failed to apply this type of analysis to their own discourse.
The “authoritarian” (Zholkovsky) and “totalizing” (Gasparov) tendencies in the semiotic movement could also be seen as half-imitation of, half-parody on the dominant discourse. Soviet structuralists’ portrayal of universal culture as a deterministic mechanism governed by strict rules without much room for freedom could serve as an implicit comment on Soviet culture and their own place in it. The turmoil in the semiotic movement in recent years could thus be interpreted as a crisis of parodic discourse when the object of parody is gone. Most recent attempts by Lotman and his colleagues to incorporate the accidental and the individual into “totalizing” semiotic models, perhaps, reflected the choices and uncertainties that semioticians themselves faced in the rapidly changing post-Soviet environment.
Despite the limited capacity of cybernetic language as a research tool for structuralists, its power was enhanced by its significant social role. Structural linguists and semioticians developed a professional language that reflected their unique position in the academic community—at the crossroads (and on the margins) of linguistics, cybernetics, the natural sciences, and the humanities. Claiming mathematics and natural science as their ideal, structuralists imitated not so much the rigor of the “exact sciences” as their discursive style, their critical mode of speaking, and their appeal to objectivity. This imitation was closely intertwined with a parodic representation of Soviet official discourse with its scientistic vocabulary and ceremonious style. When speaking cybernetic language, structuralists fashioned their singular identity as naturalists studying culture.
-----------------------
[1] Adapted from Aleksandr Zholkovskii, “O Mel’chuke,” in Idem, Inventsii (Moscow: Gendal’f, 1995), p. 170.
[2] Shukman, “Soviet Semiotics,” p. 195.
[3] Ibid., p. 191.
[4] Ibid., pp. 195-96.
[5] Viach.Vs. Ivanov, “Goluboi zver’ (Vospominaniia),” Zvezda, nos. 1-3 (1995); Idem, “Iz proshlogo semiotiki, strukturnoi lingvistiki i poetiki,” in D.A. Pospelov and Ia.I. Fet, eds. and comps., Ocherki istorii informatiki v Rossii (Novosibirsk: OIGGM SO RAN, 1998), pp. 310-40; A.D. Koshelev, comp., Iu.M. Lotman and tartusko-moskovskaia semioticheskaia shkola (Moscow: Gnozis, 1994); S.Iu. Nekliudov, comp. and ed., Moskovsko-tartuskaia semioticheskaia shkola: istoriia, vospominaniia, razmyshleniia (Moscow: Iazyki russkoi kultury, 1998); E.V. Permiakov, ed. and comp., Lotmanovskii sbornik, vol. 1 (Moscow: ITs-Garant, 1995); “Razgovory: Evgenii Breido beseduet s Aleksandrom Zholkovskim,” ; V.A. Uspenskii, “Serebrianyi vek strukturnoi, prikladnoi i matematicheskoi lingvistiki v SSSR i V.Iu. Rozentsveig: Kak eto nachinalos’ (zametki ochevidtsa),” Wiener Slawistischer Almanach, Sbd. 33: Festschrift für V.Ju. Rozencvejg (1992): 119-62; Idem, “Predvarenie dlia chitatelei ‘Novogo literaturnogo obozreniia’ k semioticheskim poslaniiam A.N. Kolmogorova,” Novoe literaturnoe obozrenie 24 (1997): 122-215; A.K. Zholkovskii, “Iz istorii vcherashnego dnia,” Rossiia/Russia, no. 1 (1998): 135-52; A. Zholkovsky, “Zh/Z: Notes of an Ex-Pre-Post-Structuralist,” Wiener Slawistischer Almanach, Sbd. 33: Festschrift für V.Ju. Rozencvejg (1992): 283-91.
[6] Uspenskii, “Serebrianyi vek,” p. 119.
[7] Zholkovsky, “Zh/Z,” p. 285.
[8] Ibidem.
[9] B.A. Gasparov, “Tartuskaia shkola 1960-kh godov kak semioticheskii fenomen,” in Koshelev, comp., Iu.M. Lotman, p. 289.
[10] Ibid., pp. 292-93.
[11] G.I. Revzin, Review of Kul’tura i vzryv, by Iu.M. Lotman, Voprosy isskustvoznaniia, no. 1 (1993): 212-16.
[12] Ibid., pp. 284-85.
[13] The literature on Jakobson is too extensive to list here. For an excellent overview of the evolution of Jakobson’s ideas, see Linda R. Waugh and Monique Monville-Burston, “Introduction: The Life, Work, and Influence of Roman Jakobson,” in Roman Jakobson, On Language, edited by Linda R. Waugh and Monique Monville-Burston (Cambridge, Mass.: Harvard University Press, 1990), pp. 1-45; see also a bibliography in the same volume.
[14] See Roman Jakobson, “An Example of Migratory Terms and Institutional Models (On the Fiftieth Anniversary of the Moscow Linguistic Circle),” in Idem, Selected Writings, vol. II: Word and Language (The Hague and Paris: Mouton, 1971), pp. 527-38.
[15] See Jakobson, “An Example”; Idem, “Efforts Toward a Means-Ends Model of Language in Interwar Continental Europe,” in Idem, Selected Writings, vol. II, pp. 522-26; Josef Vachek, The Linguistic School of Prague (Bloomington: Indiana University Press, 1966).
[16] “I am a linguist, and nothing linguistic is alien to me”; see Roman Jakobson, “Results of a Joint Conference of Anthropologists and Linguists,” [1953] in Idem, Selected Writings, vol. II, p. 555.
[17] Quoted in Roman Jakobson, “Retrospect,” in Idem, Selected Writings, vol. II, p. 711.
[18] The theory of distinctive features was formulated in its most complete form by Jakobson, C. Gunnar Fant, and Morris Halle in the early 1950s; see Roman Jakobson, C. Gunnar Fant, and Morris Halle, Preliminaries to Speech Analysis (Cambridge Mass.: MIT Press, 1952).
[19] Jakobson to Wiener, January 19, 1949; Jakobson papers, MIT MC 72, box 47.6. This letter can also be found in Wiener Papers, MIT MC22, box 6.92.
