Linguistics and Adjacent Arts and Sciences: Part 2 9783110821659, 9789027931825

Citation preview

C U R R E N T TRENDS IN VOLUME • •

LINGUISTICS 12

CURRENT TRENDS IN LINGUISTICS Edited, by T H O M A S

A.

S E B E O K

Research Center for the Language Sciences Indiana University

V O L U M E

12

Linguistics and Adjacent Arts and Sciences •* Associate Editors: ARTHUR S . ABRAMSON, DELL HYMES, HERBERT RUBENSTEIN EDWARD STANKIEWICZ

Assistant Editor: BERNARD SPOLSKY

Assistants to the Editor: ALEXANDRA D I LUGLIO LUCIA HADD ZOERCHER

1974

MOUTON THE H A G U E • PARIS

© Copyright 1974 in The Netherlands Mouton & Co. N.V., Publishers, The Hague The research reported herein was performed pursuant to a contract with the U.S. Department of Health, Education, and Welfare, Office of Education, under the authority of Section 602, Title VI, NDEA.

LIBRARY OF CONGRESS CATALOG CARD NUMBER: 64-3663

Printed in The Netherlands by Mouton & Co., Printers, The Hague

PART T H R E E

L I N G U I S T I C S A N D T H E V E R B A L ARTS

S T R U C T U R A L POETICS A N D L I N G U I S T I C S

EDWARD STANKIEWICZ

1. Structural poetics is that trend in modem literary theory and practice which tries to apply to the study of literature strict and objective methods and which starts with the premise that literary works, as verbal art, cannot be studied without reference to the linguistic material of which they are made. As a discipline, modern poetics is both old and new. It is old for at its best it is keenly aware of a tradition which goes back as far as Aristotle, the founder of a descriptive science of poetics and of rhetoric, as well as of its ties with literary scholarship insofar as that deals with the structure of literary texts rather than with their extra-literary causes or consequences. It is new in that it is conscious of the modern conception of the autonomous function of art, and in that it shares with its sister-discipline, linguistics, some of its basic concepts, methods, and concerns, as well as many of its practitioners. Although poetics is still considered by many to be a 'baby science' (called so about a hundred years ago by one of its 'modern' pioneers, G. M. Hopkins (1959: 106), who proposed to study poetry with the 'microscope and dissecting knife'), it has grown into a field which covers literature in its most diverse aspects, including verse, prose, stylistics, literary history, and the typology of literary forms. But like structural linguistics itself, to whose twist and turns it shows great sensitivity, it has multiplied into a number of different schools and approaches which profess to be 'structuralist' and in one way or another related to linguistics. Any attempt to reduce the various tenets of these schools to one common denominator could not but fail. Structuralism for these schools can be said (to use the felicitous phrase of J. Piaget) to be more a method than a doctrine. Even structural linguistics, which has an older tradition and more clearly defined boundaries, presents a variety of directions and goals which one would find hard to treat en bloc unless one were to point out their most superficial traits or identify with structuralism only one of its schools of thought (as does Piaget in his characterization of modern linguistics, 1968: 63ff.). What can be said to unify the diverse approaches of structuralist poetics is their common effort to infuse poetics with a precise methodology, the use of a common or similar terminology, and the distinction between the literary work as a message and the organizing properties of language and literary conventions as its code or codes. Poetics is indebted to modern linguistics for some of its basic

630

EDWARD STANKIEWICZ

discriminations, including the one between code and message, as well as for such distinctions as diachrony and synchrony, the 'double articulation' and levels of language, and the axes of simultaneity and succession. Attempting to become an independent science whose focus is the structure of poetic texts and their underlying codes, poetics looks on the one hand towards linguistics and on the other towards semiotics, the general science of signs which incorporates both linguistics and poetics and which explores all types and functions of signs in their multifarious variety and interaction. The connection of poetics with both linguistics and semiotics was clearly formulated over forty years ago in the Theses of the Prague Linguistic Circle: 'Only poetry,' says one of the Theses, 'enables us to experience the act of speech in its totality and reveals to us language not as a ready-made static system, but as creative energy,' while another Thesis proclaims, 'Everthing in the work of art and its relation to the outside world . . . can be discussed in terms of sign and meaning; in this sense aesthetics can be regarded as a part of the modern science of signs, of semiotics.' As the first of these Theses suggests, the relation between poetics and linguistics is not weighed to one side, with the former dependent on or subservient to the latter. Poetry and the making of poetry belong to the most universal and extensive uses of language, and any exploration of the structure of poetry is bound to broaden the linguist's understanding of his subject matter, to reveal the capacities and limits of language, and the extent to which it can be manipulated and changed. Despite the eternal plaints by poets and philosophers about the tyranny and opacity of language and about its unsuitability for artistic purposes ('Words strain, crack and sometimes break under the burden; under the tension, slip, slide, perish,' writes T. S. Eliot), the truth of the matter is that language is the poet's only instrument, and it remains a flexible and perfect instrument, since 'the major characteristics of style, inasfar as style is a technical matter of the building and placing of words, are given by the language itself, quite as inescapably as the general acoustic effect of verse is given by the sounds and natural accents of the language' (Sapir 1921:226). The poet, Sapir concludes, need not be 'an acrobat,' — 'it is enough for him to make his personality . . . felt as a presence' (Sapir 1921:227). The test of the soundness of a structural approach to poetry must ultimately be measured by the depth of the explanations it offers about the nature of a poetic text and the general properties of poetry. Poetic texts are complex but objective linguistic phenomena, whether they exist in written or oral form, whether they are new or old, and whether they are works of outstanding authors or of humble 'singers of tales'. And it is in the very treatment of the literary work as an esthetic structure that one may note the greatest divergence in the modern approaches to poetics. At one extreme of the structuralist spectrum, the literary text is viewed as a deviation from the 'ordinary', spoken language. Poetic language, for Todorov, is the antithesis of 'bon usage', for 'son essence consiste dans la violation des normes

STRUCTURAL POETICS AND LINGUISTICS

631

du langage' (1965a:305), while Guiraud defines it is an 'écart par rapport à la norme collective' (1967:4). In this approach (which is in particular favor with generative grammarians; for the general method; cf. the study by J. Katz 1964) the individual sentences of a literary text such as a poem are compared with 'well-formed' sentences of the spoken language in order to measure the degree of their deviation from the norm, with the predictable result that many sentences are indeed 'deviant', and that, as one author puts it (Levin 1964: 37), metaphor is 'content which could have been expressed in direct language without any loss'. Despite its modern terminology and intricate operations, this approach clearly harks back to those traditions of rhetoric where poetry meant 'poetic license' and the use of a certain number of tropes and figures, or 'embellishments'. No attempt is made by these students of poetry to define the difference between deviations that are poetic from those that are not poetic, nor do they provide a clear definition of what constitutes a 'neutral' norm. Poetry which does not deviate (as is often the case in simple folksongs or in classical traditions) would present no real interest for them. The entire approach, then, can be defined as an extension of linguistic analysis to literary texts which are fragmented into isolated sentences that are studied with relation to prose rather than to each other. No less atomistic, and therefore at odds with the integrated structure of a poetic text is that method which submits the text to a test of response on the part of its readers (or listeners) by means of a so-called 'stylistic device' for the purpose of determining ('directly' and 'rapidly'(!)) stylistic peculiarities and unpredictable forms. The poetic text is believed to consist of no more than a series of verbal stunts which jerk the reader out of 'automatic, semi-conscious decoding' (Riffaterre 1959:166; 168). The difference from the former method is only in that the stylistic deviations are here determined not by the linguist but by the diffused and superficial 'responses' of readers, without any attention to their different norms of expectation or to the fact that old and familiar literary texts still retain their appeal. Another approach which ignores the specificity of the poetic text is represented by the French school of structuralism which has in the last decade moved into the forefront of structural poetics. This school has taken most seriously the de Saussurian separation of langue and parole, declaring that only the former (i.e., the poetic code) lends itself to a scientific analysis, whereas the latter (i.e., the text) is a matter of subjective impressions, and consequently of no scientific interest. 'On ne peut pas diviser l'oeuvre; on ne peut parler de la structure de l'oeuvre . . . car il n'y a aucun moyen pour l'attester; on ne dispose que d'impressions contradictoires de différents lecteurs' (Todorov 1968:106). For the representatives of this school the text is as silent and insular as it was for Croce (for whom it was an act of individual expression), since the text presumably refers only to itself. Literature is, according to R. Barthes, 'a system of deceptive signification' which is 'never finally signified'; 'la littérature est au fond une activité tautologique, comme celle de ces machines cybernétiques construites pour elles-mêmes' (1964:148). Structural

632

EDWARD STANKIEWICZ

analysis must consequently become 'une choix pour la syntaxe contre la sémantique' (P. Ricoeur 1963:608) or an analysis which traverses 'la substance de l'oeuvre pour atteindre sa ossature' (G. Genette 1966:34). The literary code or 'literariness' (translated from the Russian Formalist term literaturnost') is attained by comparing and abstracting the formal properties of various texts in an infinite regression which leads from text to text, from one level to another, to a primitive 'Ur-code'. 'La poétique', writes Todorov, 'est en quelque sorte un langage — non le seul — dont dispose la littérature pour se parler. Chacune d'elles est un langage qui traite de l'autre; et en même temps chacune d'elles ne traite que d'elle même' (1968:164). The literary text serves, in this system, only to define the code, while the code serves in turn to define the text or itself. The typology of poetic features which is set up by Todorov becomes then a descriptive system, an inventory of rhetorical devices which never intersect and which remain as isolated as the text itself.1 Besides these 'deviationist', pragmatic, and neo-Formalist approaches to literature there are various schools and centers of structural poetics which pay attention both to the structure of the text and to its underlying codes, and which treat the poetic work not as a hermetically closed and immutable structure, but as a 'struttura aperta' (Umberto Eco) which is interpreted and completed by the reader (or listener) in the process of reading (or listening) in a historically defined social setting and against the background of other more or less interiorized texts. Such an open-ended and dynamic approach to literature, in which the 'text', reader, and underlying codes complement each other, was already foreshadowed in the Theses of Jakobson and Tynjanov when they broke away from the static 'immanentism' of early Formalism. To what extent the various centers and workers in the field of poetics subscribe to such an expanded and deepened program of structural poetics remains an open question. The scholarly polemics of the last years have, at any rate, not stood in the way of their mutual collaboration and of the advancement of common goals. Centers of structuralist poetics are now scattered in various parts of Europe and America, and their vitality is attested to by international conferences and a spate of new journals and publications. Particularly active are the centers of poetics or semiotics which have sprung up in the last decade or so in the Soviet Union (especially around the universities of Tartu and Moscow), in Poland (Instytut Badan Literackich), in Italy (around the journal Strumenti critici), and in the United States (around the journal Language and Style and the international journal Semiotica). An active school of structural poetics existed until recently (1968) in Czechoslovakia, and new centers have now emerged in West and East Germany and in Yugoslavia. All of these centers pursue somewhat different goals in line with their local traditions and interests (the most important studies in versification appear regularly in Poland, 1 For an incisive critique of French Structuralism, consult the works of H. Meschonnic, F. Jameson, and especially U. Eco (1968).

STRUCTURAL POETICS AND LINGUISTICS

633

whereas significant studies on esthetics and semiotics are being produced in the Soviet Union and in Italy). Despite the lack of a unified approach or methodology, poetics proceeds to probe everywhere the same basic questions of literature, thereby enriching both the study of language and that of language in the function of art. 2. Structural poetics is in many respects a direct descendant of Russian Formalism, from which it has retained some basic premises while rejecting its extreme formulations and solutions. The avowed purpose of Formalism was to transform the study of literature, which was dominated by an emotive impressionism and by utilitarianism, into a discipline of 'laws' or general principles at a time when, as a humanistic discipline (a Geisteswissenschaft), it was considered to be unamenable to such an operation. The emancipation of literary scholarship from other sciences (especially psychology and sociology) coincided with a similar development in linguistics, which began to define itself (through the works of Baudouin de Courtenay and de Saussure) as an autonomous discipline set on discovering the general laws of language and its internal development. The same trend is conspicuous in art-history, which has proclaimed to have an internal history of forms without subject-matter and without heroes (Wòlfflin's Kunstgeschichte ohne Namen). These trends have been connected at the same time with the liberation of the arts from practical pursuits and it is the poets themselves who first provided the new formulations of poetry. These developments coincided, curiously enough, with the emergence of the realistic and psychological novels which tried to get hold of 'reality' by presenting a 'slice of life', or which proclaimed the superiority of 'showing' to the art of 'telling'. This polarization of prose and poetry contributed, no doubt, to a sharper formulation of the programs of poetry as we find them in the writings of the Symbolists and later in those of the Futurists and Formalists. In declaring that 'poetry is poetry, and not another thing" (T. S. Eliot), and that 'art does not compete with elephants and locomotives' (E. E. Cummings), the poets discovered that the essence of a poem is its formal composition, and that the composition is a matter of verbal patterning, or as Mallarmé tried to put it (to his friend Degas), 'it is not with ideas that one makes a poem, but with words' (see Valéry 1958:63). Beginning with Poe's Philosophy of composition and Baudelaire's statement that 'grammar, dry grammar becomes the magic of evocation' (Le poeme de Haschisch, 1858), one can witness a steady flow of the most fruitful observations about poetry coming from the pens of poets, coupled quite often with the most extravagant claims for its ultimate meaning and mission. In rejecting the previous roles of Seer or servant to society, the poet comes now to see himself primarily as a craftsman, a 'savant austère' who must construct his poem like 'a mathematical formula' (E. A. Poe). It is subsequently from Valéry that we learn that poetry is 'the double invention of content and form' and the 'continuous oscillation between sound and meaning' (1958:74, 204), while G. M. Hopkins teaches us that a poem is based 'on parai-

634

EDWARD STANKIEWICZ

lelism of expression [which] tends to beget or passes into parallelism of thought' (1959:84). These observations and studies of poetry, which continue with the succeeding generations of poets (in the West, T. S. Eliot, Ezra Pound, Empson, the New Critic, and in Russia, Belyj and Brjusov), testify to the new and deepened concern of the poets with their craft, since they have been, in effect, engaged in creating a new experimental poetry daring in its breadth and diversity. Among its innovations which have changed the face of modern poetry are: the creation of a new type of syntax (beginning with the poetry of Mallarmé and the Futurist program of les mots ert liberté), a new type of semantically distant, 'bold' metaphors (in the works of Rimbaud and the Surrealists); the exploits of multiple linguistic codes and of puns (in Valéry's use of French words with allusions to their Latin prototypes; the punning of Lewis Carroll and James Joyce); a new emphasis on sound texture (leading from the musicality of Paul Verlaine to the poetry of pure sound in zaum'), free verse and the new poéme en prose; the exploration of the interplay between the auditory and visual aspects of a poem (beginning with the Calligrammes and Anagrammes of Apollinaire and the typographical arrangements of Majakovskij's verse up to contemporary concrete poetry); the play with and discussions on the merits of various parts of speech for poetry (as in the works of the Imagists), and the juggling of grammatical and derivational forms (as in the poetry of V. Xlebnikov, which set Roman Jakobson on his study of the 'grammar of poetry'). These developments in poetry and in literary theory were bound to put an end to academic literary theory which was biographical, sociological, historical, etc., and to the old (but never defunct) tenets of traditional Rhetoric according to which art is a reflection of Reality (conceived in one way or another) with the addition of ornament (the classical prodesse et delectare, or, in Dante's formulation, veritade ascosa sotto bella menzogna [Convivio II]). Set on abolishing any reference to external reality (on 'abolition de choses'), modern poetry became hermetic and tried to create a difficult and new language which would differ from ordinary language, or to assimilate the language of poetry to music or the other non-representational arts. It is for this reason that Valéry believes that 'pure poetry' need not carry any communicable meaning and must resemble dance and magic formulas, while the poet must try to 'draw a pure, ideal voice' from 'practical, changing and soiled language, a maid of all work' (p. 81). None of these ideas escaped the Formalists, who learned (primarily from their Symbolist mentor, A. Belyj) that 'form itself has the power to act upon us'. The belief in a formal art above and apart from meaning, or at least in the possibility of separating the formal and semantic components of art, shaped the ideas of even the more moderate Formalists. 'The material [of poetry],' wrote Zinmraskij, 'is not completely divorced from practical utility and therefore is not completely subject to the laws of purely artistic structure'; consequently 'one can speak [in poetry] of a purely linguistic structure which presents us mainly with phonetic and syntactic problems' (1966:20). These and similar

