Rubenstein-51
Accueil | Cours | Recherche | Textes | Liens

Centre de recherches en histoire et épistémologie comparée de la linguistique d'Europe centrale et orientale (CRECLECO) / Université de Lausanne // Научно-исследовательский центр по истории и сравнительной эпистемологии языкознания центральной и восточной Европы
-- Herbert RUBENSTEIN (Michigan State College) : «The Recent Conflict in Soviet Linguistics», Language, vol. 27, n° 3 (Jul-Sept. 1951), p. 281-287.

[281]  

         [The following objective report on certain aspects of linguistic study in the Soviet Union is printed here for the information of American scholars. The author has carefully refrained from injecting his personal views into what he intended as a purely factual account; but it may be proper to remark, by way of editorial comment and summary, that although the recent repudiation of Marr by Soviet linguists is undeniably a step in the right direction, so long as that repudiation remains a matter of dogma, promulgated by official edict, it is not yet a step from darkness into the light.—the editor]

        Until Stalin's condemnation on 20 June 1950, the most prominent name in Soviet linguistics was that of N. Ja. Marr (1864-1934). His ideas dominated the field because it was believed that they represented a new, Marxist orientation toward problems of language.[1]
        A few words regarding Marxism. Its set of general principles, viz. dialectical materialism, includes the following contentions: (1) matter is an objective reality and is the source of our sensations and ideas; (2) all things are knowable; (3) the knowledge of the laws of nature which we test by experiment and practice is real knowledge and objective truth. This is the material aspect of the doctrine. The dialectical aspect maintains that (4) all phenomena form a connected whole and interact with each other; (5) nature is in a state of continuous movement and change: there is always something arising and something disintegrating; (6) change is not always quantitative: it may abruptly become qualitative under the accumulation of quantitative change; (7) internal contradictions are inherent in all things, and the process of development results from these contradictions. The specific application of these principles to social phenomena results in a set of ideas called historical materialism: (8) society develops according to laws which are knowable; (9) the chief factor determining the social structure is a material one, namely the method of procuring the means of life necessary for human existence, i.e. the mode of production of material values; (10) the instruments of production and the people who operate them constitute the productive force. As the productive force changes, the relations between the people involved in the productive process must adjust to this change, though there may be a lag.[2]
        Historical materialism leads to a construct rather different from the culture of what is called 'bourgeois' science. The predominating position given to the mode of production and its relations leads to a division of culture into two unequal parts: the economic structure is the base; dependent on this is the
[ 282 ]
superstructure which Stalin defines[3] as the 'political, legal, religious, artistic, and philosophical views of society and their corresponding political, legal, and other institutions'.
        Marr attempted to set up a linguistic science that was completely in accord with his understanding of Marxian principles. His ideas changed rapidly, so that it is difficult to say what he believed and what he did not believe at the end of his career.[4] Marr's linguistic philosophy is embodied in his Japhetic theory. Before the revolution, his interest was focused on the languages of the Caucasus. Since he held that certain of these were most closely connected to the Semitic group, he called them Japhetic, after Noah's third son. In successive writings by Marr, the Japhetic group became ever more inclusive, until he was no longer able to explain its relation to other languages by the traditional techniques of historical-comparative linguistics. Thus, in 1908 Marr considered the Japhetic group to consist only of the Kartvelian languages (Georgian, Mingrelian, and Svanetian) and several extinct languages of Asia Minor; in 1916 he included also the Highland Caucasian languages; and by 1920 he had widened his class to such an extent as to include Basque, Etruscan, Pelasgian, and the various extinct languages of Asia Minor, e.g. Hittite, Urartic, and Elamitic. Marr maintained that the Japhetic speakers antedated the Indo-Europeans and the Semites in the development of Mediterranean culture. As he began to find Japhetic characteristics in the most diverse languages, he concluded that the Japhetic languages were not a genetically related group in the 'bourgeois' Indo-European sense, but rather a stage in language development, a stage which directly preceded that of the development of the Indo-European languages. Breaking with the method of historical linguistics and its assumption of proto-languages, Marr found only one path open: namely, a return to a monogenetic theory of language, i.e. to the view that all languages come from the same source.[5] Further, he stated that there is only one single process of development.
        How then do different languages come about, if they all develop from the same stock and thru the same process? Taking language as part of the superstructure, Marr declared that the process of development is made up of certain stages. The division between stages or periods corresponds to 'major changes in language and thought', which in turn correspond to 'major changes in productive technique'. Languages differ, then, according to their stage of development. Marr evidently thought this formulation inadequate to explain existing differences; for later he categorized languages according to periods of origin, i.e. he attributed special significance to the productive technique and the re-
[ 283 ]