[20] Ibidem.
[21] Wiener to Jakobson, March 2, 1949; Jakobson papers, MIT MC 72, box 47.6.
[22] Wiener to Jakobson, January 7, 1957; Jakobson papers, MIT MC 72, box 47.6.
[23] Norbert Wiener, The Human Use of Human Beings: Cybernetics and Society [1950, 1954] (New York: Avon Books, 1967), p. 255.
[24] Roman Jakobson, “Linguistics in Relation to Other Sciences,” [1974] in Idem, On Language, p. 481.
[25] Roman Jakobson, “Linguistics and Communication Theory,” [1960] in Idem, On Language, p. 494.
[26] Jakobson to Shannon, April 24, 1951; Jakobson papers, MIT MC 72, box 45.21. Shannon’s article in question most likely was C.E. Shannon, “Prediction and Entropy of Printed English,” Bell System Technical Journal, 30:1 (1951): 50-64.
[27] See E. Colin Cherry, Morris Halle, and Roman Jakobson, “Toward the Logical Description of Languages in Their Phonemic Aspect,” Language 29:1 (1953).
[28] Roman Jakobson, “Linguistics and Communication Theory,” p. 490.
[29] Ibid., p. 496.
[30] Jakobson, “Retrospect,” in Idem, Selected Writings, vol. II, p. 718.
[31] Claude E. Shannon, and Warren Weaver, The Mathematical Theory of Communication (Urbana, IL: The University of Illinois Press, 1949), pp. 33-34.
[32] Ibid., pp. 8-9 (emphasis added).
[33] Adapted from Roman Jakobson, “Linguistics and Poetics,” [1960] in Idem, Language in Literature, edited by Krystyna Pomorska and Stephen Rudy (Cambridge, Mass.: The Belknap Press of Harvard University Press, 1987), pp. 66-71.
[34] Jakobson, “Linguistics and Communication Theory,” p. 497.
[35] Jakobson papers, MIT MC 72, box 3.63-65.
[36] Ibid., box 33.9-25.
[37] Ibid., box 5.79.
[38] Ibid., box 5.81.
[39] Henshaw to Jakobson, January 25, 1965; Jakobson papers, MIT MC 72, box 5.74. Jakobson left the Society in 1974.
[40] Roman Jakobson, “Linguistics in Relation to Other Sciences,” p. 454.
[41] Ibid., p. 459.
[42] Ibid., p. 457.
[43] Ibid., p. 460.
[44] Transcript of the third lecture at the Tokyo Institute for Advanced Studies of Language, Japan, July 28, 1967; Jakobson papers, MIT MC 72, box 34.84.
[45] Transcript of the fifth lecture at the Tokyo Institute for Advanced Studies of Language, Japan, July 31 (?), 1967; Jakobson papers, MIT MC 72, box 34.86.
[46] Vsesoiuznaia Akademiia Sel’sko-Khoziaistvennykh Nauk Imeni V.I. Lenina.
[47] Nikolai Krementsov, Stalinist Science (Princeton, N.J.: Princeton University Press, 1997), p. 214. See I.I. Meshchaninov, “O polozhenii v lingvisticheskoi nauke,” Izvestiia Akademii nauk SSSR. Otdelenie literatury i iazyka, no. 6 (1948): 4-16; “O polozhenii v lingvisticheskoi nauke,” Vestnik Akademii nauk SSSR, no. 12 (1948): 71-74.
[48] On the history of Marrism and the controversies over this doctrine in Soviet linguistics, see V.M. Alpatov, Istoriia odnogo mifa: Marr i marrizm (Moscow: Nauka, 1991); Idem, “Marr, Marrism, and Stalinism,” Russian Studies in History 34:2 (Fall 1995): 37-61; M.V. Gorbanevskii, V nachale bylo slovo: maloizvestnye stranitsy istorii sovetskoi lingvistiki (Moscow: Universitet druzhby narodov, 1991); and Yuri Slezkine, “N.Ia. Marr and the National Origins of Soviet Ethnogenetics,” Slavic Review 55:4 (Winter 1996): 826-62.
[49] F.P. Filin, “O dvukh napravleniiakh v lingvistike,” Izvestiia Akademii nauk SSSR. Otdelenie literatury i iazyka, no. 6 (1948): 488.
[50] On the anti-cosmopolitanism campaign, see Gennadi Kostyrchenko, Out of the Red Shadows: Anti-Semitism in Stalin’s Russia (Amherst, NY: Prometheus Books, 1995).
[51] See, for example, vestiges of such criticism in a 1952 publication: “Structuralism neglects the specificity of national languages; it strives for creating a general grammar; it seeks cosmopolitan features valid for ‘all humanity,’” (“Zadachi sovetskogo iazykoznaniia v svete trudov I.V. Stalina i zhurnal ‘Voprosy iazykoznaniia,’” Voprosy iazykoznaniia, no. 1 [1952]: 7).
[52] A.V. Desnitskaia, “Protiv formalizma v uchenii o zvukakh rechi,” Izvestiia Akademii nauk SSSR. Otdelenie literatury i iazyka, vol. VIII, no. 4 (July-August 1949): 337.
[53] Ibid., p. 342.
[54] Ibid., pp. 338-39.
[55] Ibid., p. 342.
[56] Other obvious examples of this strategy are his 1930 article, “Dizzy with Success,” which condemned the “excesses” of forced collectivization, and the 1935 formula, “the son is not responsible for his father,” announced after the completion of large-scale systematic deportations of so-called “kulaks” (well-to-do peasants) with their entire families.
[57] There were speculations that Stalin might have confused one of the Russian dialects with the Kursko-Orlovskaia military operation of WWII. Some former Marrists, for example, Fedot Filin, immediately reeducated themselves and wrote abundantly on the “Kursko-Orlovskii dialect.”
[58] See the Russian original in Iuz Aleshkovskii, “Ne unyvai, zimoi dadut svidan’e…,” Novyi mir, no. 12 (1988): 121.
[59] L.S. Barkhudarov and G.V. Kolshanskii, Review of the American journal Language (1952-1953), Voprosy iazykoznaniia, no. 5 (1954): 131.