STRUCTURAL POETICS AND LINGUISTICS

635

pronouncements of the Formalists were of course not only a matter of purely theoretical concern, but were intended, at least in part, to justify the poetic experiments of their Futurist colleagues.2 Over the brief but agitated period of its career, Formalist doctrine moved from extreme positions which recognized in poetry the importance only of sound or of the 'sound-gesture' (Polivanov, Jakubinskij, fijxenbaum, Brik), to more moderate views of poetry as 'composition' and to the study of its technical devices. In addition to its path-breaking explorations of the formal properties of verse (Zirmunskij, Jakobson, Tomasevskij, Tynjanov), it advanced the study of prose, whose chief esthetic characteristics were seen to lie in the syntagmatic organization of the plot (sjuzet) at the expense of such elements as the theme (fabula) and the hero, which were treated only as a pretext (motivirovka) for the construction of the plot (in the studies of Sklovskij and Ejxenbaum). The analysis of individual works (to which they brought considerable critical acumen) was not for the Formalists a goal in itself, but was used rather as a springboard for the construction of a typology of forms, or (as in the case of Propp) for the reduction of a large body of texts to a finite number of invariant elements. Given the basic premises, the result could not have been more than a taxonomy of technical devices. At a later period the Formalists (especially Tynjanov and his Prague followers) introduced the notion of 'foregrounding', i.e., of the hierarchy of formal elements without, however, coming to grips with the question of the semantic organization of a work. History of literature was likewise interpreted by them as a self-contained and self-regulated process which oscillates like a pendulum from form to form. The only explanatory principle for this constant alternation of forms was the principle of 'de-automatization', i.e., of the awakening of a stultified perception that sets in with the wearing out of forms. The neglect of meaning did not prevent the Formalists from reading into poetry 'deeper' meanings, or from endowing it with a cognitive function. In the same way that some contemporary painters came to claim that the mere arrangement of color and line provides 'a bridge from the visible to the invisible' (Beckmann 1970:98), or that there is 'an innate affinity between pictorial elements and emotional states' (Kandinsky), poetry was believed to be able to provide a deeper insight into reality by jerking the mind from its natural apathy. This theory of 'de-familiarization' (ostranenie) which is sometimes considered to be one of Formalism's great insights into literature, was actually implicit in the theories of the Symbolists, and a part of the Romantic inheritance according to which poetry has the power to unveil 'analogies' and 'correspondences', i.e., transcendental truths which are not accessible to ordinary perception. Sklovskij's theory of 'making strange' (which he shared with Pound and Brecht) was formulated a century before him by Coleridge. '[His] general purpose', wrote Coleridge about his friend Words!

Exhaustive historical surveys and critical analyses of the Formalist movement can be found in the works by V. Erlich, K. Pomorska, J. Striedter, and E. Thompson. Useful remarks on Formalist tenets are also scattered in the work of Ju. Lotman.

636

EDWARD STANKIEWICZ

worth, 'was to give the charm of novelty to things of everyday and to excite a feeling analogous to the supernatural by awakening the mind's attention to the lethargy of custom' (1817:409). In addition to their achievements in the study of verse, the Formalists and their associates made important advances in the study of prose, especially in the exploration of the role of 'speech within speech', or the use of dialogue and monologue in the structure of the novel (V. Vinogradov, Volosinov, Baxtin). This line of research, which was initiated in the West by Flaubert and Henry James, the founders of the novel with 'shifting perspectives', and continued by E. M. Forster, Lubbock, Friedeman, and others, could have served the Formalists as a reminder that a work which sets itself 'realistic goals' and tries to capture reality 'as it is', in the flow of its appearances, need not relinquish its function as art and can even create new artistic values. 'A novel', wrote James in The art of fiction, 'is a living thing, all one and continuous like any other organism' (1884:5); art must therefore conceal itself, giving the reader the sense of a living presence, for otherwise it becomes 'a betrayal [of the artist's] sacred office' (1884:5). In spite of similar declarations, even the realistic novel (in which the Formalists took little interest) managed to create a new language of metonymy and metaphor, in which small, intimate and inanimate objects began to glitter with the brilliance of poetry (Weinrich 1971:35ff.). In their zeal to grasp and to defend the canons of modern, 'formal' art, the Formalists themselves fell prey to a narrow normative bias, forgetting (or anxious to forget) that entire traditions of art stubbornly pursued referential goals which they managed to harmonize with the esthetic enterprise. It is enough to think of Leonardo da Vinci, whose paintings and drawings are an inexhaustible source of anatomical, zoological, and geological information, and who glorified the power of the eye and of experience, being at the same time fully aware that painting remains 'una cosa mentale' (Trattato della pittura [ = 1890]:9a). Gombrich adduces the interesting example of Constable, who wrote about his native Suffolk: 'It is a most delightful landscape for a painter. I fancy I see Gainsborough in every hedge and hollow tree' (1960:316). The lesson of this being that man and his art is incapable of imitating nature (or society) without first performing a selection of the pertinent features of experience and without putting them into a mold which reinterprets and organizes this experience. And just as language itself is 'an organization of experience into formal patterns' (Sapir), so verbal art needs must give shape and organization to individual experience in terms of available linguistic and literary codes which are implemented and concretized in specific texts. And since the verbal material of poetry is always meaningful (if one ignores the marginal use of nonsense words in some forms of poetry), any attempt to ignore its semantic aspect becomes just like linguistics without semantics — reductive and one-sided. Archibald McLeish's requirement that 'a poem should not mean, but be' was obviously inspired by the slogans of modern 'hermetic' art, as well as by the misconception that 'meaning' means simply referential meaning, a view which was

STRUCTURAL POETICS AND LINGUISTICS

637

dear to the Positivists and neo-Positivists. Poems are, for I. A. Richards, as they are for Carnap, only 'pseudo-statements' which 'assert nothing' except the emotive, and consequently private 'attitudes' of the poet (cf. esp. Richards 1926, Ch. VI). The emotive interpretation of poetry (which was subsequently embraced by S. Langer and others, and is now again resuscitated by J. Cohen; cf. Cohen 1966: 149, 205) was never accepted by the Formalists who, on the other hand, were content to confine their definition of poetry to its formal structure ('the emphasis is on the message') without specifying the semantic nature of this message. The text thus became a tautological structure, referring only to itself. 3. The discovery that language cannot be reduced to mere reference (or denotation) is one of the most fruitful attainments of modern linguistics, and it was an integral part of Prague structuralist thought in linguistics and poetics. At the IVth International Congress of Linguists (1938), Mukafovsky proposed to expand Biihler's model of language, which included a referential, expressive, and appellative function, by adding to it a fourth, esthetic function which he defined, however, as being self-referential, i.e., as constituting a meta-language. This proposal had the advantage of placing 'poetic language' not only in opposition to merely 'referential language' (as was done in the West by I. A. Richards and the New Critics, and in Russia by the Formalists), but of seeing it in the broader context of the variegated uses of language. Biihler's model had already made it clear that language is not used simply to convey information about the outside world, but also as the foremost vehicle of social interaction. That language is not merely a tool for the description and manipulation of empirical facts had already been fully realized by the British empiricists, those assiduous and tireless students of language as an instrument of cognition. 'The communicating of ideas marked by words is not the chief and only end of language, as is commonly supposed', wrote Berkeley; 'there are other ends, as the raising of some passion, the exciting to or deterring from an action, the putting the mind in some particular disposition; to which the former is in many cases barely subservient, and sometimes entirely omitted, when these can be obtained without it, as I think doth not unfrequently happen in the familiar use of language' (1965: § 20, p. 57). These insights of Berkeley, which show his awareness of the various functions and styles of language, have often been forgotten in modern linguistics, which has tended to treat language as a closed and monolithic system which is independent of speech and society. The treatment of language as a set of 'negative, relative and oppositional terms' which are related to each other, like in a game of chess, in tight, internal configurations, is particularly characteristic of de Saussure's conception of language, and came about as a natural, though extreme reaction to the older theories of language which had treated it only as the 'speech activity' of the individual speaker. It was de Saussure's outstanding merit to attempt to separate the invariant and constant properties of language as a code from the variable and fluctuating phenomena which are displayed in the concrete

638

EDWARD STANKIEWICZ

message. A more flexible position with respect to the langue/parole dichotomy was taken, as we know, by Baudouin de Courtenay, who coined the term 'collective individuality' to indicate the interaction which exists between the obligatory, collective norm and the creative, individual speech-art (1972:20ff.). The relation langue/ parole cannot, however, be reduced (as thought by de Saussure and as the Geneva school assumed) to the question of a collective norm and individual expression, for the code and speech present dialectically interrelated and complementary aspects. The separation of langue and parole is also a shortcoming of some contemporary theories of language. 'The defect of the Chomsky's theory', writes J. Searle in a recent review 1972:23), 'arises from the failure to see the essential connection between language and communication, between meaning and speech-acts. The picture that underlies Chomsky's whole theory of language is that sentences are abstract objects that are produced and understood independently of their role in communication'. Only by viewing the code and the message in their interdependence, or, in Jakobson's terms, by adopting the model of means and ends, is the linguist able to give a full and correct interpretation of language. It is the merit of Jakobson to have refined and broadened Biihler's and Mukarovsky's models by re-defining the nature of some linguistic functions and by adding to them two more functions. Thus Jakobson points out that in addition to the referential, expressive (or emotive) and appellative (or conative) functions, language can also be used in a metalinguistic function (as was made clear by the works of modern logicians), i.e., when the message is used to refer to the code (or to other codes), and in a phatic function which serves to establish, maintain, or sever contact between the interlocutors (1960:356-57). The poetic function is, in turn, defined in opposition to the metalinguistic function: in the former 'the equation is used to build a sequence' whereas in the latter 'the sequence is used to build an equation' (1960:358). Jakobson further points out that the various functions do not occur in isolation but interact with each other in concrete messages in which one or the other function becomes dominant. This comprehensive and cohesive model of Jakobson, nevertheless, raises certain questions which must be taken up in further investigations. First it may be pointed out that the established functions do not exhaust all the possible uses of the message. One might, for example, add a 'deferential' (or 'distancing') function which defines the social status of the speakers in the speech-act (a function which is especially conspicuous in languages which employ honorifics and deferential sub-codes, i.e., special grammatical and lexical forms to express this function), and an emphatic function (the distinctive properties of which were studied by such linguists as Gy. Laziczius (1936), Trubetzkoy (1949:16ff.), and Bolinger. Second, it is important to indicate that the nature and hierarchy of the various functions are not only different, but incommensurable. The referential function is obviously the foundation of any complete message or sentence which arises through the use of a predicate and which necessarily establishes (by means of verbal shifters) a relation between the

STRUCTURAL POETICS A N D LINGUISTICS

639

narrated event and the speech-event. This predicative or propositional function of language makes possible the question about the truth value or modality of any message which distinguishes language from all other sign systems. The other functions of language (e.g., the expressive or phatic functions) are, on the other hand, concomitant with or superimposed upon the referential function and are expressed by limited sets of devices. If verbal art employs sentences whose foremost feature is their propositional function, the claim that poetry is 'neither true nor false' (as argued by Carnap and his followers) cannot be maintained without further qualification. It is equally misleading to assume that the poetic function is simply 'superimposed' upon the referential function, for such a view would lead to the impasse of the old rhetorical doctrine according to which poetry expresses 'truth' with the addition of, or despite its ornaments. If poetry invariably appears to us as fiction (whether it refers to true or, more often, to imaginary events) and prevents us from raising the question of its direct reference and truth-value, it is rather because poetry is based on the principle of 'multiple exposure' or of simultaneous multi-dimensional reference, i.e., of the multivalence (the so-called 'ambiguity') of its meanings, which is created in the poetic message through the internal relations of its verbal signs. Poetry may tell both profound truths and bald lies (of which folk poets in particular like to boast), but these questions always remain marginal and subordinate to the basic question of opposing and blending different aspects of reality. This marginal, artistically irrelevant question of reference was very well understood and expressed by Cervantes, who has Don Quixote say to the Duchess, 'God knows whether Dulcinea does or does not exist in the world, and whether she is the product of phantasy or not; these are not things whose investigation can be carried through to the end'.3 The 'poetic function' cannot, furthermore, be put on a par with the other functions of language, since the code employs no special features to render this function, as it does in the rendering of the other functions. The 'poetic function' is thus always a function of the message itself, and involves poetic use of language or poetic speech, rather than a special poetic function of language. Poetry is, on the other hand, always able to expand the boundaries of the 'ordinary' linguistic code (see below, 5.) just as it tends, far more than 'ordinary' language to combine with other systems of signs. This tendency towards syncretism of various types of signs is one of the prominent features not only of verbal art but of all arts. According to Sapir, 'poetry everywhere is inseparable in its origins from the singing voice and the measure of dance' (1921:229fn. 11), and it is interesting to note that many languages lack 3

The passage in Spanish reads as follows: 'Dios sabe si hay Dulcinea o no en el mundo, y si es fantastica o no es fantastica; y estas no son de las cosas cuya averigacion se ha da llevar hasta el cabo,' Don Quijote de la Mancha, part 2, Ch. 32.