suiting social structure existing at the time at which a language first arose.[6] On this assumption he based the following hierarchy: (1) language systems of the primary period (Chinese, some African); (2) language systems of the secondary period (Finno-Ugric, Turkic, Mongolian); (3) language systems of the tertiary period (surviving Japhetic languages, Hamitic); (4) language systems of the quaternary period (Semitic, Indo-European). These four classes of systems constitute the whole course of linguistic development. If a linguistic community deviates from the main current of world progress, its language will similarly deviate from the development of the system to which it originally belonged, and hence will grow apart from the other languages in that system; and this will remain true even if the community later returns to the main current of progress.[7] Although Marr did not accept the construct of the familiar proto-languages, he did acknowledge the fact that certain languages were more similar than certain others — a circumstance which he explained as due to social convergence, i.e. to contact with resulting hybridization, or to an original similarity of social conditions followed by social divergence.[8] Marr even went so far as to believe that he could reveal the four linguistic elements from which all languages developed: sal, ber, yon, roš. At first he identified these elements with the self-designations of four Japhetic tribes; later he advanced the idea that these elements also had a totemic significance; in his last works he suggested that there was an inseparable connection between the formal and the ideological aspect of these elements, but left this unclarified.[9] Each one of these elements has a number of variants, e.g. sal ~ žal, tsal, dal, gal, tkal, dgal, tskal, dzgal, etc.[10] Every word in every language consists of one or two, more rarely of three such elements. Following these ideas, Marr was able to compare words from languages completely unrelated in the historical-comparative sense—the more easily since he was not at all concerned with the histories of the words in question. Thus he considered Chuvash pus 'head', Basque buru 'bead', and Latin i-pse 'self' to contain the same element;[11] and he established kinship between German Himmel 'sky' and Russian zemlja 'earth' by breaking them down into two elements hi-mel and ze-mel.[12] Marr's linguistic paleontology rests entirely on a semantic basis; he devoted a great deal of energy to the investigation of semantic clusters, such as head ~ self, sky ~ earth, which he based on anthropological and archeological data. To Marr, language was from its very inception a class phenomenon — to such an extent that he declared[13] that 'languages of the same class in different
[ 284 ] 
countries with identical social structures exhibit greater typological kinship with each other than languages of different classes in one and the same country, one and the same nation.' As an example, he points out that Armenia and Georgia each have two national languages, one feudal and one popular, and that the Georgian feudal language is more closely related in system to the Armenian feudal language than either of them is to the popular language of its own country.
        These ideas were welcomed by many as the new Marxist linguistics. The reasons for this are apparent: Marr attacked the problem of the origin of language ('all things are knowable'), which had been given up as incapable of solution by 'bourgeois' linguists; he exploited the idea that changes in the base produce changes in the superstructure, by regarding linguistic change of every kind as due to social change, which in turn was due to a change in the mode of production; and in declaring that language development follows only one line, he drew a scheme parallel to the Marxist single economic development —primitive communal > slave-holding > feudal > capitalist > socialist > communist. When Marr divided this linguistic process into four discrete stages, he was following the dialectic of quantitative change becoming qualitative. In his insistence on the class nature of language, Japhetidology reflects the concept of the class struggle. In short, Marr attempted to incorporate all aspects of language into the Marxist view of social phenomena. There was only one doctrine in which Marr showed outright disagreement with Marxian tenets: in holding that language originated with a specific segment of the community. According to Marr, not all the members of the primitive community were able to say the four primary words. Those who could, acted as medicine men and used these words in their ritual to communicate with the totem and to gain the upper hand over the mutes.[14] Thus language was from its very beginning a class matter. Marx and Engels, on the other hand, believed[15] that 'language, like consciousness, arises only from the need, the necessity of intercourse with other men'; and Marxists in general hold that class distinctions did not arise until the slave-holding period.[16]
        As recently as January 1949, the official seal of approval was again placed on the main body of Marr's work when a special meeting was called in Moscow by the N. Ja. Marr Institute of Language and Thought, on a joint resolution of the Presidium of the USSR Academy of Sciences and representatives of institutes of the national republics working on language.[17] 'Bourgeois' formal-comparative linguistics was assailed as 'reactionary' and 'racist', 'justifying the imperialist colonial policy', and invalid because it considers language a 'closed entity' instead of examining it in its economic and social matrix. An editor's note to the account of the proceedings states that while the overwhelming majority of Soviet linguists base their work on the principles of dialectical and historical materialism, 'the work of the session revealed the generally unsatis-
[ 285 ] 
factory situation in the field of linguistics' and adds that there are 'major shortcomings' in this field. 'The Institute of Language and Thought was unable to present a single report throwing light on the basic questions of Marr's teaching.' No linguists in the field of the Russian language took part in the work of the session.
        On 9 May 1950, Pravda, the official newspaper of the Communist Party, began publishing articles by leading Soviet linguists for and against the Japhetic theory. The adverse articles, which were by far the more numerous, tore so at Marr's work that hardly a shred of it survived. It is of course obvious that none of the tenets of Marr's linguistic theory is supported by sufficient facts to satisfy the minimum requirements of objectivity. Marr's work, aside from his pro-revolutionary studies on Caucasian languages, must be considered, from the viewpoint of empiric science, purely programmatic. The conflict in Soviet linguistics was merely the question whether Marr's theories should be accepted as a program for future work or not: a question which arose because of the sterility of these ideas both for practical linguistics (composing grammars for the various nationalities within the USSR) and for linguistic theory.[18]