[60] Quoted in Jakobson, “An Example,” p. 535.
[61] O. Leshka, “K voprosu o strukturalizme (Dve kontseptsii grammatiki v Prazhskom lingvisticheskom kruzhke),” Voprosy iazykoznaniia, no. 5 (1953): 90.
[62] Chairman of the Polish American Congress Sigmund J. Sluszka, as quoted in an anonymous 1950 newspaper article, “Reds Admit Role in Columbia Deal”; Jakobson papers, MIT MC 72, box 38.56.
[63] Kirillin to the Central Committee, April 28, 1956; the Center for Preservation of Contemporary Documentation (Tsentr Khraneniia Sovremennoi Dokumentatsii [TsKhSD]), f. 5, op. 35, d. 22, ll. 86-87.
[64] Nevozvrashchenets is a person who stayed abroad despite strict orders to return. According to a 1929 law, such actions were officially declared a crime deserving capital punishment.—S.G.
[65] Ivanov, “Goluboi zver’,” no. 3, p. 159. The article in question was A.Ia. Khinchin, “Ob osnovnykh teoremakh teorii informatsii,” Uspekhi matematicheskikh nauk 11:1 (1956): 17-75.
[66] Ivanov to Jakobson, October 22, 1956; Jakobson papers, MIT MC 72, box 42.30.
[67] Jakobson to Ivanov, July 30, 1957; Jakobson papers, MIT MC 72, box 42.30.
[68] Uspenskii, “Serebrianyi vek,” p. 127.
[69] I.I. Revzin and V.Iu. Rozentsveig, Osnovy obshchego i mashinnogo perevoda (Moscow: Vysshaia shkola, 1964), pp. 46-54.
[70] For an overview of the impact of Jakobson’s work on Soviet linguistics, see T.V. Gamkrelidze, T.Ia. Elizarenkova, and Viach.Vs. Ivanov, “Lingvisticheskaia teoriia R.O. Iakobsona v rabotakh sovetskikh lingvistov,” in Daniel Armstrong and C. H. van Schooneveld, eds., Roman Jakobson: Echoes of His Scholarship (Lisse: Peter de Ridder Press, 1977), pp. 91-121.
[71] Ibid., p. 95.
[72] On this failed attempt to publish Jakobson’s volume, see I.A. Mel’cuk, “Roman Jakobson and Studies of the Russian Language,” in Wiener Slawistischer Almanach, Sbd. 39: I.A. Mel’cuk, The Russian Language in the Meaning-Text Perspective (1995): 607-608.
[73] See Roman Jakobson, Izbrannye raboty, comp. and ed. V.A. Zveginitsev (Moscow: Progress, 1985).
[74] Uspenskii, “Serebrianyi vek,” p. 123.
[75] This reference pertains to Hans Reichenbach’s works on symbolic logic; see Hans Reichenbach, Elements of Symbolic Logic (New York: Macmillan Co., 1947)—S.G.
[76] Ibid., p. 121.
[77] Gamkrelidze, Elizarenkova, and Ivanov, “Lingvisticheskaia teoriia,” p. 92.
[78] Uspenskii, “Predvarenie,” p. 180.
[79] Ibid., p. 130.
[80] Uspenskii, “Serebrianyi vek,” p. 124.
[81] D.M. Segal, “Mezhvuzovskaia nauchnaia konferentsiia po prikladnoi lingvistike,” Mashinnyi perevod i prikladnaia lingvistika, no. 5 (1961): 96.
[82] For a more detailed discussion of Kolmogorov’s definition, see I.I. Revzin, Models of Language, trans. N.F.C. Owen and A.S.C. Ross (London: Methuen, 1966), p. 66; Uspenskii, “Predvarenie,” pp. 180-81.
[83] Roman Jakobson, “Appendix: Case Revisited,” [1958] in Idem, On Language, p. 384.
[84] I.I. Revzin, Sovremennaia strukturnaia lingvistika: Problemy i metody, ed. Viach.Vs. Ivanov (Moscow: Nauka, 1977), pp. 140-44; see also A.A. Zalizniak, Russkoe imennoe slovoizmenenie (Moscow: Nauka, 1967).
[85] Uspenskii, “Predvarenie,” p. 136. Ivan Pavlov’s physiological studies of conditional reflexes, which involved salivation, were then considered an exemplar for the study of “higher nervous activity.”
[86] See a comprehensive bibiliography of Kolmogorov’s works on this subject in A.N. Shiriaev, ed., Kolmogorov v vospominaniiakh (Moscow: Nauka, 1993), p. 698.
[87] A.V. Prokhorov, “O sluchainoi versifikatsii (K voprosu o teoreticheskikh i rechevykh modeliakh stikhotvornoi rechi),” in V.E. Kholshevnikov, ed., Problemy teorii stikha (Leningrad: Nauka, 1984), p. 95, table 8.
[88] Uspenskii, “Predvarenie,” p. 134.
[89] M.L. Gasparov, Sovremennyi russkii stikh. Metrika i ritmika (Moscow: Nauka, 1974), pp. 13-14.
[90] See A.S. Monin, “Dorogi v Komarovku,” in Shiriaev, ed., Kolmogorov v vospominaniiakh, p. 482; I.I. Revzin, “Soveshchanie v g. Gor’kom, posviashchennoe primeneniiu matematicheskikh metodov k izucheniiu iazyka khudozhestvennoi literatury,” in T.N. Moloshnaia, ed., Strukturno-tipologicheskie issledovaniia (Moscow: AN SSSR, 1962), p. 287.
[91] Revzin, “Soveshchanie v g. Gor’kom,” p. 287 (emphasis original).
[92] Monin, “Dorogi v Komarovku,” p. 482.
[93] S.P. Bobrov, “Opyt izucheniia vol’nogo stikha pushkinskikh ‘Pesen zapadnykh slavian,’” Teoriia veroiatnostei i ee primeneniia 9:2 (1964): 262-72.
[94] Ivanov, “Goluboi zver’,” no. 3, p. 169. Ivanov has preserved a number of Kolmogorov’s unpublished manuscripts on this subject.