640

EDWARD STANKIEWICZ

a special word for a poem or the poet and use instead the word for 'song' or the 'singer' (for example in Serbo-Croatian the folk-singer is called pevai or guslar). Ethnographers have observed that without musical accompaniment, the execution of oral poetry becomes severely distorted, or cannot be performed at all (Ruzicic 1934:232). The role of music and of mimicry is also well known from the modern 'chanson' (Weinrich 1971:124ff.) which explains why the actual performance (and performer) is often far more important than the verbal text. There are entire genres of folk-literature which lie on the borderline of or oscillate between music and poetry (children-songs, lullabies, dirges, anthems). Modern poetry, which is meant primarily for visual consumption, is also ambiguous as to the proper nature of its signs, and a long tradition of emblems and ideograms, and the use of sets, mimicry and song in the theatre, all testify to the perennial syncretism of the various arts. The ubiquitous and simultaneous use of language for various functions makes it difficult to draw a sharp line between its esthetic and non-esthetic functions, and one can speak only of a continuous scale which goes from densely structured poetic texts to the use of poetic devices in everyday communication. This question can never be decided in internal terms alone, since the definition of art and of a work of art also depends on fashions and styles, and on the 'intention' of the reader as much as on that of the artist (e.g., a work of art executed primarily for a nonesthetic purpose may be 'read' as a purely esthetic product when the original purpose is ignored or forgotten). The use of formal devices, such as verse, is certainly insufficient for the definition of verbal art, as was recognized by Aristotle, who saw that Empedocles and Homer have nothing in common except verse (Poetics, I, 1447b). Artistic prose tends, on the other hand, always to be rhythmically organized, or to vary the use of prose with that of verse (as in The Tale of Genji, or Boccaccio's Decameron), or finally to be set apart from non-poetic prose in its oral delivery (Chinese artistic prose, for example, is chanted). Certain forms of 'ordinary' language (especially scientific prose) are totally set on a referential function though the use of poetic figures or metaphors has never been a hindrance in the expression of 'clear and distinct ideas' (Descartes, for that matter, used more metaphors than some of his contemporary French playwrights (Weinrich 1963: 340), while other forms of discourse which pursue pragmatic functions, are usually couched in poetic form (as, for example, ritual formulas, charms, and children's verse which is used for didactic or meta-linguistic purposes, and contemporary advertisements)). There are, on the other hand, entire poetic genres, such as the proverb, which are inextricably woven into the concrete speech-act and serve as a metaphor for the given non-poetic message. Truly poetic texts, however, tend to assert their independence from concrete and practical contexts, and are recognized as such through the unity and density of their internal structure. Such works are generally marked by a maximal integration of their form and meaning, i.e., by the use of form for the structuring of meaning and by the dependence of meaning on the structured form. Such works are frequently set in the form of verse, so that

STRUCTURAL POETICS AND LINGUISTICS

641

verse itself has come to be seen as the paradigm and embodiment of the poetic principle. 4. Although poetry does not need to have recourse to any special language or to deviate from the norms of a given linguistic code, it presents 'the innate art of language intensified or sublimated' (Sapir 1921:225), and involves in a deeper sense a complete reinterpretation of the 'neutral', non-poetic use of language. This reinterpretation effects in poetry: 1) the syntagmatic character of the message (see below, 6.1.), 2) the status and participants of the speech-act (6.2.), and 3) the relations between the levels and elements of language (6.3.), producing, in effect, a new kind of code. The Russian structuralists are thus essentially right when they call this code 'a secondary modeling system', as Novalis did before them when he spoke of poetry as 'a second language', or 'Sprache in der zweiten Potenz' (1945-46: vol. 3, p. 93; cf. Stfrensen 1963:201). 5. Before discussing the transformations which language undergoes in poetry, we should note that poetry does indeed at times expand the limits of a given linguistic code by using certain features which are not otherwise encountered. These features belong to various levels of language, but are not used at random: they are either a part of older poetic traditions which are zealously preserved, or they serve to perform special poetic functions. Thus it is known that the oral delivery of Russian poetry has been (since the 17th century, when Russian acquired from Polish its syllabic verse) on a special, archaic pronunciation of the unaccented vowels (in particular o), while the 'declamatory style' has also been marked (as we know from Turgenev) by a nasalized reading of the vowels (Tomasevskij 1948:235ff.). The Yugoslav singers deliver their songs in a 'Turkish' manner that involves a shift in the quality of various vowels (M. Braun 1961:48, fn. 1). The use of archaic grammatical forms and of neologisms and foreign, 'exotic' names is extremely widespread, especially in modern poetry. Certain modern poetic schools (e.g. the Imagists) have tried to construct an entire poetics on certain parts of speech (on nouns or on verbs), while copious examples of grammatical and lexical peculiarities have been recorded in the folklore of Russia and Macedonia (Evgen'eva, Koneski). Equally common is poetic freedom in word-order for the purpose of semantic foregrounding or sharper syntactic parallelism. Folk-poetry is particularly prone to use onomatopoeia and interjections, exhortative and nonsense words (of the hey nonny nonny type). These elements frequently perform a compositional function delimiting the opening or ends of stanzas or entire songs, and complement or support the use of other, non-verbal devices (melody, clapping of hands, mimicry). The choice of language is sometimes used to distinguish entire genres or to mark poetry in opposition to prose (e.g., the choice of different dialects for different genres in Greece, the use of a superdialectal koine in Serbo-Croatian epic poetry, the special languages of the Icelandic and Irish poetic guilds).

642

EDWARD STANKIEWICZ

Written poetry presents a host of special problems, especially in verse, where the use of capital letters, the spacing of stanzas and a complex (though more or less individual) system of punctuation serves to underscore the semantic significance of words or lines, to mark the boundaries of lines and to create a tension (as in modern poetry) between the metrical lines and syntactic units. 6.1 The classical division of the arts into arts of time (e.g. poetry and music) and the arts of space (e.g. painting and sculpture) is perhaps best known from Lessing's defense of poetry in Laocoon. The main virtue which Lessing found in poetry (in opposition to the classic doctrine of ut pictura poesis) was that, being a temporal and sequential art, it had the power to grasp movement and action, and he praised Homer for having succeeded in converting 'coexistent into successive' (in his description of the shield of Hephaistos, Laocoon, = 1965 sec. 17). Aristotle's requirement that a literary work (he spoke specifically of tragedy) must have a beginning, a middle and an end applies, to all temporal, or successive arts and is signalled in poetry, and especially in verse, on all levels of its composition, beginning with the basic units (the line) and ending with the more complex ones (couplets, stanzas, parts of cycles). Poetry delimits the beginning and the end of a work, creating poetic 'closure', or a frame which contains the work and sets it off from 'ordinary' language as an artifact which has itself as its purpose and context. The frame of a text, which functions like the curtain and stage of the theatre or the frame of a painting, is generally verbal (see the remarks above), but it may exhibit peculiarities which distinguish it from the rest of the text. In Slavic epic poetry, the preamble and the close (Russ. koncovka) often dwell on themes which have little in common with the main text. The prologues and epilogues of prose fulfill similar functions. The first and final lines of poems are often identical, though the end of a poem is generally more clearly defined than its beginning, just as the end of a line is more sharply delimited than its beginning (cf. the loose character of the anacrusis as opposed to the marked form of the clausula, especially when the latter is rhymed). The end of a poem (like the couplet in the Shakespearean sonnet or the refrain) may be set off against the rest of the text by its peculiar form, and convey in addition a thematic cadence which is expected to resolve the accumulated tensions of the work (e.g., the epigrammatic endings of lyrical poems, or the proverb in Serbo-Croatian epic songs). Between the two outer boundaries of the work, the poetic message unfolds sequentially, with a forward thrust that compels us to see each part (except for the opening) as the successor of an earlier part, and as an antecedent to the following part (excluding the close), imparting to the text the qualities of anticipation and surprise. This successive movement of the poetic message is simply a corrollary of the sequential, linear character of language which de Saussure viewed as one of its basic characteristics. While the linear progression is in ordinary language endowed with an iconic function (e.g., the irreversibility of the verbs in the phrase veni, vidi,

STRUCTURAL POETICS AND LINGUISTICS

643

vici, which reflects the actual order of events), it is in poetry exploited in far more complex ways, involving not only the arrangement of the time sequence, but the structure of plot, the vicissitudes of the heroes, and the formal structure itself. The successive order is persistently felt, even when the author rearranges (by means of flashbacks) the sequential order of time, or when the causal sequence is complicated by minor plots or counterplots. The successive principle dominates in the theater, and in the classical epic which evolved, as in Homer, along a single temporal line, and is at the basis of entire literary forms, such as the catena — epiploke (e.g., Heine's Ein Jüngling liebt ein Mädchen / Die hat einen andern erwählt / Der andre liebt eine andre . . ., the Passover song Chad Gadyo, fairytales with a spiralling movement towards a certain target). The basic irreversibility of the sequence is only corroborated by the palindrome, which is a parody of the principle of sequential order. But in addition to the principle of succession, poetic language involves also the principle of simultaneity, forcing us to grasp the elements of succession in their simultaneous presence, as if they were elements in space. The artistic text is thus marked not only by an anticipatory development, but also by a retroactive impulse which integrates the elements of succession. This forward and backward movement (which is implied by the Latin term versus or German gebundene Rede for verse) applies also to the other temporal arts. Its significance in music was vividly described by Mozart, who claimed to compose his works 'in the mind' and to perceive the parts 'all at once . . . like a fine picture or a beautiful statue'.4 The arts of space, on the other hand, which are believed to be comprehended all in one glance (as Leonardo would have it, 'in un subito' — Trattato, § 21) by no means dispense with the principle of progression, but exploit it either in the composition of their works (e.g., the device of contrapposto in the Renaissance or the dramatic spirals of Baroque art), or require for their perception progression in time (as in the 'unfolding' of Chinese scrolls, of medieval panels or of modern comic strips). The different treatment of the two principles is in the respective arts only a matter of their hierarchy: poetic works (and the 'arts of time') cannot eliminate the rule of ordered progression, while visual art cannot suppress the perception of a simultaneous presence. The principle of simultaneity combined with the principle of succession makes the poetic message into an integrated structure in which each element is perceived both as an autonomous part of a sequence and a dependent part of a whole. All the parts function as elements which are subordinated to higher and more complex parts that become progressively more independent from each other, and they 4

The full passage of Mozart's letter reads: 'My object enlarges itself, becomes methodised and defined, and the whole, though it be long, stands almost complete and finished in my mind, so that I can survey it, like a fine picture or beautiful statue at a glance. N o r do I hear in my imagination the parts successively, but I hear them as it were all at once' (Letters ( = 1958), p. vii).

644

EDWARD STANKIEWICZ

are all, at the same time, coordinated as equivalent parts within the broader whole. Each progressive part thus restates, as in a moving picture, the properties of its antecedents, and presents, like a montage, the distinctive properties of the global text. This dynamic pattern of a unity in succession (Goethe's Bauer im Wechsel) is most clearly felt in the structure of verse, where each line repeats the obligatory features of the succeeding and antecedent lines, but has no existence without the other lines. The same pattern asserts itself also in artistic prose, where the occurrence of any individual segment or motive may anticipate the occurrence of a similar segment, or serve to illuminate the structure of the whole. Thus, the nail which is hammered in Act One of a play should, according to Chekhov, be the same nail on which the hero hangs himself in Act Five, just as the train in one of the early sections (Chap. 18) of Anna Karenina is the very train under which she finds death at the end of the work. The opening line of Mickiewicz' Crimean Sonnets, 'Wjechalem na suchego przestwór oceanu' ('I sailed out into the expanse of the dry ocean') — indeed, the entire first sonnet — anticipates the dynamic movement of the cycle of sonnets (which describe alternately a voyage by sea and by land) and mirrors the semantic opposition which organizes the entire composition. The plays within Hamlet are a part of the sequential unfolding of Hamlet, and function at the same time as a metaphor for the whole play. The combination of the successive and simultaneous dimensions is differently exploited in various genres and traditions which assign a different importance to the one or the other principle. Longer works with a narrative intent can hardly be expected to dwell on the simultaneity of their various components, whereas short lyrical poems exploit to the hilt the simultaneous order of all their parts. There is also a difference between works which are destined for reading and those destined for oral delivery, where the mind cannot easily retain the totality of the successive parts. This limitation of memory is compensated by the use of supplementary, integrating features such as music, and it is further, as in oral poetry, overcome by the listeners' familiarity with the text. What makes all poetry, however, an art of 'difficult reading' is precisely the fact that the principle of succession is at every step complicated and as if resisted by the principle of simultaneity, which compels attention to the structure as a whole. The difficult form of modern poetry is but a consequence of the increased emphasis on the role of simultaneity at the expense of the principle of continuity. The ideal of Valéry to combine 'le simultané de la vision avec le successif de la parole' (1957-60: I, 625) was not merely a quest for a new form, but was connected with the search for a new language of 'daring' and complex metaphors which could be grasped only within the confines of a highly condensed text. In the hands of modern poets, the traditional verse-forms took on the shape of fragments and of torsos (to mention but the works of Rilke or Cvetaeva), which depend for their comprehension as much on implied reference to other texts as on their own multidimensional implications.

STRUCTURAL POETICS AND LINGUISTICS

645

6.2 The sharp opposition between langue and parole established by de Saussure has, as we have seen, been an obstacle in recognizing the diverse functions of language in speech, and has indirectly contributed to the exploration of verbal art solely from the viewpoint of the code. It is only thanks to the recent studies of the speech-act that this confining antinomy is being overcome, and that the relations of the speech-act, code, and message appear in their proper perspective. The basic modalities of the sentence, such as question, answer, and command (i.e., the appellative function) are properties of the code which imply reference to the speechact and which establish a correlation between the participants of the speech-act and the narrated event of the message. The obligatory participation of speaker and hearer in every practical message has prompted some scholars (such as Weinrich [1966] and before him, Volosinov) to insist that the basic unit of the utterance is not the sentence, but the dialogue, and to point out that the quality and comprehension of a given message depend directly on the context of the speech-act and the relation between the participants. Through this new interest in the role of the dialogue and of the speech-situation linguistics comes to share one of the central concerns of contemporary poetics, where questions of the 'strategy of the point of view' (E. M. Forster) and 'polyphony of voices' (Baxtin) have imposed themselves especially through the study of the modern novel, though there is hardly a literary genre in which these problems can be ignored. In poetry, where the 'attention is on the message' rather than on external reference, the role of the speech-act and the relation between the participants of the speech-act are given a different significance and acquire dimensions other than in 'ordinary' language. The author, or the addresser of the poetic message is not, in the strict sense of the word, an addresser, for, despite the claims of the Romantics, the author has no privileged place within that message except as a literary 'device', i.e., as one of the voices that alternates with the voices of his characters. The reader, on the other hand, or the addressee of the poetic message, is not a real addressee, but an ubiquitous and unknown interpreter of the message, who is not expected to reply; he is merely an eavesdropper, not an interlocutor. Poetry is thus, as Valéry said, 'strange discourse, as though made by someone other than the speaker and addressed to someone other than the reader. In short, it is a language within a language' (1958: 63). The independence of the poetic message from the actual speech-act complicates its character and rearranges the relations of its grammatical categories, and especially those between the 'addresser' and 'addressee'. The non-poetic message itself varies and fluctuates according to the relation of the interlocutors to each other and to the context of the speech-act. The degree of intimacy between the interlocutors and their familiarity with the reference affect at every step the character of the message, the use of ellipsis or redundancy, the recourse to repetition or paraphrase, and the very choice of the code (or codes). The