        There are three main theses in the Japhetic theory: (1) language arose mono-genetically and develops along a single path; (2) language is part of the superstructure; (3) language is a class phenomenon. The first idea — that of monogenesis and a single line of linguistic development—was completely rejected by Marr's critics. Even his own student and foremost defender, 1.1. Meščaninov, admits that in his own work he passes over in silence the question of the existence of the four primary elements, and concedes that the theory of development by stages must be reworked.[19] F. Filin, Secretary of the Presidium of the USSR Academy of Sciences and a strong proponent of Marr's theories, grants that Marr's stage scheme is incomplete and erroneous in so far as it is only based on morphological data. He defends Marr against the charge of racism by declaring that though Marr assigned languages to higher and lower places on the evolutionary scale, he intended no judgment regarding their ability to express content.[20]
        Stalin did not concern himself with this question in either of his two articles, but devoted himself almost entirely to the remaining two theses. There was no clear-cut majority opinion on these questions among the writers in the Pravda series, beyond the unanimous judgment that language was not a class phenomenon at the time of its inception. From the articles and letters following Stalin's discussions, it is clear that their doubts concerning the incorrectness of Marr's position on these two points are now entirely resolved.
        From the Marxist point of view, the whole Japhetic theory rests on the contention that language is part of the superstructure; for only then can language be so closely connected with the rest of the social fabric and, ultimately, with
[ 286 ] 
the mode of production. Once the superstructural conception of language is denied, as it is by Stalin, Marr's whole theory collapses. Stalin's main objection to considering language part of the superstructure is that changes in the grammar of a language do not coincide with changes in the base. The only linguistic changes that coincide with these are lexical: changes in the meanings of words, addition of new words, loss of words which have been rendered useless by changes in the social structure. But even on this level there is a core which is constant, the 'basic lexical fund'. Stalin points out[21] that 'the structure of language cannot be regarded as the product of any one epoch. The structure of language, its grammar and basic lexical fund, is the product of a number of eras.' Language, therefore, provides a certain national continuity. As a further feature serving to differentiate language from the superstructure, Stalin declares that the superstructure is only indirectly bound to the productivity of man (the mode of production influences the base, which in turn influences the superstructure), while language, at least as far as the vocabulary is concerned, is directly bound to productive activity, i.e. reflects changes in production without waiting for changes in the base.[22]
        Consistent with this opinion, Stalin rejects Marr's contention that language has always been and remains a class phenomenon — that a single non-class language common to a whole society and a whole people does not exist. He admits the existence of social dialects, but regards these as subordinate to the national language, whose significance is not at all weakened by them. We can see how great a significance the concept of national language has for Stalin from its use in one of his definitions:[23] 'A nation is a historically formed, stable community of persons which has emerged on the basis of a common language, territory, economic life, and psychological make-up and which manifests itself in a common culture.' Further, language resembles the means of production in being classless, i.e. in serving one class as well as another.
        Stalin advances a number of other ideas in his linguistic discussions. Since language, a means of communication, is a social phenomenon whose existence is bound up with that of a particular society, we can understand the development of a language only in connection with the history of that society. Linguistic change is not abrupt, 'not thru an explosion', but gradual. Grammar (morphology and syntax) is a set of rules not for particular words and sentences but for classes of words and sentences. The grammatical structure of a language changes even more slowly than the basic lexical fund; these two are 'the foundation of a language, the essence of its identity'. 'The main task of linguistics is to study the internal laws of languages.'[24] Semantics falls within the realm of linguistics and
[ 287 ] 
deserves investigation; but we must bear in mind that thought can only arise on the basis of linguistic material: it does not normally exist independently of linguistic material.[25]
        Judging from the letters and articles that followed Stalin's discussions, his ideas were warmly accepted by the leading Soviet linguists. Some of them, after making known their repentance for having followed the ideas of Marr even in part, suggested problems for future investigation: the internal laws of language, the history of word formation, an author's style and its relation to the literary language, the nature and means of utilizing language to express the superstructure, regularities in the development of a given language under socialist and former 'bourgeois' conditions, conditions for improving a language in a socialist state, the fate of the dialects of a given language, the changing psychological make-up of speakers under socialist conditions as reflected in language.
        Despite the official re-acceptance of the historical-comparative method and the breaking of ties with the Japhetic theory,[26] it is clear that the orientation of Soviet linguists is quite different from that of most American linguists. Soviet linguists show a far greater interest in correlations between linguistic and social facts; they are concerned with the general laws of linguistic change, and with principles to guide them in the practical work of setting up normative grammars for various minority groups hitherto illiterate. On the other hand, they seem to be relatively unconcerned with problems of pure description, and accordingly less interested in developing objective techniques for that purpose.