[95] Uspenskii, “Predvarenie,” p. 143.
[96] Shannon and Weaver, The Mathematical Theory, p. 45.
[97] A.A. Markov, Example of a statistical investigation of the text of “Eugene Onegin” illustrating the dependence between samples in chain, [1913] trans. Morris Halle, 1955.
[98] Shannon, “Prediction and Entropy.”
[99] A.A. Piotrovskaia et al., “Entropiia russkogo iazyka,” Voprosy iazykoznaniia, no. 6 (1962): 124.
[100] See description and discussion of Kolmogorov’s experiments in A.M. Iaglom and I.M. Iaglom, Veroiatnost’ i informatsiia (Moscow: Nauka, 1973), pp. 257-60; Uspenskii, “Predvarenie,” pp. 184-85.
[101] A.S. Monin, “Dorogi v Komarovku,” p. 484.
[102] Ivanov, “Goluboi zver’,” no. 3, p. 168.
[103] R.L. Dobrushin, “Matematicheskie metody v lingvistike,” Matematicheskoe prosveshchenie, no. 6 (1961): 45-46.
[104] See Revzin, “Soveshchanie v g. Gor’kom,” p. 289; Walter Rewar, “Tartu Semiotics,” Bulletin of Literary Semiotics, no. 3 (May 1976): 3-4; A.N. Shiriaev, “Andrei Nikolaevich Kolmogorov: Biographicheskii ocherk o zhizni i tvorcheskom puti,” in Idem, ed. Kolmogorov v vospominaniiakh, pp. 115-17; quotations are from L. Fatkin, “Teoriia informatsii,” in F.V. Konstantinov, ed., Filosofskaia entsiklopediia, vol. 5 (Moscow: Sovetskaia entsiklopediia, 1970), p. 213.
[105] Viach.Vs. Ivanov, “Nekotorye problemy sovremennoi lingvistiki,” Narody Azii i Afriki, no. 4 (1963): 171; see also Idem, Ocherki po istorii semiotiki v SSSR (Moscow: Nauka, 1976), pp. 141-42.
[106] Revzin, “Soveshchanie v g. Gor’kom,” p. 289.
[107] A.N. Kolmogorov, “On Works in Information Theory and Some of Its Applications,” in A.N. Shiryayev, ed., Selected Works of A.N. Kolmogorov, vol. III: Information Theory and the Theory of Algorithms, trans. A.B. Sossinsky (Dordrecht: Kluwer Academic Publishers, 1993), p. 220.
[108] A.N. Kolmogorov, “Three Approaches to the Definition of the Notion of Amount of Information,” [1965] in Shiryayev, ed., Selected Works of A.N. Kolmogorov, vol. III, p. 188.
[109] Ibid., pp. 188-89.
[110] Kolmogorov, “On Works in Information Theory,” p. 221 (emphasis added).
[111] Kolmogorov, “Three Approaches,” p. 190. Igor’ Mel’chuk developed somewhat similar ideas with respect to the complexity of linguistic descriptions; see I.A. Mel’chuk, “K voprosu o termine ‘sistema’ v lingvistike,” Zeichen und System der Sprache, vol. II (Berlin, 1962).
[112] A.N. Kolmogorov, “The Combinatorial Foundations of Information Theory and the Probability Calculus,” [1983] in Shiryayev, ed., Selected Works of A.N. Kolmogorov, vol. III, p. 210.
[113] Kolmogorov, “Three Approaches,” p. 192.
[114] N. Macdonald, “Language Translation by Machine—A Report of the First Successful Trial,” Computers and Automation 3:2 (1954): 6-10.
[115] Institut nauchnoi infomatsii.
[116] D.Iu. Panov, “Perevod s odnogo iazyka na drugoi pri pomoshchi mashiny: Otchet o pervom uspeshnom ispytanii,” Matematika: Referativnyi zhurnal, no. 10 (1954): 75-76.
[117] Institut tochnoi mekhaniki i vychislitel’noi tekhniki.
[118] D.Iu. Panov, A.A. Liapunov, and I.S. Mukhin, “Avtomatizatsiia perevoda s odnogo iazyka na drugoi,” in V.A. Trapeznikov, ed., Sessiia Akademii nauk SSSR po nauchnym problemam avtomatizatsii proizvodstva, 15-20 oktiabria 1956 g.: Plenarnye zasedaniia (Moscow: AN SSSR, 1957), p. 194.
[119] See O.S. Kulagina, “A.A. Liapunov i mashinnyi perevod,” in Pospelov and Fet, eds. and comps., Ocherki istorii informatiki, pp. 341-50; Mel’chuk, I.A. “Kak nachinalas’ matematicheskaia lingvistika,” in Pospelov and Fet, eds. and comps., Ocherki istorii informatiki, pp. 358-70.
[120] A.A. Liapunov, G.P. Bagrinovskaia, and O.S. Kulagina, “O nekotorykh metodologicheskikh voprosakh, otnosiashchikhsia Kolmogorov mashinnomu perevodu,” in A.A. Liapunov, Problemy teoreticheskoi i prikladnoi kibernetiki, ed. S.L. Sobolev (Moscow: Nauka, 1980), p. 183.
[121] See I.I. Revzin, “Vospominaniia,” in T.M. Nikolaeva, comp., Iz rabot moskovskogo semioticheskogo kruga (Moscow: Iazyki russkoi kul’tury, 1997), p. 794.
[122] Pervyi Moskovskii Gosudarstvennyi pedagogicheskii institut inostrannykh iazykov.
[123] Revzin, “Vospominaniia,” p. 794.
[124] Ibid., p. 810.
[125] Ob’edinenie po problemam mashinnogo perevoda.
[126] Uspenskii, “Serebrianyi vek,” p. 125.
[127] Ten issues of Biulleten’ Ob’edineniia po problemam mashinnogo perevoda came out in 1957-59; this edition was succeded by the series Mashinnyi perevod i prikladnaia lingvistika (1959-80), see V.Ju. Rozencvejg, ed., Machine Translation and Applied Linguistics, 2 vols. (Frankfurt am Main: Athenaion Verlag, 1974).