646

EDWARD STANKIEWICZ

live speech-act is a process of incessant adjustment and accommodation which aims both at the greatest clarity and economy of the message. More formal discourse, which reduces its dependence on the context of the speech-act, obtains these goals by stipulating the precise meaning of its terms, or by coining (as in scientific language) special terms, and by restricting the field of reference to a particular subject. None of these adjustments are possible in the case of the poetic message, which is severed from the context of any speech-act, and in which the meanings of the terms refer primarily to each other, creating a structure which is intentionally multi-valent and polysemous. This independence of the poetic message from the speech-act also modifies the value of syntactic categories which 'normally' pertain to the speechact. Thus questions and commands become 'rhetorical questions' and pseudocommands, whereas the categories of tense are exploited for their oppositions within the text (e.g., in structuring the narrative or plot), without relinquishing their inherent grammatical meanings. Particularly complex is the role of 'addresser' and that of 'addressee', which split, especially in the narrative forms of fiction, into multiple addressers and addressees, speaking in a variety of styles and in intra- and inter-personal forms. Although the author and the reader are basically in the background, they may always assert their presence, as when the author addresses the audience directly or when the stage is not sharply separated from the audience (e.g., in the open stage of the Shakespearean or modern theater). Every intervention of the author can, on the other hand, be interpreted as a further development of the play, or as the voice of one more character or actor. The modern attempts by novelists to adopt the 'dramatic mode' and to reduce the role of the author have only shown in how many different guises the author may appear, delegating his role (as in the skaz) to a fictitious narrator, assuming the poses and styles of several speakers, or resorting to style indirecte libre. Only lyrical poetry, which is based on the monologue, seems to speak only in the voice of a lyrical T . This monophony, however, can be deceptive, as it is, for example, in Pasternak's poem Gamlet, where the overt monologue fuses in consecutive order the voices of Hamlet, Christ, and the poet. Between the original addresser of the text, i.e., the 'omniscient' author, and his ultimate addressee, the reader, there invervenes a series of voices which all assert their right to be heard as the diverse elements of a polyphonic structure. 6.3 One of the fundamental features of poetry is the unity of the 'inner and outer' (Goethe), the 'sensual and intellectual' (Dante), or the interdependence of sound and meaning. 'Form', write Wimsatt and Brooks, 'embraces and penetrates "message" in a way that constitutes a deeper and more substantial meaning . . . the poetic dimension is just that dramatically unified meaning which is coterminous with form' (1957: 748). The unity of meaning and form is most transparent in verse: the metrical structure of verse integrates all the elements of a poem into a unified whole, where every part contributes to the meaning of the whole, and the whole clarifies

STRUCTURAL POETICS AND LINGUISTICS

647

the meaning of the parts. Every line or stanza of a poem carries its own meaning, but leads inevitably to the meaning of another line or stanza with which it forms a higher meaning. The idea that 'poetry is made up of beautiful details' (Voltaire), and that no single part can be altered or removed without destroying the whole therefore is as pertinent as the idea that no detail has a value in and by itself, and that the meaning of a poem is only the meaning of the whole. The close interrelation between meaning and form compels us, indeed, to agree with Valéry that 'ce qui est la "forme" pour quiconque est le "fond" pour moi', i.e., that meaning and form are two complementary aspects of any poetic work. The 'symmetry between form and content' (Valéry 1958:74) or 'parallelism of expression' coupled with 'parallelism of thought' makes poetry into a code of its own kind, in which the formal elements of language, such as the selection and arrangement of sound, the length of a sentence, and word-order all contribute to the meaning of the whole and acquire particular semantic weight. Poetry explores and tests to the hilt the potentialities of language and represents the maximal utilization of the linguistic code. The interrelation between form and meaning is also apparent in the interplay between thematic structure and the use of metrical forms. Thus narrative fiction is generally less prone to be put into the form of verse than lyrical poetry, which is almost always versified. As has been repeatedly confirmed by poets, meter itself has the power to evoke the appropriate theme, and the selection of certain themes goes hand in hand with the choice of meter. The mere rhythm of the ten-syllable line (and its association with Dante's endecasillabo) triggered for Valéry the theme of his Le Cimetière marin (1958:78), while Aristotle maintains that 'Nature herself teaches us to choose the appropriate meter' (Poetics, 1460a). The use of the 'discursive' iamb in dialogues alternates in the Greek theater with the use of anapests in the appearances of the chorus, while the hexameter, or 'heroic meter' is the 'natural' meter of the Greek epic tale. The close correspondence between sound and meaning must not, however, be interpreted in an oversimplified way that imputes to sound direct reference to meaning. Poetry has no power to achieve that, i.e., it cannot suspend the arbitrary nature of the linguistic sign, despite the age-old belief of poets and mystics that that is what poetry (or language in general) is called upon to do. 'Poetry must try to raise', wrote Lessing (in a letter to Nicolai, 1769), 'its arbitrary signs to natural signs: that is how it differs from prose and becomes poetry' (Wellek 1955:164). It is indeed surprising that this idea of the evocative power of sound is still shared by some scholars, who try to reduce it (like Fónagy) to synesthetic or emotive effects, or who seek in it direct associations with meaning (Levin's theory of 'coupling'), ignoring the overall and intricate interplay of sound and meaning in a text (cf. esp. Levin 1962:37ff.). Similarly, one must not look at meter as if it were a purely euphonic device, or a 'convention' which serves 'to impart to a stretch of language a characteristic impression . . . but do[es] not by [it]self impart to a poem the sense of unity which poems produce' (Levin 1962:59). Nor is it a

648

EDWARD STANKIEWICZ

pattern which is imposed upon the poem in order to set it apart iconically from prose. Every metrical pattern is, in fact, a compositional 'device' which organizes the semantic structure of a poem by combining the principle of succession with that of simultaneity, and which forces us to perceive the various elements of a work as parts of a whole (or wholes). It permits each part to establish its independence from, and at the same time its solidarity with the other parts of a text. Every metrical pattern contains, in consequence, two interdependent components: a system of obligatory features (the so-called 'constants') which unify the entire poem (e.g., isosyllabism of lines, syntactic divisions, rhyme), and a system of more or less optional variables (or so-called 'tendencies') which put into relief the diversity of its parts (e.g. variation in the number of syllables and of accent in metrical systems built on an equal number of moras; divergence between the ictus and word-boundary in systems based on the regular alternation of stress; variation of syntactic boundaries, or enjambement in systems with a basic parallelism of lines). The chief properties of verse are, as in a nutshell, concentrated in rhyme — to wit, the combination of succession and simultaneity, similitude and difference, necessity and freedom. The major shortcoming of the traditional approach to metrics has been to treat these two components apart, as if one (the 'constants' or 'metrical schemes') were a matter of system, and the other (the 'rhythmic tendencies') a matter of individual 'deviation' or performance, an approach which is now being revived in some quarters, albeit in the fashionable terminology of 'competence" and 'performance' (cf. the pertinent remarks on this 'new' method by J. Lotz contained in this volume). The ability of the obligatory and optional components to interact with each other explains, as a matter of fact, their power for historical survival and utility, and clarifies, in turn, the preferences which different languages show (apart from cultural borrowing) in the selection and use of various syntactic and phonological features in the construction of verse: all the features of a given linguistic system are used with an eye to giving full range to the 'constants' and 'tendencies' and allowing maximal variation within the parts without sacrificing the unity of the whole. (Borrowed metrical systems modify this picture, but they are either adapted to the possibilities inherent in the given language, or they are, after some time, discarded as too confining.) The modern concern with form at the expense of meaning has distracted modern poetics from the study of the function of form, and more broadly, from the study of the very function of poetry.5 The insistence on the 'symmetry between sound and meaning', the definition of poetry as 'the emphasis on the mesage', and the inclination to see poetry as purposeless activity which can be likened to the turns of a 5

It should be noted in this context that linguistics became a proper semiotic discipline only when it discovered the multiple functions of sound in the speech chain, and not when it succeeded in separating, as Todorov believes (1968:106), the abstract phoneme from the 'crude* phonetic matter, or the variants. For a historical discussion of this problem, see Baudouin de Courtenay Anthology (— 1972), p. 21 ff.

STRUCTURAL POETICS AND LINGUISTICS

649

dance (the metaphor used by Sklovskij after Valéry, and by the latter after Malherbe) as opposed to the linear progression of prose, have tended to skirt the central issue of its function: the meaning of the poetic message as structured form. For even verse, the most pervasive and systematic of all formal devices, is not merely a structure of unfolding parallelisms of sound and meaning, but a system in which the parallelisms of form serve to create parallelisms of meaning. The function of poetry, in opposition to non-poetic language, may be said to lie in the multidimensionality of its meanings, or in what has been called the 'metaphoric process' (Foss, Wheelright). The essence of the metaphoric process is given in nuce in the metaphor, which represents the nature of poetic meaning in the same way that rhyme represents the formal properties of verse. Neither rhyme nor verse are, however, indispensable features of poetry, while the metaphoric process is essential to all its forms. The proper function of the metaphor (which was perhaps most clearly grasped by the Romantics who, however, imputed to it all kinds of transcendental connotations or, like Potebnja and the Imagists, identified it narrowly with the evocation of 'images') is the discovery of similarity in difference, and of difference in similarity, and the conjuction of these seeming contradictions into a more abstract comprehensive meaning. The metaphor contained in the above quoted first line of Mickiewicz' Crimean Sonnets (and which is developed throughout the cycle of sonnets), Wplynqlem na suchego przestwór oceanu, compels us to see the similarity and the difference between the steppe and the sea, and to view them as variants of a more abstract, unified meaning. The exploration of likeness in unlikeness, and of unlikeness in likeness, is in verbal art supported by the structure of language itself, since every meaningful unit is both opposed and similar to other meaningful units. The pattern of semantic differences and similarities encompasses all levels of language (lexical, morphological, syntactic), so that any identify or resemblance on one level (e.g., lexical) entails a difference on another level (e.g., morphological or syntactical), producing in poetry (as in language in general) a constant interplay of similarities in dissimilarities. This alternation of linguistic parallelisms can be clearly observed in those almost universal forms of folk-poetry which are built on syntactic 'parallelism', in which the syntactic resemblance of the lines is counterbalanced by lexical variation (see the interesting form of tuei ['antithesis'] in Chinese verse, cited by Liu 1967: 146). As an example of the interplay of all levels (lexical, grammatical, and syntactic) we may take the fairly simple first stanza of Blake's A Poison Tree: I was angry with my friend / I told my wrath, my wrath did end / I was angry with my foe / I told it not, my wrath did grow. In this stanza, 1) the syntactic parallellism of the odd lines is of a different kind from that of the even lines, 2) the grammatical dissimilarity in the first two rhymes (friend / end) is paralleled by that in the second two rhymes (foe / grow), and 3) the rhyming pairs are lexically opposed as anto-

650

EDWARD STANKIEWICZ

nyms but similar in their semantic range (friend /foe, end /grow). The metaphoric process engages the poet in an incessant exploration of the resources of his language, but the choices of what is similar and dissimilar are his own, and it is they that make up the 'vision' of his poetry. But he further transcends the limits of language by creating entire works whose themes, plots, actions, and heroes are constructed on the same principle of similarity in difference. The application of this principle to the structure of narratives is again most tangible in folklore (and in the literature based on it), since folklore reduces to the simplest schemes the relations which in original (and especially in great) literature tend to become highly complicated and intertwined. These narratives include such universal topics as the fight of the hero and the dragon, the return of a hero after many years of exile, the transformations of man into animal (like in the Metamorphoses of Ovid), the strange yet familiar worlds of giants and dwarfs (as in Gulliver's Travels), the blindness of the seeing and the seeing of the blind (as in Sophocles' Oedipus Rex). All these 'stories' contain the elements of wonder and surprise which come with the recognition that the split realities which appear before us are basically the same, and that the conflicting diversities conceal a basic unity. The forcing of oppositions into a single focus, or the interpretation of invariants as variants of a more abstract invariant, lend the metaphoric process the qualities of creativity and discovery which defy the claim that poetry has 'nothing to tell' (unless by 'telling' we mean only the acquisition of new facts) in that it refers only to itself. The interplay and conjunction of parallelisms, furthermore, is not a static process, as may appear from the contemplation of single metaphors, the parlor game of traditional Rhetoric. The metaphoric process unfolds with the work as a whole, so that resemblances established in one part of the work are transformed into differences in another part, creating shifting perspectives which compete for a unified vision. This dynamic character of the metaphoric process was most cogently defined by G. M. Hopkins: '[poetry] makes of each resemblance a reason for surprise in the next difference, and of each difference a reason for surprise in the next resemblance . . . and resemblances and antitheses themselves are made to make up a wider difference' (1959:105). The constant openness to new differences and syntheses ('Aufhebung') goes beyond the text itself, for the reader is always free, and indeed prompted to look for further implications, as in the riddle, whose structure suggests that the question posed in the comparison allows for more than one or two seemingly obvious answers. The poetic text is thus the most dynamic and open-ended type of message, and not a hermetically closed and self-referential structure. Nor can it be reduced to a fixed number of interpretations, as it was in the exegetic tradition of the Middle Ages, with its four methods of literary criticism (Colish 1968:256). On the contrary, it can always be interpreted, metonymically, as a part of a larger text (as a

STRUCTURAL POETICS A N D LINGUISTICS

651

poem in a cycle, or as a specific work of a certain school), or, metaphorically, as an echo of another text (like a parody with relation to its prototype, or Joyce's Ulysses with respect to Homer's Odyssey). The elements of a poetic text combine, as a result, both a centripetal and centrifugal force: they pull towards each other to form a unified text, and lead invariably outside to be enlarged and comprehended in the context of other texts. 6.4 Thus we may conclude that poetry 1) restructures the character of the message by superimposing simultaneity upon succession; 2) reinterprets the nature of the speech-act by detaching it from the speech-situation and by altering the relation between the 'interlocutors', and 3) modifies the relations within the linguistic code by interlocking the formal and semantic levels and by integrating the inherent meanings of the linguistic units with the meanings generated in the poetic message into a more complex and comprehensive meaning. 7. In choosing the terms of the metaphoric process which organize a given literary text, the poet is not entirely free: he must put them into the mold of one of the available types or genres which make up the literary conventions of his time. Although the types are to some extent isomorphic and translatable into each other (e.g., a novel can be made into a play, and a play can be transformed into a poem), each type is governed by its own 'laws', which define the text semantically and formally. In creating a poetic text, the poet is thus bound, as it were, by two conflicting requirements: he must create a message that says something new (if it is not to be a replica of another text), and must adhere to the norms of the prevailing poetic code. The tension between these two requirements may vary according to the cultural standards of the times, as well as according to the type of text. Popular art and kitsch which aim at immediate recognition tend to adhere maximally to the norms, whereas original literature may depart considerably from the literary code and create standards for a new code (or codes). Total originality is impossible, as is would lead to total incomprehensibility, which could hardly be the aim of any communicable message. Only modern poetry, as we have seen, has elevated this aim to an esthetic ideal (e.g., the attempts of Symbolists and of some contemporary poets to create a 'silent, white poem' [Mallarmé] to which the reader may ascribe any meaning he wants [Yeats, Valéry]). What is more curious is that this ideal should be embraced also by modern critics who, beginning with Croce down to the French Structuralists, proclaim (in theory, at least) the uniqueness and nonaccessibility of the literary text and who would thereby seem to remove the ground on which they stand (in practice, these scholars indulge a great deal in the study of texts). The analysis of literary types, or a typology of literature, is one of the foremost tasks of a structural poetics, although attempts to classify the various forms of literature are probably as old as literature itself (to wit, the classificatory schemes