[1] This paper was originally presented to the Michigan Linguistic Society in East Lansing, 2 December 1950.

[2] This summary is derived from Joseph Stalin, Dialectical and historical materialism (New York, 1940).

[3] J. Stalin, Pravda, 20 June 1950; translated in the Current digest of the Soviet press 21.3. (This work will hereafter be abbreviated CD; all references are to Vol. 2.)

[4] According to I. I. Meščaninov in Pravda, 16 May 1950 (CD 19.7), Marr declared that his works prior to 1926 should be revised or, preferably, not read. [On Marr and the Japhetic theory see also Yakov Malkiel, Lg. 20.157 fn. 2; W. K. Matthews, The Slavonic review 27.172-92 (1948-9).]

[5] N. Ja. Marr, Vstupitel'naja řeč' k kursu obsčego učenija ob jazyke [Introductory lecture to a course in general linguistics] (1927), Izbrannyje raboty [Selected works, hereafter abbreviated IR], Vol. 2.16.

[6] Marr, Počemu tak trudno stať lingvistom-teoretikom [Why it is so difficult to become a theoretical linguist] (1928) 2.405.

[7] Marr, Jazyk [Language] (1927), IR 2.135.

[8] 'Mesˇcˇaninov, Pravda, 16 May 1950 (CD 19.5); V. V. Vinogradov, Pravda, 6 June 1950 (CD 24.16); S. Nikiforov, Pravda, 13 June 1950 (CD 25.17).

[9] Meščaninov, op.cit. (CD 9).

[10] Marr, Obščij kurs učenija ob jazyke [General course in linguistics] (1927), IR 2.96 ff.

[11] Marr, Gottentoty —  sredizemnomorcy [Hottentots—Mediterraneans] (1927), IR 4.115-6.

[12] The I in Himmel represents a special German development from n; cf. Goth. himins, and Mod. Ger. Orgel < OHG organa. The l in zemlja, on the other hand, is the so-called epenthetic l which arose in Proto-Slavic in the sequence labial plus j and is preserved in East and South Slavic.

[13] Marr, Počemu tak trudno stať lingvistom-teoretikom (1928), IR 2.415.

[14] Marr, Jazyk i myšlenie [Language and thought] (1931), IR 3.116.

[15] K. Marx and F. Engels, The German ideology, Marxist-Leninist library 17.19 (London, 1942).

[16] Marr explained that he was not speaking of classes in this Marxist sense. See his K bakinskoj diskusii o jafetidologii i marksizme [Contributions to the Baku discussion on Japhetidology and Marxism] 39 (Baku, 1932).

[17] Described by A. G. Spirkin, Voprosy filosofii, No. 3 (1949).

[18] There were of course other considerations which led to this controversy, e.g. the impossibility of getting a hearing for any views contrary to those of Marr. See G. Sanžeev, Pravda, 23 May 1950 (CD 21.9).

[19] Meščaninov, Pravda, 16 May 1950 (CD 19.9).

[20] F. Filin, Pravda, 30 May 1950 (CD 22.6).

[21] Stalin, Pravda, 20 June 1950 (CD 2.17).

[22] Stalin's argument proves at most that the 'structure of the language' (as distinct from the language itself) is not part of the superstructure. It is worth asking whether the structure of the language = the language. The answer is No if structural analysis is unable to deal with the problem of the distribution of lexical morphemes in terms of other such morphemes; for it is precisely in this distribution that the culture is most clearly reflected.

[23] Stalin, Sočinenija [Works] 2.296.

[24] Stalin, Pravda, 20 June 1950 (CD 21.7 and 8).

[25] Stalin, Pravda, 4 July 1950 (CD 28.3).

[26] A number of linguists, especially those primarily interested in the Indo-European languages, never accepted the Japhetic theory—e.g. Seliščev, Bubrix, Vinogradov, Ušakov, Reformatskij, and Freiman.