[128] Stenogramma zasedaniia rektorata MGU, January 14, 1957; the Central Municipal Archive of Moscow (Tsentral’nyi munitsipal’nyi arkhiv goroda Moskvy [TsMAM]), f. 1609, op. 2, d. 428, l. 46.
[129] Ibid., ll. 46-48.
[130] Viach.Vs. Ivanov, “Komitet po prikladnoi lingvistike,” Voprosy iazykoznaniia, no 3 (1958): 136-37.
[131] See Mashinnyi perevod i prikladnaia lingvistika, no. 1 (1959); Tezisy konferentsii po mashinnomu perevodu (15-21 maia 1958 g.) (Moscow, 1958).
[132] Uspenskii, “Serebrianyi vek,” p. 133.
[133] Ibid., p. 130.
[134] Laboratoriia elektromodelirovaniia.
[135] Uspenskii, “Serebrianyi vek,” pp. 137-41.
[136] Soveshchanie po kompleksu voprosov, sviazannykh s razrabotkoi i postroeniem informatsionnykh mashin s bol’shoi dolgovremennoi pamiat’iu, May 28-31, 1957, Moscow.
[137] Igor’ Aleksandrovich Mel’chuk, letter to author, 20 August 1998.
[138] “Reshenie Uchenogo Soveta filologicheskogo fakul’teta MGU,” December 24, 1958, as quoted in Ivanov, “Goluboi zver’,” no. 3, p. 160.
[139] Ivanov, “Goluboi zver’,” no. 3, p. 167.
[140] A.E. Kibrik and A.I. Koval’, “Voprosy prikladnogo iazykoznaniia v sbornike ‘Nauchno-tekhnicheskaia informatsiia,’” Voprosy iazykoznaniia, no. 5 (1965): 127 (emphasis original).
[141] Ivanov, “Nekotorye problemy,” p. 159.
[142] Ol’ga Sergeevna Kulagina, conversation with author, Moscow, July 3, 1997.
[143] Protokol zasedaniia partiinogo biuro ITMVT, March 27, 1962; the Central Archive of Social Movements of Moscow (Tsentral’nyi Arkhiv Obshchestvennykh Dvizhenii goroda Moskvy [TsAODM]), f. 7341, op. 1, d. 15, l. 98.
[144] Revzin, “Vospominaniia,” p. 815.
[145] Adapted from M.I. Steblin-Kamenskii, “Znachenie mashinnogo perevoda dlia iazykoznaniia,” in Tezisy konferentsii po mashinnomu perevodu, p. 23; and Idem, “Znachenie mashinnogo perevoda dlia iazykoznaniia,” Materialy po mashinnomu perevodu, vol. 1 (Leningrad: Leningrad University, 1958), pp. 3-9.
[146] I.I. Revzin, “Formal’naia teoriia predlozheniia,” in Tezisy konferentsii po mashinnomu perevodu, p. 50.
[147] Aleksandr E. Kibrik, “Iz istorii sovetskogo iazykoznaniia 60-kh—70-kh godov (o nashikh dostizheniiakh i poteriakh),” in André Clas, ed., Le mot, les mots, les bons mots: hommage a Igor A. Mel’cuk par ses amis, collegues et élèves a l’occasion de son soixantième anniversaire (Montréal: Presses de l’Université de Montréal, 1992), p. 17.
[148] I.A. Mel’chuk, “O terminakh ‘ustoichivost’’ i ‘idiomatichnost’,’” Voprosy iazykoznaniia, no. 4 (1960): 78.
[149] Mel’chuk, “K voprosu o termine ‘sistema.’”
[150] I.A. Mel’chuk, Review of Avtomatizatsiia v lingvistike, ed. and comp. L.N. Zasorina, Voprosy iazykoznaniia, no. 1 (1968): 141.
[151] Ibidem.
[152] I.I. Revzin and V.Iu. Rozentsveig, “Nekotorye voprosy teorii perevoda, sviazannye s obshchei problemoi avtomatizatsii informatsionnykh protsessov,” Nauchno-tekhnicheskaia informatsiia, no. 9 (1963): 34.
[153] See E.V. Paducheva, “Nekotorye voprosy perevoda s informatsionno-logicheskogo iazyka na russkii,” Nauchno-tekhnicheskaia informatsiia, no. 2 (1964).
[154] G.E. Vleduts et al., “Semiotika,” in A.I. Berg, ed., Kibernetiku—na sluzhbu kommunizmu, vol. 5 (Moscow: Energiia, 1967), p. 390.
[155] Ivanov, “Nekotorye problemy,” p. 164.
[156] Viach.Vs. Ivanov, “Lingvistika kak teoriia otnoshenii mezhdu iazykovymi sistemami i ee sovremennye prakticheskie prilozheniia,” in Lingvisticheskie issledovaniia po mashinnomu perevodu, vol. II (Moscow: VINITI, 1961), p. 21.
[157] D.M. Segal, “Konferentsiia po obrabotke informatsii, mashinnomu perevodu i avtomaticheskomu chteniiu teksta,” in T.N. Moloshnaia, ed., Strukturno-tipologicheskie issledovaniia (Moscow: AN SSSR, 1962), p. 280.
[158] Later Kulagina returned to machine translation research; see O.S. Kulagina, Issledovaniia po mashinnomu perevodu (Moscow: Nauka, 1979).
[159] Mel’chuk to Liapunov, January 8, 1965 (emphasis original); Natal’ia Liapunova’s archive.
[160] On evolution of the basic ideas of the “meaning-text” model, see Iu.D. Apresian, Izbrannye trudy, vol. I: Leksicheskaia semantika (Moscow: Iazyki russkoi kul’tury, 1995), pp. 36-55.
[161] A.K. Zholkovskii and I.A. Mel’chuk, “O semanticheskom sinteze,” Problemy kibernetiki 19 (1967): 181.
[162] E.g., the word prestupat’ would be included, meaning sovershat’ prestuplenie (literally, the verb “to crime” would be invented to provide the meaning “to commit a crime”).