652

EDWARD STANKIEWICZ

proferred by Plato and Aristotle). In our own century such attempts have to some extent fallen into disrepute, partly because of the greater fluidity between the genres and partly because of the attacks launched against them by students who believe in the sanctity of the text (especially Croce) and reject on principle the study of systems and 'laws'. Classification is, however, the necessary first step in any scientific description, and a typology of literary forms can be extremely fruitful if it does not restrict itself to a mere enumeration of certain haphazardly chosen features. Recent classifications of genres suffer, like some of the older attempts, either from an excessive schematism, which reduces the structure of genres to isolated linguistic or psychological elements (such as the difference in the use of the categories of person or time), or from the tendency to impute to the genres an immutable existence of 'virtualities' given for all time (like Goethe's Urphdnomena). The goal of typology of literary forms is, however, like that of linguistic typology, far more comprehensive, and ideally it should attempt to reveal the invariant properties of literary forms, the hierarchy of genres, and the possibilities of occurrence and nonoccurrence of certain forms, for as Wolfflin reminded us, 'not everything is possible at the same time' (1929:xi). Such typologies must not, furthermore, be conceived, as until now, as inventories of individual genres, since all the genres of a given period form an integrated system in which the qualities of each genre are defined by their relation to those of other co-existing genres, i.e., by their positive as well as negative features (the so-called 'negative devices' which were ignored by the Formalists but are now emphasized by the Russian Structuralists). The empirical investigation of the literary genres which are found in various places and times can considerably broaden our conception of the genres and of their division which have been transmitted to us by traditional Poetics. Boas, who studied the forms of oral poetry used by the American Indians, arrived at the conclusion that 'the two fundamental forms of literature, song and tale, are found universally and must be considered the primary forms of literary activity' (1955: 301). Other forms are less widely spread and appear to be variants of these two fundamental types. According to Boas, epic poetry is found only in Europe and in Central Asia, whereas the riddle and proverb are practically unknown in America, but highly popular in Africa, where even court decisions are rendered in the form of proverbs (1955:338). The epic and tragedy are almost absent in Chinese poetry, which, according to Liu (1967:152), also makes scarce use of narrative verse. Similar restrictions in the use of or in the combination of various forms are well known from the history of European literatures. Boas' observations compel us to reconsider the 'ultimate status' of the celebrated three genres, the epic, lyric, and drama, which to Goethe represented immutable 'natural forms', and which are, in turn, upheld by existentialist or phenomenological explanations (see Wellek's critique of Staiger in 1971:225ff.). The universality of the prose narrative and the song (or versified 'lyrical' form) permits us to view them as the polar types of verbal art, with the former being the

STRUCTURAL POETICS AND LINGUISTICS

653

semantically 'marked' term of the opposition, and the latter the 'unmarked' term, since the narrative always has a 'story' to tell ('Yes — oh, dear, yes —', E. M. Forster says whimsically, 'the novel tells a story' [1927:26]), which involves the progress of the action, its heroes, and the point of view of the author, whereas the 'song' or verse has no definite thematic commitments. This difference entails a number of consequences: in the narrative, the metaphoric process is built around actions, plots, the relation between the narrator and heroes, and propels the narrative along the axis of succession, whereas the 'song' builds its metaphoric tensions around the most heterogeneous semantic material and attains its unity by emphasizing the principle of simultaneity. Consequently the narrative tends to avoid versified form (or at least to avoid complex stanzaic forms), whereas the 'song' cannot do without a tight metrical organization, and even the support of musical elements, which integrate its content. The drama, which is a less universal form (or a relative newcomer, as in classical Greek poetry) can, in turn, be viewed only as a variant of the narrative: it shares with the narrative proper the elements of action and heroes, but replaces the voice of the narrator with that of the actors and the verbal presentation of a part of the action with its visual presentation on the stage. Like the 'narrative', drama further emphasizes the principle of progression, avoiding the use of 'stalling' metaphors and of complex metric form. Between these basic types of verbal art there is a whole gamut of intermediary forms in which the relation of the elements is constantly in flux, competing for the performance of the same basic functions in a large variety of different forms. A scientific typology of literary forms must also pay due attention to the fundamental difference between oral and written poetry. The access of the modern reader to a vast variety of texts and historically different types enlarges the context within which he must interpret any literal text, and introduces complications, due to historical depth, that do not arise in the sphere of oral poetry, which is essentially synchronic. The other basic characteristic of oral poetry (which is connected with its synchronic nature) is that it exhibits maximal variation of the individual text and minimal variation of the collective code. The 'singer' recites basically what everybody knows, but has never heard before. Every text appears thus to be another variant of the same type, and every performer becomes the author of a new text. This uniformity and basic stability of oral literature has made it the favorite ground for the study of its 'laws', and for the modern scientific attempts to define its invariant properties (Olrik, Jolles, Skaftymov, Propp). Unfortunately, some of these attempts are either one-sided and purely formal (e.g., Olrik and Propp), or obscured with unconvincing metaphysical speculations, which are inherited from the Romantics (Jolles). Written poetry, on the other hand, has an inherent tendency to codify (to 'fix') the text and, unlike oral poetry, to transcend the limits of the prevailing synchronic norms. Furthermore, the written text and especially the printed book have made

654

EDWARD STANKIEWICZ

available to the reader entire traditions of literature which, in the words of Eliot, constitute 'a simultaneous existence and compose a simultaneous order' (1932:4). These various traditions with their various literary norms may all contribute and compete for the proper interpretation of a text. The modern poet, as we have seen, has been the first to renounce his authority over his text by declaring (like Valéry) that 'once published, a text is like an apparatus that anyone may use as he will according to his ability' (Valéry 1958:152). This drift into relativism, which denies the text any fixed meaning can, however, hardly be corrected by the opposite view according to which the text has an 'immanent' and immutable meaning which exists independently of other texts and norms. Nor can one hope to rescue 'scientific' objectivism by claiming, like Milic, that 'successful reading' is primarily a scholarly enterprise in which 'the reader decodes a text by recovering the author's code' (Chatman 1971:23). No such recovery is, of course, necessary for the normal enjoyment of a work, nor can the reader be expected to isolate such a code without the intereference from the dominant code (or codes) of his own time. The simultaneous existence of historically different codes creates in literature a situation similar to the metaphoric tension which arises in the poetic text itself: the different codes of the past are interpreted and enriched by the knowledge of the dominant code(s) of the present, whereas the code of the present is deepened and expanded by the knowledge of the codes of the past.

REFERENCES

1958. Poetics, transi. G. M. Brube. New York. Essais critiques. Paris. BAUDOUIN DE COURTENAY, JAN. 1 9 7 2 . Jan Baudouin de Courtenay anthology, ed. and trans. E. Stankiewicz. Bloomington, Ind. BAXTIN, MIXAIL. 1 9 2 9 . Problemy tvorcestva Dostoevskogo. Leningrad. Partial English translation, Discourse typology in prose, in L. Matejka and K. Pomorska, eds. ( q . v . ) , 1 9 7 1 , 1 7 6 - 9 6 . BECKMANN, MAX. 1 9 7 0 . Lecture delivered at the New Burlington Galleries in 1 9 3 8 , cited by H. Osborne, Aesthetics and art theory: An historical introduction. New York. BERKELEY, GEORGE. 1 9 6 5 . The principles of human knowledge, Berkeley's philosophical writings, ed. D. M. Armstrong. New York. BOAS, FRANZ. 1 9 5 5 . Primitive art. New York. BOLINGER, DWIGHT. 1 9 6 1 . Generality, gradience, and the all-or-none. JanL, series minor 14. BRAUN, MAXIMILIAN. 1 9 6 1 . Das serbokroatische Heldenlied. Göttingen. BRIK, OSIP. 1 9 1 9 . Zvukovye povtory. Poètika: Sborniki po teorii poèticeskogo jazyka, 5 8 - 9 8 . Petrograd. ARISTOTLE.

BARTHES, R . 1 9 6 4 .

STRUCTURAL POETICS AND LINGUISTICS

655

K. 1934. Sprachtheorie. Die Darstellungsfunktion der Sprache, Jena. 19652, Stuttgart. CARNAP, RUDOLPH. 1935. Philosophy and logical syntax. London. CERVANTES, S. M. DE. 1965. Don Quixote, ed. Mateu. Barcelona. CHATMAN, SEYMOUR, ed. 1 9 7 1 . Literary styles: A symposium. London-New York. CHATMAN, SEYMOUR, and S. LEVIN, eds. 1 9 6 7 . Essays on the language of literature. Boston. COHEN, JEAN. 1 9 6 6 . Structure du langage poétique. Paris. COLERIDGE, SAMUEL TAYLOR. 1 8 1 7 . Bibliographia literaria. Repr. in Wordsworth and Coleridge, Lyrical ballads, 1805, ed. Derek Roper. London and Glasgow BÜHLER,

1968.

The mirror of language: A study in the medieval theory of knowledge. New Haven. Eco, UMBERTO. 1962. Opera aperta: Forma e indeterminazione nelle poetiche contemporanee. Milan. . 1968. La struttura assente. Milan. . 1972. Einführung in die Semiotik. Munich. ÈJXENBAUM, B . M . 1 9 1 9 . Kak sdelana 'Sinei' Gogolja. Poètika: Sbornik po teorii poèticeskogo jazyka, 1 5 1 - 6 5 . Petrograd. . 1922. Melodika russkogo liriceskogo stixa. Petrograd. . 1924. Hljuzija skaza. Skvoz' literaturu, 152-56. Leningrad. . 1927. Leskov i sovremennaja proza. Literatura. Teorija. Kritika. Polemika, 210-25. Leningrad. ELIOT, T. S. 1932. Tradition and the individual talent. Selected essays. New York. ERLICH, VICTOR. 1 9 6 9 . Russian Formalism — History and doctrine. The Hague. EVGEN'EVA, A. P. 1963. Ocerki po jazyku russkoj ustnoj poèzii v zapisjax XVII-XX w . Moscow. FÓNAGY, IVAN. 1 9 6 4 . L'information du style verbal. Linguistics 4 . 3 4 . . 1971. The functions of vocal style. In S. Chatman, ed. (q.v.), 1971, 159-76. FORSTER, E. M. 1927. Aspects of the novel. London. FOSS, M. 1949. Symbol and metaphor in human experience. New York. FRIEDEMANN, KÄTE. 1910. Die Rolle des Erzählers in der Epik. Leipzig. GANTNER, JOSEPH. 1 9 5 4 . Lionardo de Vinci. Discours universitaires. Saarbrücken. GENETTE, GÉRARD. 1966. Structuralisme et critique littéraire. L'Arc, 26. GOMBRICH, E. H. 1960. Art and illusion. Princeton. GUIRAUD, PIERRE. 1 9 6 7 . Linguistique et critique littéraire. Cours d'été et colloques scientifiques, Université du Bucarest, Sinaia, 25 juillet — 25 août. . 1971. Immanence and transitivity of stylistic criteria, In S. Chatman, ed. (q.v.), 1971, 16-23. HAMBURGER, KÄTE. 1957. Logik der Dichtung. Stuttgart. HOPKINS, GERARD MANLEY. 1 9 5 9 . Poetic diction and On the origin of beauty. The Journals and Papers, ed. H. House, 8 4 - 1 1 4 . London. COLISH, MARCIA L . 1 9 6 8 .

656

EDWARD STANKIEWICZ

1921. Novejsaja russkaja poèzija. Nabrosok pervyj. Xlebnikov. Prague. . 1923. O cesskom stixe, preimuscestvenno v sopostavlenii s russkim. Berlin. . 1953. The kernel of comparative Slavic literature. Harvard Slavic Studies I, 1-81. Cambridge, Mass. . 1960. Linguistics and poetics. Style in language, ed. T. A. Sebeok, 350-77. Boston and New York. . 1961. Poèzija grammatiki i grammatika poèzii. Poetics — Poetyka — Poètika, 397-417. Warsaw. JAKUBINSKIJ, L. D. 1919. O zvukax stixotvornogo jazyka. Poètika: Sbornik po teorii poèticeskogo jazyka, 37-49. Petrograd. JAMES, HENRY. 1884. The art of fiction. London. JAMESON, FREDRIC. 1972. The prison house of language. A critical account of structuralism and Russian Formalism. Princeton. JOLLES, ANDRÉ. 1 9 2 9 . Einfache Formen. Halle. Darmstadt. 1 9 5 8 3 . KANDINSKY, WASSILY. 1912. Über das Geistige in der Kunst. Munich. KATZ, J. 1964. Semi-sentences. The philosophy of language, ed. by J. A. Fodor and J. J. Katz, pp. 400-16. Englewood Cliffs, N.J. KONESKI, BLAÌE. 1971. Jazikot na makedonskata narodna poezija. Skopje. LANGER, SUSANNE. 1 9 6 7 . Mind: An essay on human feeling. New York. LAZICZIUS, G Y . 1936. A new category in phonology. Proceedings of the Second International Congress of Phonetic Sciences, 57ff. Ghent. LEMON, L . , and M. REIS, eds. and transls. 1965. Russian Formalist criticism: Four essays. Lincoln, Neb. LEONARDO DA VINCI. 1890. Trattato della pittura, ed. G. Milanesi. Rome. LESSING, GOTTHOLD EPHRAIM. 1 9 6 5 . Laocoon oder über die Grenzen der Malerei und Poesie, ed. D. Reich. London. 2 LEVIN, SAMUEL. 1964 . Linguistic structures in poetry. JanL, series minor 23. Liu, JAMES J . Y. 1967. The art of Chinese poetry. Chicago. LOTMAN, JURIJ. 1970. Struktura xudozestvennogo teksta. Moscow. German transl. by Rolf-Dietrich, Die Struktur literarischer Texte. Munich, 1972. LUBBOCK, PERCY. 1929. The craft of fiction. New York. MATEJKA, L. and K. POMORSKA, eds. 1971. Readings in Russian poetics: Formalist and structuralist views. Cambridge, Mass. MESCHONNIC, HENRI. 1970. Pour la poétique. Paris. MILIC, LOUIS T. 1971. Rhetorical choice and stylistic option: The conscious and unconscious poles. In S. Chatman, ed. (q.v.), 1971, 77-94. MOZART, WOLFGANG AMADEUS. 1 9 6 8 . Letters, ed. H . Mersmann. New York. MUKA&OVSKY, JAN. 1 9 3 8 . La dénomination poétique et la fonction esthétique de la langue. Actes du Quatrième Congrès International de Linguistes, 9 8 - 1 0 4 . Copenhagen. JAKOBSON, ROMAN.

STRUCTURAL POETICS AND LINGUISTICS

657

. 1967. Standard language and poetic language. In S. Chatman and S. Levin, eds. (q.v.), 1967. NOVALIS. 1 9 4 5 - 4 6 . Gesammelte Werke, ed. Carl Seeling. Vols. 1 - 5 . OLRIK, AXEL. 1 9 6 9 . Epische Gesetze der Volksdichtung, Zeitschrift für das Altertum und deutsche Literatur, 2.2ff. Berlin. 3 PIAGET, JEAN. 1968 . Le structuralisme linguistique. Le Structuralisme, 6 3 - 8 1 . Paris. POLIVANOV, E . D. 1919. Po povodu 'zvukovyx zestov' japonskogo jazyka, Poètika: Sbornik po teorii poèticeskogo jazyka, 27-36. Petrograd. French transi, in La poétique, Seuil, 6, 1966. POMORSKA, K . 1 9 6 8 . Russian Formalist theory and its poetic ambiance. The Hague. PROPF, V. I . 1 9 2 8 . O morfologii skazki. Leningrad. English trans., Morphology of the folktale. Bloomington, 1958. R E I D , LOUIS ARNAUD. 1 9 6 8 . New notes on Langer. The British Journal of Aesthetics 8/4, Oct. RICHARDS, I. A. 1926. Science and poetry. New York. RICOUER, PAUL. 1963. Structure et herméneutique. Esprit 11, Nov. RIFFATERRE, M. 1959. Criteria for style analysis. Word 15/1.154-74. . 1960. Stylistic context. Word 16/2.207-18. . 1961. Vers la définition linguistique du style. Word 17/3.318-44. Ruiiöic, GOJKO. 1934. Narodna epika u Velesu. Prilozi proucavanju narodne poezije 1, 2. SAPIR, E . 1921. Language. New York. DE SAUSSURE, F. 1968. Cours de linguistique générale, ed. Rudolf Engler. Wiesbaden. SEARLE, JOHN. 1 9 6 5 . Speech acts. Cambridge, Mass. . 1972. Chomsky's revolution in linguistics. New York Review of Books XVIII/12, June 29, 16-24. SKAFTYMOV, A. P. 1924. Poètika i genezis bylin. Moscow-Saratov. SKLOVSKIJ, VIKTOR. 1914. Voskresenie slova. St. Petersburg. Repr., with German transi., in W.-D. Stempel, ed. {q.v.), 1972, 2-17. . 1919. Iskusstvo kak priëm. Poètika: Sbornik po teorii poèticeskogo jazyka, 101-14. Petrograd. English transi, in L. Lemon and M. Reis, eds. (q.v.), 1965, 3-24. . 1925. O teorii prozy. Moscow. S0RENSEN, BENGT ALGOT. 1 9 6 3 . Symbol und Symbolismus in den ästestischen Theorien des 18. Jahrhunderts und der deutschen Romantik. Copenhagen. STAIGER, EMIL. 1 9 6 1 5 . Grundbegriffe der Poetik. Zurich. STEMPEL, WOLF-DIETER, ed. 1972. Texte der russischen Formalisten, II. Munich. STRIEDTER, JURIJ, ed. 1 9 6 9 . Texte der russischen Formalisten, I . Munich. Theses of the Prague Circle — Thèses présentées au Premier Congrès des philologues slaves. 1929. TCLP 1.5-29.