[163] See Zholkovskii and Mel’chuk, “O semanticheskom sinteze,” p. 182. The idea of creating “Basic Russian” had already floated around for several years. Besides the references to “Basic English,” Soviet linguists also cited the French efforts to create “français élémentaire”; see V.Iu. Rozentsveig, “Iazykovaia praktika i lingvisticheskaia teoriia,” Voprosy iazykoznaniia, no. 2 (1966): 31-32. In February 1961, the famous physicist academician Petr Kapitsa addressed the presidium of the Academy of Sciences with a call for creating Basic Russian (“bazisnyi russkii iazyk”). A few days later he sent a letter to the secretary of the Party Central Committee Mikhail Suslov, raising this issue. One of the arguments Kapitsa gave was that “basic languages are much more susceptible to machine translation than literary languages”; see Kapitsa to Suslov, February 27, 1961; TsKhSD, f. 5, op. 35, d. 176, l. 9. The idea of creating Basic Russian, however, did not appeal to the authorities and the Central Committee Department of Science deemed Kapitsa’s proposal “unacceptable.”
[164] At a seminar on machine translation in the Department of Mathematics and Mechanics at Moscow University in 1966, Mel’chuk reportedly said, “I am not interested in machine translation at all. I want to understand how people really speak and to build a formal model of human language”; Vladimir Mikhailovich Alpatov, letter to author, 27-29 July 1998.
[165] I.A. Mel’chuk, Opyt teorii lingvisticheskikh modelei “Smysl—Tekst” (Moscow: Nauka, 1974), p. 20 (emphasis original).
[166] Ibid., p. 21; see also A.V. Gladkii and I.A. Mel’chuk, Elementy matematicheskoi lingvistiki (Moscow: Nauka, 1969), pp. 9-11.
[167] Aleksandr K. Zholkovskii and Igor’ A. Mel’chuk, “K postroeniiu deistvuiushchei modeli iazyka (‘Smysl-Tekst’),” in A.J. Greimas, et al., eds., Sign. Language. Culture (The Hague: Mouton, 1970), p. 159 (emphasis original).
[168] See Mel’chuk, Opyt teorii, esp. pp. 23, 31, 72-73. The concept of “semantic axioms” was inspired by Rudolf Carnap’s book Meaning and Necessity: A Study in Semantics and Modal Logic (Chicago: The University of Chicago Press, 1958) [the Russian translation, 1959]; Alexander Zholkovsky, letter to author, 13 August 1998.
[169] Igor’ A. Mel’cuk, Cybernetics and Linguistics: Some Reasons for as Well as Some Consequences of Bringing Them Together (Vienna: Osterr. Studienges. f. Kybernetik, 1977), p. 15.
[170] Ibidem.
[171] Roman Jakobson, “Linguistics and Communication Theory,” p. 496. Shannon’s original quote can be found in Claude E. Shannon, “The Redundancy of English,” in Cybernetics: Transactions of the 7th Macy Conference (New York, 1951), p. 157.
[172] Mel’chuk, Opyt teorii, p. 12.
[173] Mel’cuk, Cybernetics and Linguistics, p. 15.
[174] Ibidem.
[175] Ibidem.
[176] See A.K. Zholkovskii, N.N. Leont’eva, and Iu.S. Martem’ianov, “O printsipial’nom ispol’zovanii smysla pri mashinnom perevode,” Mashinnyi perevod: Trudy ITMiVT, no. 2 (1961): 17-46; Mashinnyi perevod i prikladnaia lingvistika, no. 8 (1964).
[177] “Protokol soveshchaniia po organizatsii v sostave Otdeleniia fiziko-matematicheskikh nauk AN SSSR instituta s kiberneticheskoi tematikoi, June 25, 1960,” in Pospelov and Fet, eds. and comps., Ocherki istorii informatiki, pp. 166-67.
[178] N.D. Andreev, “Soveshchanie po matematicheskoi lingvistike,” Voprosy iazykoznaniia, no 1 (1960): 131-37.
[179] Ibid., pp. 133, 137.
[180] Institut iazykoznaniia AN SSSR.
[181] Institut russkogo iazyka AN SSSR
[182] Institut slavianovedeniia AN SSSR.
[183] Institut vostokovedeniia AN SSSR.
[184] See V.P. Grigor’ev, “O razvitii strukturnykh i matematicheskikh metodov issledovaniia iazyka,” Voprosy iazykoznaniia, no 4 (1960): 153-55; Viach.Vs. Ivanov, “Akademik A.I. Berg i razvitie rabot po strukturnoi lingvistike i semiotike v SSSR,” in V.I. Siforov, ed., Put’ v bol’shuiu nauku: akademik Aksel’ Berg (Moscow: Nauka, 1988), pp. 164-86. Most of those divisions were closed by the early 1970s.
[185] Viach.Vs. Ivanov and S.K. Shaumian, “Lingvisticheskie problemy kibernetiki i strukturnaia lingvistika,” in A.I. Berg, ed., Kibernetiku—na sluzhbu kommunizmu, vol. 1 (Moscow and Leningrad: Gosenergoizdat, 1961), p. 220.
[186] Ibid., p. 218.
[187] Keldysh to the Central Committee, September 19, 1961, TsKhSD, f. 5, op. 35, d. 176, l. 73.
[188] Kukin and Kuznetsova to the Central Committee, September 26, 1961, TsKhSD, f. 5, op. 35, d. 176, l. 77.
[189] “Predislovie,” in Simpozium po strukturnomu izucheniiu znakovykh sistem: tezisy dokladov (Moscow: AN SSSR, 1962), p. 3.
[190] Viach.Vs. Ivanov, “Lingvistika i issledovanie afazii,” in Moloshnaia, ed., Strukturno-tipologicheskie issledovaniia, p. 93.
[191] Ivanov, “Goluboi zver’,” no. 3, p. 167.
[192] Protokol soveshchaniia po organizatsii v sostave Otdeleniia fiziko-matematicheskikh nauk AN SSSR instituta s kiberneticheskoi tematikoi,” June 25, 1960; Liapunov papers, ARAN.
[193] See Vleduts et al., “Semiotika,” pp. 394-99; Uspenskii, “Serebrianyi vek,” pp. 137-141.
[194] This term was widely popularized by logician Viktor Finn, see Revzin, p. 820.
[195] Ivanov and Shaumian, “Lingvisticheskie problemy,” p. 226.
[196] On Soviet semiotics, see also E. Meletinskij and D. Segal, “Structuralism and Semiotics in the USSR,” Diogenes, no. 73 (1971): 88-115; T.M. Nikolaeva, comp., Iz rabot moskovskogo semioticheskogo kruga (Moscow: Iazyki russkoi kul’tury, 1997); Stephen Rudy, “Semiotics in the USSR,” in Thomas A. Sebeok and Jean Umiker-Sebeok, eds., The Semiotic Sphere (New York: Plenum Press, 1986), pp. 555-82.
[197] Uspenskii, “Predvarenie,” pp. 199-200.
[198] A.A. Zalizniak, Viach.Vs. Ivanov, and V.N. Toporov, “O vozmozhnosti strukturno-tipologicheskogo izucheniia nekotorykh modeliruiushchikh semioticheskikh sistem,” in Moloshnaia, ed., Strukturno-tipologicheskie issledovaniia, p.134.
[199] Ibid., p. 141.
[200] Ivanov, Ocherki po istorii semiotiki, pp. 10, 145.
[201] Ibidem.
[202] On this landmark symposium, see Uspenskii, “Serebrianyi vek,” pp. 143-46; Ivanov, “Akademik A.I. Berg,” pp. 177-80. See the proceedings of the symposium in Simpozium po strukturnomu izucheniiu znakovykh sistem: tezisy dokladov (Moscow: AN SSSR, 1962).
[203] A.K. Zholkovskii and Iu.K. Shcheglov, “O vozmozhnostiakh postroeniia strukturnoi poetiki,” in Simpozium po strukturnomu izucheniiu znakovykh sistem: tezisy dokladov (Moscow: AN SSSR, 1962), p. 138.
[204] A.K. Zholkovskii, “Ob usilenii,” in Moloshnaia, ed. Strukturno-tipologicheskie issledovaniia, pp. 169-71 (emphasis original).
[205] Aleksandr K. Zholkovskii, “Deus ex Machina,” in Greimas, et al., eds., Sign. Language. Culture, p. 539.
[206] Ibid., p. 542.
[207] A.K. Zholkovskij and Ju.K. Sceglov, “Structural Poetics Is a Generative Poetics,” [1967] in Daniel P. Lucid, ed. and trans., Soviet Semiotics (Baltimore: The Johns Hopkins University Press, 1988), p. 182.
[208] Yuri Shcheglov and Alexander Zholkovsky, Poetics of Expressiveness: A Theory and Applications (Amsterdam: John Benjamins, 1987), p. 5.
[209] The authors counted ten elementary expressive devices: concretization, augmentation, repetition, division, variation, contrast, combination, preparation, coordination, and reduction. More complex devices, such as “recoil movement” or “sudden turn,” involved a combination of elementary ones.
[210] A.K. Zholkovskii and Iu.K. Shcheglov, Raboty po poetike vyrazitel’nosti: Invarianty—Tema—Priemy—Tekst (Moscow: Progress, 1996), p. 15.
[211] Zholkovsky, “Zh/Z,” p. 284.
[212] “Razgovory.”
[213] On Lotman, see A.D. Koshelev, comp., Iu.M. Lotman and tartusko-moskovskaia semioticheskaia shkola (Moscow: Gnozis, 1994); E.V. Permiakov, ed. and comp., Lotmanovskii sbornik, vol. 1 (Moscow: ITs-Garant, 1995); Ann Shukhman, Literature and Semiotics: A Study of the Writings of Yu.M. Lotman (Amsterdam: North-Holland Publishing Company, 1977).
[214] Uspenskii wrote in his memoirs that he had suggested the term “secondary modeling systems,” for the following reasons: (1) it sounded very scientific; (2) it was totally incomprehensible; and (3) in case of great urgency, an explanation could be cooked up: since natural languages modeled reality, they could be called primary modeling systems; other modeling systems of culture built on top of natural languages, therefore, could be called “secondary modeling systems”; see V.A. Uspenskii, “Progulki s Lotmanom i vtorichnoe modelirovanie,” in E.V. Permiakov, ed. and comp., Lotmanovskii sbornik, vol. 1 (Moscow: ITs-Garant, 1995), p. 106.
[215] Lotman praised the “great wealth” of Jakobson’s ideas, which pointed the “direction of further development for lingustic-semiotic studies,” (Lotman to Jakobson, July 20, 1980; Jakobson papers, MIT MC 72, box 43.25).
[216] Iurii Lotman, “Lektsii po struktural’noi poetike,” [1964] in Koshelev, comp., Iu.M. Lotman, p. 23.
[217] Ann Shukhman, “Soviet Semiotics and Literary Criticism,” New Literary History 9:2 (Winter 1978): 193.
[218] Lotman, “Lektsii,” p. 26.
[219] Yurij Lotman, The Structure of the Artistic Text, trans. Ronald Vroon (Ann Arbor: The University of Michigan, 1977), p. 28.
[220] Revzin, Models of Language, p. 150.
[221] See I.A. Mel’chuk, “Vsesoiuznoe soveshchanie po matematicheskoi lingvistike,” Mashinnyi perevod i prikladnaia lingvistika, no. 3 (1959): 84.
[222] Revzin, Models of Language, pp. 2-3.
[223] Zholkovskii and Mel’chuk, “O semanticheskom sinteze,” p. 177.
[224] Ivanov and Shaumian, “Lingvisticheskie problemy,” p. 232.
[225] I.I. Revzin, “Strukturnaia lingvistika i edinstvo iazykoznaniia,” Voprosy iazykoznaniia, no. 3 (1965): 45.
[226] Ivanov, “Goluboi zver’,” no. 3, p. 167.
[227] Revzin, “Strukturnaia lingvistika,” p. 45.
[228] Iu. Lotman, “Literaturovedenie dolzhno stat’ naukoi,” Voprosy literatury, no. 1 (1967): 95.