EDWARD STANKIEWICZ

658 THOMPSON, E W A .

1971. Russian Formalism and Anglo-American New Criticism.

The Hague. TODOROV, TZVETAN.

1965a. Les poètes devant le 'bon usage'. Revue d'esthétique

3/4, 301-05. . 1965b. L'héritage méthodologique du Formalisme. L'Homme 5/1.64-83. . 1968. Poétique. Qu'est-ce-que le structuralisme, ed. O. Ducrot, T. Todorov, D. Sperber, et al., 99-166. Paris. TOMASEVSKIJ, B. V. 1929. O stixe, Stat'i. Leningrad. . 1948. K istorii russkoj rifmy. Trudy Otdela novoj russkoj literatury, 233-80. TRUBETZKOY, N. S. 1949. Principes de phonologie. Paris. TYNJANOV, JURIJ. 1924. Problema stixotvornogo jazyka. Leningrad. Reprinted, The Hague, 1963. Partial English transi, in L. Matejka and K. Pomorska, eds. ( I . V . ) , 1971, 126-45. . 1929. Literaturnyj fakt. Arxaisty i novatory, 5-29. Leningrad. TYNJANOV, JURIJ, and R . JAKOBSON. 1928. Problemy izucenija literatury i jazyka. Novyj Lef 12.36-37. English transi, in L. Matejka and K. Pomorska, eds. (q.v.), 1971, 79-81. VALÉRY, PAUL. 1 9 5 7 - 6 0 . Oeuvres, ed. Jean Hytier. Bibliothèque de la Pléiade, Paris. . 1958. The art of poetry, transi. Denise Folliot. New York. VINOGRADOV, V . V . 1926. Problema skaza v stilistike. Poètika: Sbornik statej (Vremennik Otdela Slovesnyx Iskusstv), 1.124-40. Leningrad. Reprinted, The Hague, 1966. . 1927. K postroeniju teorii poèticeskogo ja2yka. Ucenie o sistemax reci literaturnyx proixvedenij. Poètika (Vremennik Otdela Slovesnyx Iskusstv) 3.524. Leningrad. . 1930. O xudozestvennoj proze. Moscow-Leningrad. VOLOSINOV, V . N. 1930. Marksizm i filosofija jazyka. Leningrad. Partial English transi., Reported speech, in L. Matejka and K. Pomorska, eds. (q.v.), 1971, 149-75. Reprinted, The Hague, 1972. WEINRICH, HARALD. 1 9 6 3 . Semantik der kühnen Metapher. Deutsche Vierteljahrschrift 3 7 . 3 2 5 - 4 5 . . 1964. Tempus: Besprochene und erzählte Welt. Stuttgart. . 1966. Linguistik der Lüge. Heidelberg. . 1971. Literatur für Leser. Essays und Aufsätze zur Literaturwissenschaft. Stuttgart-Berlin-Cologne-Mainz. WELLEK, R E N É . 1 9 5 5 . A history of modern criticism: 1 7 5 0 - 1 9 5 0 , I . New HavenLondon. . 1971. Genre theory, the lyric, and 'Erlebnis'. Discriminations: Further concepts of criticism. New Haven. WHEELWRIGHT, PHILIP. 1962. Metaphor and reality. Bloomington-London.

STRUCTURAL POETICS AND LINGUISTICS

659

. 1958. The burning fountain: A study in the language of symbolism. [Esp. Ch. 6.] Bloomington-London. WIMSATT, WILLIAM K., JR., and CLEANTH BROOKS. 1957. Literary criticism: A short history. New York. WOLFFLIN, HEINRICH. 1 9 2 9 7 . Kunstgeschichtliche Grundbegriffe. Munich. ZIRMUNSKU, VIKTOR. 1 9 2 3 . Rifma, ee istorija i teorija. Petrograd. . 1928. Voprosy teorii literatury. Leningrad. . 1966. Introduction to metrics: The theory of verse, ed. E. Stankiewicz and W. N. Vickery. The Hague.

LINGUISTICS A N D FOLKLORISTICS

WILLIAM O. HENDRICKS

Association between the disciplines of linguistics and folkloristics has been meager and largely one-sided.1 Even in the nineteenth century, when linguistic and folkloristic study were pursued in several cases by the same scholars, linguistic methodology provided the model for folkloristics.2 For instance, the so-called Finnish method, which entails the study of the geographical distribution of variants of tales and the attempt to reconstruct the archetypal ('proto') form, bears obvious similarities with the comparative method as developed in linguistics. (However, one eminent practitioner of this method has cautioned against pushing the linguistic analogy too far; see Thompson 1953:270.) As Sebeok (1953:378) has pointed out, 'linguistic preoccupation with Proto-Indo-European reconstructions led the brothers Grimm to assume tales to be a detritus of early Indo-European myths... The early decades of the twentieth century marked a growing divergence of the two disciplines. In the early part of the century Boas and his followers collected and published a large number of full texts of American Indian tales. Not only were these texts valued as linguistic documents, they were also subjected to a type of folkloristic analysis. However, with the death of Boas, enthusiasm for such collections waned (cf. Fischer 1963:238). But insofar as linguists find folktales and myths useful as a means for collecting material for linguistic analysis, a marginal association between linguistics and folkloristics continues to exist. Beginning in the '20s and '30s a new intellectual current — the structural approach — began to have impact on various disciplines, including linguistics. It arose in opposition to nineteenth century historicism, with its positivistic emphasis on the cataloguing of atomistic facts. As Dundes (1964:37) has noted, however, 'most folklorists remained oblivious of the new worlds about them'. Jacobs (1966: 417), in a somewhat stronger vein, has asserted that folkloristics even to this day 1

The term folklore can refer to either the discipline or the materials studied by discipline. In this paper the term will refer only to the materials studied; folkloristics be used to denote the discipline. Although folklore may be verbal or non-verbal, only former will be considered here — what some refer to as 'verbal art*. 1 See Sebeok (1953), Dundes (1964: 38). However, as Jespersen (1922 :41) has pointed Jacob Grimm's interest in language was initially subordinate to his interest in folklore. further historical discussion, see von Sydow (1948).

the will the out, For

662

WILLIAM O. HENDRICKS

'has not changed substantially since the 1890s because its specialists are loath to search for more revealing aims and better methodological tools'. Although it is true that folkloristics largely remains committed to a historical approach, it is nevertheless the case that Jacobs overstates the situation when he asserts that the structural approach has 'stirred only minor anxieties among professional folklorists' (ibid.). More balanced (or certainly more optimistic) views are presented by Pop (1968) and by Voigt (1969). They discern a growing interest in the structural approach among folklorists — though Voigt (249) concedes that a fully structural folkloristics remains largely a task for the future. The current status of the structural approach to folklore can perhaps best be ascertained from Meletinskij's (1969) extensive survey of the field.3 As is apparent from the survey, the structural approach began to take root only after the 1958 English translation of Propp's Morphology of the folktale. The reason for this, according to Meletinskij (1969:6-7), was the success of structural linguistics and anthropology. The structural approach in folkloristics thus represents a renewal of the association between linguistics and folkloristics — but again it is largely a one-sided relationship. Pop (1968:1) asserts that 'Wie früher, so greifen auch die gegenwärtigen Versuche um eine Erneuerung der folkloristischen Methodologie zahlreiche Elemente aus den von der Sprachforschung in dieser Hinsicht schon erzielten Resultaten auf'. Voigt (1968:249) has similarly noted that as yet there is no independent methodology of folklore structuralism and that 'in the United States most of the structural folklore conceptions are based directly on linguistic analogies'.4 Despite the appeal to linguistics on the part of several folklorists, there often exists simultaneously an anti-linguistic bias, in the form of a sharp differentiation between 'linguistic' and 'folkloristic' structure. Dundes (1964), for instance, has drawn this distinction on the basis of the fact that the structure of, say, a riddle remains unchanged when it is translated into another language, whereas the linguistic features will more than likely undergo change. That is, the structure of a folkloristic work or genre is asserted to be independent of any given language. Dundes (1964:49) does state, 'of course, it is extremely interesting to study both the folkloristic and the linguistic structural aspects of texts, especially to see if the linguistic stylistic features reinforce the folkloristic structure, e.g. rhyme, assonance, or metrical features may parallel and even underscore folkloristic structural unit divisions. However, it is important to remember that folklore structure can be analyzed WITHOUT reference to a particular language.' As is obvious from these remarks, Dundes thinks of linguistic features solely in terms of sub-sentence units. Likewise, Fischer (1963:249) has stated that 'the 3

This is a German translation of the afterword Meletinskij contributed to the new Russian edition of Propp's Morphology of the folktale. Another recent survey is that of Nathhorst (1969); it is marred, however, by an excessively negative attitude. Drobin (1969) attempts to put Nathhorst's critique into perspective. 4 The leading American folklorist in this regard is Alan Dundes, who has adapted Pike's notions of the etic and the emic to folktale analysis; see especially Dundes (1962, 1964).

LINGUISTICS AND FOLKLORISTICS

663

order of words within the sentence is primarily a matter of linguistic syntax, and has nothing to do directly or necessarily with the tale structure proper'. What these scholars fail to appreciate is the possibility for extending linguistics so as to be able to account for relations beyond the sentence. They share this failure with most linguists, who assume that linguistic structure does not exist beyond the sentence boundary. But it is a thesis of this paper that structuralism in folkloristics represents an attempt to achieve something comparable to a 'syntax' (and 'semantics') of discourse. 'Folkloristic' structure may well be that aspect of linguistic structure that pertains to supra-sentential relations. Although linguistics has been primarily a donor to folMoristics, the position taken in this paper is that folkloristics potentially has much to offer linguistics, particularly in the area of discourse analysis. The major research problem for the linguist is exactly how text ('folkloristic') structure might be integrated into linguistic structure. This is an extremely complex subject, and only certain tentative suggestions can be presented in this paper.5 More general, but related, issues will receive primary attention here. While the focus will be on the contribution of folkloristics to linguistics, some tentative proposals will be offered as to the value of certain linguistic notions for the advancement and development of folkloristic theory. Given the present situation in linguistics, the linguist interested in integrating the work of the folklorists back into linguistics must come to terms with one important issue. To what extent is structuralism in folkloristics vulnerable to the same criticisms Chomsky has made of the structural school of linguistics? According to Chomsky (1968:16-17), modern structural linguistics, in departure from the tradition of philosophical grammar, 'restricts itself to the analysis of what I have called surface structure, to formal properties that are explicit in the signal and to phrases and units that can be determined from the signal by techniques of segmentation and classification. This restriction is a perfectly self-conscious one, and it was regarded — I believe quite erroneously — as a great advance.' In the heyday of linguistic structuralism this euphoric feeling of accomplishment was shared by other disciplines, which looked with awe upon linguistics. As early as 1945, Lévi-Strauss wrote that 'linguistics occupies a special place among the 5

For elaboration, see Hendricks (1967, MS). Incidentally, the possibility of integrating text structure into linguistic structure has been vaguely foreseen as a possibility by Fischer — despite the fact that he too has differentiated between the 'gross structure' ( = Dundes' 'folkloristic structured and the linguistic structure of folktales on the basis of the fact that the former does not change with translation (see Fischer 1960). As Dundes (1964:45) has acutely noted, 'Fischer's equation of 'gross structure' with 'the relations among sentences and sentence groups' indicates that he has not completely escaped the linguistic bias'. In a later article Fischer (1963 :248-9) is somewhat more explicit on this point: 'We may, if we wish, regard the structure of the tale as consisting of more inclusive levels - i.e. above the level of the sentence - imposed on linguistic structure. If we regard tale structure as linguistic, however, we should be clear that we are dealing not with linguistic form, but with linguistic meaning.'

664

WILLIAM O. HENDRICKS

social sciences.... It is not merely a social science like the others, but, rather, the one in which by far the greatest progress has been made. It is probably the only one which can truly claim to be a science and which has achieved both the formulation of an empirical method and an understanding of the nature of the data submitted to its analysis. This privileged position carries with it several obligations. The linguist will often find scientists from related but different disciplines drawing inspiration from his example and trying to follow his lead' (Lévi-Strauss 1963:29). What particularly impressed scholars in other disciplines was the linguist's ability to segment his material into discrete units. The folklorist was no exception. Dundes, writing in the early '60s, stated that 'a linguistics model proves very suggestive for folkloristic analysis, at least with respect to units' (Dundes 1964:55; cf. also Dundes 1962). Jacobs has made similar assertions, including as well emphasis on the importance of fieldwork (1966:421): 'No language can be easily if at all analyzed after return from a field situation. . . . An oral literature can be probed, its significant parts identified, and classes of them set up, . . . only in the continued presence of informed, articulate, and alertly cooperative natives.' Another similarity between structuralism in linguistics and in folkloristics is concern for finite corpora of data, rather than productive rules (cf. Hymes 1968: 175-6, Sperber 1968 :208). This, as Hymes notes, is a consequence of defining the object of folkloristics as traditional materials — as a collection of things rather than as a process. (This issue will be touched upon later.) The preceding considerations seem to suggest rather close parallels with just those aspects of structural linguistics that Chomsky finds most objectionable. However, Jacobs (1963 :278) has noted, 'it is curious that although some writers have invoked a linguistic analogy of correct method in oral literature research, it is not really employed' (cf. also Colby 1963:275). Jacobs implies that this is regrettable, but in one sense it is advantageous in that structural analysis in folklore may thus be exempt from the valid criticisms of structuralism made by Chomsky. Insofar as folklorists differentiate between folkloristic and linguistic structure, that is, insofar as they are not concerned with the actual linguistic surface of tales but rather with the images, events, etc. conveyed by the language, then to that extent they cannot literally follow structuralist procedures of segmentation and classification (cf. in this regard Fischer 1963:289). In any case, it would be a mistake to identify structuralism in linguistics too specifically with a concern for segmentation and classification of the speech signal. At its most general level, linguistic structuralism may be characterized in negative terms as an opposition to the diachronic analysis of isolated phenomena. More positively, the structural 'revolution' shifted emphasis from conscious linguistic phenomena to their unconscious basis; it focused, not on terms as independent entities, but rather on the relations between terms; it introduced the notion of system; it stressed the uncovering of general laws. The preceding characteristics are cited by Lévi-Strauss (1963:31), following

LINGUISTICS AND FOLKLORISTICS

665

Trubetzkoy, in the same article in which he speaks of the 'special place' linguistics occupies among the social sciences. Because of his avoidance of an appeal to techniques of segmentation and classification, Lévi-Strauss's work would thus seem to be immune to Chomsky's criticisms of structuralism. This would be important in that Lévi-Strauss, while not primarily a folklorist, is undoubtedly the single most influential person today who advocates the structural approach to folklore, particularly myth. Chomsky himself has briefly discussed the work of Lévi-Strauss as one of the attempts to find something comparable to language structure in such nonlinguistic domains as kinship and folk taxonomy.6 Noting that Lévi-Strauss models his work on that of Trubetzkoy and Jakobson, Chomsky ( 1 9 6 8 : 6 6 ) states that 'several reservations are necessary when structural linguistics is used as a model in this way'. His general conclusion is that 'the problem of extending concepts of linguistic structure to other cognitive systems seems to me, for the moment, in not too promising a s t a t e . . . ' (Ibid.). Chomsky's cursory and rather negative discussion of Lévi-Strauss is puzzling in light of the fact that numerous scholars have discerned striking similarities in their thinking. For instance, Sperber (1968:207) has asserted that although LéviStrauss is faithful to the language of structural linguistics, his conception of myth has little to do with the structuralist model of language. Sperber sees such basic similarities to certain ideas of generative grammar as the following. Both Chomsky and Lévi-Strauss reject the view prevailing among the behavioral sciences that one can study solely the manifestations of human action and not the mechanisms of which they are the observable output. For both, the observable actions depend on underlying mechanisms. Similarly, Piaget (1968:90) has suggested that behind concrete relations, Lévi-Strauss searches for 'la structure sous-jacente et "inconsciente" ne pouvant être atteinte que par la construction déductive de modeles abstraits'.7 Hymes (1968:176), in suggesting the relevance for folkloristics of Chomsky's notions of descriptive and explanatory adequacy, has noted that 'in the sphere of linguistics Chomsky, Lenneberg, and others consider explanatory adequacy to involve innate, species-specific universals of grammatical structure. In the sphere of folklore the capacities are no doubt derived from innate abilities, and the work of Lévi-Strauss would seem to point directly to what some of them might b e . . . . ' Buchler and Selby (1968:21) have also correctly noted similarities between Chomsky makes no mention of Lévi-Strauss's work in myth, but indications are that he would regard this as an area lying outside the domain of linguistics proper (see Chomsky 1968 : 2 ) . 7 Piaget does, however, discern some differences between the views of Chomsky and LéviStrauss: 'Mais cet esprit humain invariant ou "activité inconsciente de l'esprit" occupe dans la pensée de Lévi-Strauss une position précise, qui n'est ni l'innéité de Chomsky, ni surtout le " v é c u " . . . , mais bien un système de schemes s'intercalant entre les infrastructures et les superstructures* (93). 6

666

WILLIAM O. HENDRICKS

Lévi-Strauss's views and 'certain epistemological foundations of . . . the theory of generative grammar'. They note that The historically dominant view in British social anthropology and American cultural anthropology . . . is that one ougjht to begin with die empirical facts of performance and work one's way backward. In psychology, linguistics, and anthropology this strategy often crystalizes into some form of neobehaviorism. Lévi-Strauss and the generative grammarians, on the other hand, hold to the view that a theory of competence must of necessity precede a theory of performance. Attempts to interpret Lévi-Strauss' studies within a neobehavioristic frame of reference have led to much confusion and incomprehension concerning Lévi-Strauss' conception of social structure and the structural study of myth (22-4). Buchler and Selby further note that Lévi-Strauss has agreed to a suggestion that his thought is anchored in certain aspects of eighteenth century thought, such as Leibniz's notion of a universal grammar. Although there are some obvious similarities in the thought of Lévi-Strauss and of Chomsky, these exist more at the general epistemological level than at the technical level of formalization (cf. Matthews 1967). It is basically at the technical level that Chomsky in fact criticizes Lévi-Strauss's work. Commenting on LéviStrauss's frequent use of Jakobsonian phonology as a model, Chomsky (1968:66) states that the structure of a phonological system is of very little interest as a formal object; there is nothing of significance to be said, from a formal point of view, about a set of fortyodd elements cross-classified in terms of eight or ten features. The significance of structuralist phonology . . . lies not in the formal properties of phonemic systems but in the fact that a fairly small number of features that can be specified in absolute, languageindependent terms appear to provide the basis for the organization of all phonological systems. . . . Furthermore, to a greater and greater extent, current work in phonology is demonstrating that the real richness of phonological systems lies not in the structural patterns of phonemes but rather in the intricate systems of rules by which these patterns are formed, modified, and elaborated. The structural patterns that arise at various stages of derivation are a kind of epipihenomenon. Two points need to be made. The first concerns the putative epiphenomenality of structural patterns. Phonological rules have an input — abstract phonological units — which they relate to actual phonetic entities — the output. The abstract units are interrelated in terms of opposition and correlation — the very type of relations Lévi-Strauss emphasizes. Furthermore, these relations are not on the 'surface' as they are in structuralist phonology. A more basic issue concerns Lévi-Strauss's use of phonological analysis as a model for his own investigations. Before examining this particular issue, let us first consider the question of whether the formalization Chomsky has developed for the description of syntax can be applied to Lévi-Strauss's achievements in myth analysis. Sperber (1968:208) has asserted that 'La formalisation, avant l'élaboration d'hypotheses plus spécifiés, n'a, à notre sens, guere de raison d'être. Aussi se gardera-t-on de postuler un parallélisme avec le modele génératif du langage, qui a

LINGUISTICS AND FOLKLORISTICS

667

le mente d'être trop spécifié pour qu'on puisse l'appliquer a d'autres objets que le sien.' The soundness of Sperber's judgment is borne out by Buchler and Selby's (1968) abortive attempt at a 'formal' study of myth, by which they intend more or less what Chomsky means by 'formal linguistics', namely, an attempt at formalization in the mathematico-logical sense. More specifically, they try to reformulate LéviStrauss's (1967) analysis of the Story of Asdiwal, using such Chomskian notions as deep and surface structure, phrase-marker, rewrite rule, etc. As a preliminary to discussing Buchler and Selby's proposals, a very brief exposition of Lévi-Strauss' approach will follow. Fundamental to Lévi-Strauss's analytic approach is the distinction between sequence and schema (cf. Lévi-Strauss 1967:17). The former is temporal and the latter is non-temporal. The former may also be more or less equated with plot — it pertains to the sequence of events or actions that make up a myth. Schemata, however, pertain to 'mythological' meaning. Another way of characterizing the distinction is in terms of the two levels of signification recognized by literary critics — the literal and the symbolic. On the literal level, that of the narrative sequence, a king is a king and a shepherdess is a shepherdess. But on the symbolic level, the segments of the narrative are bearers of a signification different from their usual linguistic meaning. Furthermore, these can be characterized in terms of binary oppositions, such as high/low on the dimension of social status. Various schemata — geographical, cosmological, sociological, etc. — represent, in Lévi-Strauss's view, so many different 'codes' for transmitting the same message. The various schemata can be integrated into a global one consisting of a number of binary oppositions and representable in a diagram. (The following is a slight modification of a diagram given by Lévi-Strauss 1967:19.)

High

Low

/ Water

/

Sea hunting

\ Land

A

Mountain hunting

A

Peak Valley

668

WILLIAM O. HENDRICKS

Buchler and Selby attempt to equate sequence and schema in various ways to Chomskian notions: sequence = sentence, schema = structural description; sequence = performance, schema — competence; sequence = surface structure, schema = deep structure or base phrase-marker. Further, 'just as in linguistic theory, LéviStrauss' schemata are a system of rules that relate signals to semantic interpretations of these signals' (21-2). The above equations evince a number of misunderstandings. The diagrams of schemata given by Lévi-Strauss, while superficially in the form of phrase-structure trees, are not in fact equivalent to them. Rather, they are more comparable to taxonomic diagrams (of logical hierarchization). Buchler and Selby state that the schemata are 'a system of rules', but in generative grammar phrase-markers are NOT a system of rules — the phrase-markers are generated by a system of rules. This may merely be inexact language on the part of Buchler and Selby, for they do give a set of rules, in the rewrite format, which they claim generates schemata representing Lévi-Strauss's analysis of the Story of Asdiwal. However, these rules do not have the same constraints on them as Chomsky's. For instance, they allow rules of the form X + Y Z, resulting in trees with the following configuration:

X

Y Z

Levi-Strauss's schemata are non-sequential in nature, but the trees produced by Chomsky's base component are not. The base component generates concatenated strings, not unordered sets of elements. Further, the schemata indicate oppositions that are not necessarily found within a single given myth. Thus, some of Buchler and Selby's diagrams refer to two narratives, not just one. They label the initial node 'S', for 'sequence', but what is actually diagramed are oppositions underlying the initial sequence of two versions of the same myth. Although supposedly drawing upon the form of generative grammar presented by Chomsky (1965), Buchler and Selby still assume that several phrase-markers may underlie (be the basis of) a given sequence — in other words, they still operate with generalized transformations. They give three phrase-markers which they say are the basis of a 'well-formed mythical sequence' (the opening sequence of the Skeena version of the Story of Asdiwal), along with the following 'transformational history', showing the steps involved in mapping the 'deep structure' (represented by the three phrase-markers) onto the 'surface structure' (the opening sequence): 'First, apply a "Geographical" Transformation T g to the base Phrase-marker (1); embed the result in the base Phrase-marker (2) by a generalized substitution transformation T e ; next apply an Alimentary Transformation T a to the base Phrase-marker (2);

LINGUISTICS AND FOLKLORISTICS

669

embed the result in the base Phrase-marker (3) by T e ' (38). Buchler and Selby use the term transformation in a way devoid of any content. They fail to give any sort of indication of the operations the various 'transformations' are supposed to effect. Despite the major criticisms one can make of Buchler and Selby's effort, it nevertheless remains of interest as one of the first attempts to develop a formalized study of myth. Part of their difficulty perhaps results from the fact that LéviStrauss's schemata come closer to corresponding to the representation of sentences on the sememic stratum, in Lamb's stratificational theory, than they do to deep structure representations of sentences in Chomsky's syntax-centered approach. To develop this idea, let us first return to Chomsky's criticism of Lévi-Strauss's use of phonological analysis as a model for his investigations. Chomsky asserted that 'there is nothing of significance to be said, from a formal point of view, about a set of forty-odd elements cross-classified in terms of eight or ten features'. However, Lévi-Strauss is not directly interested in phonological elements per se, but rather in the possibility of analyzing semantic phenomena in an analogous fashion into simultaneous bundles of minimal components. (Cf. Hjelmslev's (1963) treatment of the plane of content along the lines of his treatment of the plane of expression; though, as Lamb (1966:194-5) has noted, there is a great disparity in the number of minimal units on the two planes — semantic components might well number in the thousands.) Lévi-Strauss has noted that presently it is recognized that language is structured at the phonological and grammatical levels, but that it is less certain that it is structured at the level of vocabulary, except for certain privileged domains (1960: 146). His position is that in myths and tales there is structure at all levels — including vocabulary: Mais une autre dimension s'ajoute à l'habituelle, parce que règles et mots y servent à construire des images et des actions qui sont, à la fois, des siginificants 'normaux' par rapport aux signifiés du discours, et des éléments de signification, par rapport à un système significatif supplémentaire, qui se situe sur un autre plan; disons, pour éclairer cette thèse, que, dans un conte, un 'roi' n'est pas seulement un roi, et une 'bergère', une bergère, mais que ces mots, et les signifiés qu'ils recouvrent, deviennent des moyens sensibles pour construire un système intelligible formé des oppositions: mâle/femelle (sous le rapport de la nature) et: haut/bas (sous le rapport de la culture) . . . Tout en restant des termes du discours, les mots du mythe y fonctionnent comme des paquets d'éléments différentiele. Du point de vue de la classification, ces mytfaèmes se situent, non pas sur le plan du vocabulaire, mais sur celui des phonèmes; avec cette différence qu'ils n'operent pas sur le même continuum (ressources de l'expérience sensible, dans un cas, de l'appareil phonateur, dans l'autre); avec cette ressemblance, aussi, que le continuum est décomposé, et recomposé, selon des règles, binaires ou ternaires, d'opposition et de corrélation.

Lévi-Strauss's ideas are rich in possibilities for development, although they need to be qualified in certain ways. For instance, we would reject his implication (in remarks not quoted above) that a 'hyper-structural' plane of signification is a distin-

670

WILLIAM O. HENDRICKS

guishing feature of myths and folktales and not narrative discourse in general. Further, Lévi-Strauss apparently does not believe that words on the plane of language can be analyzed into components — a componential analysis presumably is possible only on the hyper-plane of signification. This at least seems to be the import of his remark that on one level of signification a 'king' is a king and a 'shepherdess' a shepherdess, whereas on a higher level it is a matter of the components male/female and high/low. It would be more exact to say that semantic components exist on both levels of signification. In Lévi-Strauss's example, the components male/female would belong on the level of language; the components high/low (social status) would belong on the level of text structure. That is, the first pair would result from an analysis of the language in general, whereas the second pair would result only from the analysis of these particular words in a given text. In an analysis of vocabulary in general (and not in a given text), king and shepherdess would not likely be directly paired; rather, the following pairings would be more likely: king-queen shepherd-shepherdess prince-princess waiter-waitress etc. Here a male/female opposition would be easily noted, but an opposition between high and low social status would not emerge — that is, one would probably not pair queen with waiter, etc. But these two terms could occur in the same text and be in opposition. The recognition of two levels of signification may perhaps obviate some of the objections raised against the componential approach to semantics. For instance, assume that man is asserted to contain the component 'male' as opposed to 'female', which implies that man is paired with woman. However, if man were paired with God, then in that case the components mortal/immortal would be relevant. The components of man would thus seem to be infinitely variable, as a function of what particular word(s) it is contrasted with. The above argument was advanced in the course of the discussion of a paper delivered by Lamb on sememic analysis (1966:196). However, some of Lamb's remarks in his paper can be cast into a new light and used to refute this particular objection to componential analysis. In the course of a sample stratificational analysis of the sentence the man caught the tiger, Lamb posits as semantic components of tiger the following: 'thing', 'animate', and '(tiger)'. The element enclosed in parentheses indicates what is left of tiger after the first two components are extracted. It is what distinguishes a tiger from all other animals. But, according to Lamb (1966:194), 'linguistically, all we have is a single structural element which it is not our responsibility to relate to the real world. One of the factors here is that

LINGUISTICS AND FOLKLORISTICS

671

there is a great deal of individual difference: different people have varying amounts of information on what it is that differentiates tigers from other animals.' Rather than assuming that the variability is due to individual differences of knowledge, we can hypothesize that this is a function of the different words that could potentially be contrasted with tiger in a given text. There is a range of possible differentiating features, but the context of the text sharply limits the possibilities. We may reinterpret Lévi-Strauss's example in these terms. The components of king might be 'male' and '(king)' on the level of language, but on the level of discourse the diffuse component '(king)' is specified as 'high social status' (in contrast to the 'low' of shepherdess). The above suggestions are obviously only first approximations, but hopefully they do indicate some of the potential linguistic implications of certain folkloristic research. In particular, this line of inquiry seems to point in the direction of a generative semantics approach to language description.8 Although work in the structural analysis of folklore, particularly that of LéviStrauss, seems to favor one type of linguistic description over another, this in one sense is a secondary issue. Of greater importance is the implication of this work for the fundamental distinction in linguistic theory between langue and parole, and the related notion, stressed by Chomsky, of linguistic creativity. In a famous essay published over four decades ago, Bogatyrev and Jakobson (1929) suggested that Saussure's distinction between langue and parole is applicable to folklore. The linguistic analogy was said to offer an explanation for the striking similarities that can be observed in tale plots throughout the world: In folklore as well as in language, only a part of the similarities can be explained on the basis of common patrimony or of diffusion (migratory plots). And, since the fortuity of the other coincidences is impossible, there arises imperatively the question of structural laws that will explain all these striking coincidences. . . . The explanation that we have tried to develop in relation to corresponding linguistic phenomena suggests itself similarly for the tale patterns. The folk tale is a typically collective ownership. The socialized sections of the mental culture, as for instance language or folk tale, are subject to much stricter and more uniform laws than fields in which individual creation prevails.9

Although Bogatyrev and Jakobson appeal to the notion of 'collective creation' (kollektives Schaffen), they have nothing mystical in mind. Saussure's langue is a 8

Most work along this line, while not yet very developed, still largely remains sentence oriented. Despite the title of his 1966 paper, Lamb himself dealt only with the sentence. Real progress in developing a generative semantics would seem to entail refining Hjelmslev's (1963) notion of connotation. • These remarks occur in Jakobson (1945:640-1), but they closely correspond to the discussion in the joint essay written earlier. Cf., for instance, this remark: 'Ähnliche Sujets entstehen auf Grund der allgemeinen Gesetze der dichterischen Komposition (These von Viktor Sklovsky), die gleich den Strukturgesetzen der Sprache gegenüber dem kollektiven Schaffen gleichförmiger und strenger sind, als gegenüber dem individuellen' (911). Jakobson cites Propp's Morphology as an example of a study of structural (or compositional) laws. However, more suggestive in this regard is Olrik (1909). For an extended comparison of Propp and Olrik, see Hendricks (1970).