[229] A.N. Kolmogorov, “Intervention at the Session,” [1956] in Shiryayev, ed., Selected Works of A.N. Kolmogorov, vol. III, p. 32.
[230] Roman Jakobson, “O khudozhestvennom realizme,” in Idem, Raboty po poetike (Moscow, 1987), p. 387.
[231] Ivanov to Jakobson and Krystyna Pomorska, February 1, 1968; Jakobson papers, MIT MC 72, box 42.30.
[232] Ivanov to Jakobson, April 9, 1977; Jakobson papers, MIT MC 72, box 42.30.
[233] Lotman, “Literaturovedenie,” p. 100.
[234] Ibid., p. 90.
[235] S.K. Shaumian, “O sushchnosti strukturnoi lingvistiki,” Voprosy iazykoznaniia, no. 5 (1956): 44.
[236] Ibidem.
[237] Ibid., p. 46.
[238] Ibid., p. 47.
[239] Zholkovskii and Shcheglov, Raboty po poetike, p. 13.
[240] Lotman, “Lektsii,” p. 241.
[241] See Priscilla Johnson, Khrushchev and the Arts: The Politics of Soviet Culture, 1962-1964 (Cambridge, Mass.: MIT Press, 1965), esp. pp. 7-10, 105-105; Petr Vail’ and Aleksandr Genis, 60-e: Mir sovetskogo cheloveka (Moscow: Novoe literaturnoe obozrenie, 1996), pp. 190-92.
[242] Quoted in Uspenskii, “Serebrianyi vek,” p. 143. This critique was tempered in the final version of his speech.
[243] Quoted in Uspenskii, “Serebrianyi vek,” p. 145.
[244] Ivanov, “Goluboi zver’,” no. 3, p. 174.
[245] G.P. Serdiuchenko, “O nekotorykh filosofskikh voprosakh sovetskogo iazykoznaniia,” in V.V. Vinogradov, ed., Teoreticheskie problemy sovremennogo sovetskogo iazykoznaniia (Moscow: Nauka, 1964), pp. 126-27.
[246] F.P. Filin, “K probleme sotsial’noi obluslovlennosti iazyka,” Voprosy iazykoznaniia, no. 4 (1966): 32-34.
[247] Ibid., p. 44.
[248] Ibid., p. 36.
[249] Ibid., p. 38.
[250] E.M. Mednikova, “Metody lingvisticheskoi semantiki,” Voprosy iazykoznaniia, no. 3 (1969): 46.
[251] On charges of “formalism” in music, see, for example, Sheila Fitzpatrick, “The Lady Macbeth Affair: Shostakovich and the Soviet Puritans,” in Idem, The Cultural Front: Power and Culture in Revolutionary Russia (Ithaca: Cornell University Press, 1992), pp. 183-215.
[252] V.I. Abaev, “Lingvisticheskii modernizm kak degumanizatsiia nauki o iazyke,” Voprosy iazykoznaniia, no. 3 (1965): 27-28.
[253] Ibid., pp. 24, 28.
[254] Ibid., p. 32.
[255] Ibid., p. 34.
[256] Ibid., p. 33.
[257] Ibid., pp. 39, 42-43.
[258] Revzin, “Strukturnaia lingvistika,” p. 44.
[259] See “V Institute iazykoznaniia AN SSSR,” Voprosy iazykoznaniia, no. 4 (1957): 124-29.
[260] Revzin, “Vospominaniia,” p. 802.
[261] Yurij D. Apresjan, “Igor Mel’cuk: Brushstrokes for a Portrait,” in Clas, ed., Le mot, les mots, les bons mots, p. 2.
[262] Mel’chuk to Jakobson, July 12, 1974; Jakobson papers, MIT MC 72, box 44.17.
[263] “An Open Letter to Colleagues,” May 1976; Jakobson papers, MIT MC 72, box 44.17.
[264] Mel’chuk to Jakobson, May 30, 1977; Jakobson papers, MIT MC 72, box 44.17.
[265] Uspenskii, “Serebrianyi vek,” p. 159.
[266] Iu.D. Apresian, “Predislovie ko vtoromu izdaniiu,” in Idem, Izbrannye trudy, vol. I, pp. II-III.
[267] Ivanov, “Goluboi zver’,” no. 3, p. 166.
[268] Mel’chuk to Jakobson, June 20, 1977; Jakobson papers, MIT MC 72, box 44.17.
[269] Revzin, “Vospominaniia,” p. 804.
[270] Uspenskii, “Serebrianyi vek,” p. 127.
[271] Quoted in Uspenskii, “Serebrianyi vek,” p. 127.
[272] Revzin, Sovremennaia strukturnaia lingvistika, p. 214.
[273] Ibidem.
[274] I.I. Revzin, “Nekotorye trudnosti pri postroenii semanticheskikh modelei dlia estestvennykh iazykov,” in Simpozium po strukturnomu izucheniiu znakovykh sistem: tezisy dokladov (Moscow: AN SSSR, 1962), p. 22.
[275] Ibidem.
[276] Revzin, Sovremennaia strukturnaia lingvistika, p. 226.
[277] See Mikhail Bakhtin, Rabelais and His World, trans. Helene Iswolsky (Cambridge, Mass.: MIT Press, 1968); Viach.Vs. Ivanov, “The Semiotic Theory of Carnival as the Inversion of Bipolar Opposites,” trans. R. Reeder and J. Rostinsky, in Umberto Eco, Viach.Vs. Ivanov, and Monica Rector, Carnival! (Berlin: Mouton, 1984), pp. 11-35.
................
................
In order to avoid copyright disputes, this page is only a partial summary.
To fulfill the demand for quickly locating and searching documents.
It is intelligent file search solution for home and business.
Related searches
- six philosophies of education
- stock market last six months
- six grade writing worksheets
- six categories of new products
- chapter by chapter bible summary
- chapter by chapter summaries
- novel outline template chapter by chapter pdf
- chapter 9 chapter review answers
- the outsiders chapter 5 and six summary
- outsiders chapter summary chapter 5
- genesis chapter by chapter summary
- chapter by chapter bible study