672

WILLIAM O. HENDRICKS

matter of collective creation in their sense — it is a matter of supra-individual conventions that underlie the speech activity of individual speakers (parole). The expression is not meant to be taken literally.10 Saussure's recognition of langue in addition to parole implied a rejection of the Junggrammatiker thesis that, in the final analysis, only the speech of a definite person at a definite time represents true reality. This thesis is a specific instance of a more general doctrine that Bogatyrev and Jakobson call 'naive realism'. In the case of folkore, this doctrine rejected the Romantic notion of collective creation; since it is not given to us in any perceptual experience, it was presumed necessary to assume an individual creator, who happens to be anonymous (see Bogatyrev and Jakobson 1929:903). Chomsky's notion of linguistic creativity is actually comparable in many respects to Bogatyrev and Jakobson's notion of collective, as opposed to individual, creation. That is, the individual's creative competence to produce or understand an unlimited number of novel sentences is due to his internalization of a set of generative rules, which with minor exceptions are assumed to be the same for all speakers of the given language. From one point of view, the existence of such socialized rules implies the lack of freedom and creativity for the individual. Much of the difficulty here is terminological. For Saussure, apparently, creativity was solely a matter of individual (non-socialized) action in the exécution ('performance') of language: activating the vocal organs and conceiving what one wants to say. Insofar as synonymy exists, the choice of linguistic signs to match the extralinguistic situation is not determinate. Since Chomsky does not concern himself with performance — the use of language in specific communicative situations — it is clear that he is not interested in creativity in this sense. Of course, as is well known, Saussure was ambivalent as to whether the sentence was an aspect of langue or of parole. If the latter, then langue is basically a collection of (paradigmatically related) simple linguistic signs, whose (syntagmatic) combination into complex signs (i.e. sentences) is a matter of individual 'creative' acts. On the other hand, if the sentence belongs in langue, then langue is not solely a lexicon, but also comprises patterns which regulate and restrict the combination of linguistic signs. By 'linguistic creativity' Chomsky means rule-governed creativity; and he speaks of Saussure's view of sentence formation as an aspect of parole, as 'a process of FREE creation, unconstrained by linguistic rules except insofar as such rules govern the forms of words and the patterns of sounds' (Chomsky 1968 : 17; emphasis added). But insofar as Saussure included sentence formation in langue, Chomsky means by linguistic creativity what Saussure would consider a mechanical process.11 10

Sapir (1963) made much the same point in his discussion of unconscious patterning of behavior in society; see particularly p. 544. 11 Chomsky contrasts generative with mechanical. By the latter he means phenomena accountable for in terms of S-R conditioning or concepts proper to the interaction of physical bodies (devoid of 'mind'). A generative model (which implies linguistic creativity) assumes that the sentences produced by individuals cannot be accounted for in terms of memorized or fixed phrases restricted to stereotyped situations.

LINGUISTICS AND FOLKLORISTICS

673

To the extent that the production of discourse is not a totally free process on the individual's part, then to that extent the text is no more a part of parole than the sentence. Part of the resistance to the acceptance of this idea may be due to a confusion of rule-governed creativity with what Chomsky (1968:8) calls 'TRUE creativity, an exercise of the creative imagination in ways that go beyond normal intelligence . . . ' (emphasis added). There is a tendency to think of texts only as written discourse, and it is typically only the product of more or less original thinking that gets committed to print in the first place. However, verbal art folklore is connected discourse which in one wide-spread view presents only average, unexceptional thought. Consider the following remark, which occurs in an introductory text on folklore: 'Folklore comprises the unrecorded traditions of a people. The study of folklore records and analyzes these traditions because they reveal the common life of the mind below the level of "high" or formal culture, which is recorded by civilizations as the learned heritage of their times' (Brunvand 1968 : 1). Of course, there is also the view expressed in one definition of the proverb as 'the wisdom of many, the wit of one' (attributed by Dundes (MS) to Bertrand Russell). And Labov and Waletzky (1967:12) have recently noted that 'most attempts to analyze narrative have taken as their subject matter the more complex products of long-standing literary or oral traditions.... In our opinion it will not be possible to make very much progress in the analysis and understanding of these complex narratives until the simplest and most fundamental narrative structures are analyzed . . . We suggest that such fundamental structures are to be found in oral versions of personal experiences: not the products of expert story tellers that have been re-told many times, but the original productions of a representative sample of the population.' Labov and Waletzky analyze narratives elicited from a wide range of speakers — who have in common the fact that none finished high school. Although they do not explicitly say so, their study is folkloristic in that narratives of personal experience constitute one genre of verbal folklore, termed memorate by von Sydow (1948). However, one does not have to turn to inexperienced story tellers to escape the problem of 'true creativity'. According to Jacobs (1960 : 123), 'the dogma of signal creativity of the rare genius may have deflected research away from the quest of processes which reside in community manipulations of stories and in impact of the populace upon narrators'. The processes Jacobs refers to are labeled by Bogatyrev and Jakobson the 'preventive censorship of the community' (die Praventivzensur der Gemeinschaft), an important notion auxiliary to that of collective creation. According to Bogatyrev and Jakobson (1929:901), 'Die Existenz eines Folkloregebildes als solches beginnt erst, nachdem es von einer bestimmten Gemeinschaft angenommen wurde, und es existiert von ihm nur das, was sich diese Gemeinschaft angeeignet hat.' They differentiate between folklore and written literature on the basis of the fact that preventive censorship is not at work in the case of the latter. This means that the work comes into existence the moment the writer puts it down

674

WILLIAM O. HENDRICKS

on paper. Further, unlike the creator of folk literature, the literary writer can go against the grain of his social milieu. The major emphasis of this essay has been upon the contribution folkloristics can make to linguistics. As a conclusion, some tentative suggestions will be made as to possible contributions of linguistics to folkloristics. Bogatyrev and Jakobson's essay will provide the framework for these suggestions. An attempt will be made to demonstrate that their appeal to linguistic concepts represents no real advancement, maintaining to a large extent the status quo. Before doing so, however, it should be noted that the survey of potential contributions of folkloristics to linguistics has been somewhat narrow. An interest in analysis beyond the sentence represents only one major trend towards the expansion of the scope of linguistics (or the establishment of interrelations with other disciplines). Such an expansion continues the linguist's traditional focus on the underlying code, and not its use in concrete communicative situations. A second major trend involves a concern for what Hymes (1965) calls the 'ethnography of communication'. As with trans-sentence analysis, a case can be made that folklore offers a privileged domain for such studies. In ordinary linguistic communication one cannot typically abstract 'creativity' in a generative sense (the production of sentences) from 'creativity' in applying certain messages to a particular (novel) situation. However, generative creativity is not at issue in the case of a fixed-phrase folkloristic genre such as the proverb.12 As Arewa and Dundes point out (1964: 80), 'nor is use of proverbs conditional upon skill in creative and adaptive change in message form. By the very nature of fixed-phrase genres of verbal art, the messages are culturally standardized in form and content. The creativity and adaptation lie rather in the successful application of these traditional materials to new situations.' Arewa and Dundes examine a dozen Yoruba proverbs and attempt to establish rules for applying them. The most important of these have to do with the identity of the participants in the speech situation. Of primary consideration is the age of the person speaking relative to the age of the addressee. The identity of the addressee is crucial with regard to whether a particular proverb is appropriate or not. Let us turn now to a critical examination of the proposals of Bogatyrev and Jakobson. In characterizing folklore, they give major importance to the criterion of oral transmission. In the opinion of many, however, this represents an unprogressive view. According to Ben-Amos (1971), this criterion has become the last citadel of folklore scholars in defending the uniqueness of their materials. Yet, as he notes, the medium of transmission does not define what folklore really is — it merely provides a qualifying statement about its form of circulation. In Hymes's opinion (1962: 688), 'to understand the continuity in the cultural use of verbal art, 18 By saying that the proverb is a 'fixed-phrase' genre, one means that the exact wording of proverbs does not vary; e.g. we have coffee boiled is coffee spoiled, but not such variants as coffee is spoiled if it's boiled, spoiled coffee is boiled coffee, etc.

LINGUISTICS AND FOLKLORISTICS

675

we have to transcend the common dichotomy between oral and written transmission. That ostensible difference must be seen as often superficial, if we are to preserve the relevance of folklore as a discipline. . . . Yet if folklore can only be oral and traditional in the ordinary sense of being anonymous, then it is about to all but disappear from human society.' The view is occasionally expressed that folklore is in fact a dead trait in the modern world (e.g. Benedict, quoted by Dundes 1969:13). This view is due, at least in part, to the mistaken assumption that the folk are illiterates in a literate society. According to Dundes (1969:14-15), 'it is wrongly assumed that literate people have no folklore.... It is certainly doubtful whether increased literacy and education have seriously affected the quality and quantity of folk speech or jokes, at least in American culture.' Fischer regards jokes as relatively trivial, as compared to the folktales of most non-literate groups. In his view (1963:291), 'in modern society the comic strip, the television serial, and the best-selling paperback may all be more nearly functional equivalents of the folktale than the short humorous anecdote'. A seemingly reasonable hypothesis is that as literacy increases, folklore increasingly becomes something transmitted by the graphic medium. The putative disappearance of folklore in modern urban society might be compared with Piaget's view that egocentric speech disappears from the child's speech when he reaches a certain maturity. Vygotsky has criticized this view, noting that, for the child, speech initially serves several undifferentiated functions. As these functions become differentiated, egocentric speech becomes silent (inner speech). 'The inner speech of the adult represents his "thinking for himself" rather than social adaptation: i.e., it has the same function that egocentric speech has in the child. It also has the same structural characteristics' (Vygotsky 1962:18). In an analogous fashion, one can point out that in non-literate societies everything — including folklore — is orally transmitted. If the society acquires an orthography, functional uses of language may be differentiated as to mode of transmission, oral or written. The fact that writing may now be used for certain functions should not suggest that these functions have disappeared. As Fischer notes (1963:290), in a literate society the longer and more serious narratives are now all written down. The function of the surviving short oral forms, such as the joke, is somewhat different, therefore, from the function of the typical non-literate tale. Aside from the issue of medium of transmission, Bogatyrev and Jakobson's sharp differentiation between folklore and written literature has another aspect of questionable validity. While not going so far as to suggest a mutual exclusion between the social (langue) and the individual (parole), they do come close to this by characterizing folklore in terms of focus on the former; literature, by focus on the latter.18 As Sapir has pointed out (1963 :544), 'the terms "social" and "individual" 1S

'Ein wesentlicher Unterschied zwischen der Folklore und der Literatur besteht darin, dass fur die erstere die Einstellung auf die langue, fur die letztere die auf die parole spezifisch ist' (905).

676

WILLIAM O. HENDRICKS

are contrastive in only a limited sense.... Strictly speaking, each kind of behavior is individual, the difference in terminology being entirely due to a difference in point of view.' What this suggests is that, rather than differentiating folklore and literature on this basis, one can regard each from both the social and the individual viewpoint. In reality both aspects are more or less equally copresent. A pervasive bias in Western literary criticism stresses the unique and the innovative — at the expense of a proper recognition of the persistence of certain rather fundamental conventions that make a narrative a narrative. But if, as Bogatyrev and Jakobson suggest, literature is basically a matter of parole, then the implication is that no (or hardly any) supra-individual conventions restrict the production of writers. The situation in folkloristics is the inverse of the one in literary criticism. Jakobson (1945:639) has noted that 'it is true that the hypothetical^ reconstructed archetype of a tale interested [the Russian folklorist] Afanas'ev perhaps more than its actual, individual variants, but he did not follow the fanatic principle that the eminent historian of literature, A. Pipin, ascribed to him, namely, that everything expressing the arbitrary manner of the individual should be weeded out from the presentation of the tale as "twaddle that is only personal".' But Jakobson later goes on to say that 'variations must not be overestimated. Afanas'ev avoided the danger of missing the tale itself behind its variants' (641). A more balanced view has been expressed by Fischer (1963:258): the 'use of myth by individuals to justify individual goals . . . and to win social support for "selfish" acts is probably supported by the field experience of every ethnographer who has paid any attention to folktales'. He goes on to point out that the individual, to affect the behavior of others, must restrain himself if he is to gain acceptance of at least some other members of his society. He concludes that a tale can persist only if it attains some balance between the social and the individual. Folklorists' traditional notion of 'variant' demands careful consideration. In their treatment of tales as variants of a historically earlier form (reconstructed archetype), traditional folklorists are vulnerable to the charge of committing a conceptual error analogous to one Saussure perceived in traditional linguistics. According to Godel (1966:483), Saussure 'very soon discovered that all the errors and inconsistencies in traditional linguistics originate in a confusion of two sets of connections or relationships: those which hold between contemporary forms belonging to the same system (e.g. Eng. father, fathers, fatherly; mother, son; he, etc.)

and those which hold between forms that succeed each other in time (e.g. IE *pstérand Eng. father).' These two types of relations correspond to two types of identity. On the one hand, there is diachronic identity, 'involving the preservation of a sign through centuries' (485). That is, IE * pater- is identical to father, despite all changes, because it has not ceased to be a linguistic sign, distinct from all other English words inherited from IE. On the other hand, there is synchronic identity: 'father is of course identical with father, whenever it recurs in a talk or a tale, it is

LINGUISTICS AND FOLKLORISTICS

677

similarly understood by the partners or the reader, regardless of the fact that its pronunciation may vary from one utterance to another . . .' (485). Saying that a number of tales are 'variants' of a historically earlier one is comparable to saying that Eng. father is 'identical' to IE *pstér-. Such an identification is a consequence of stressing the traditionality of folklore. But as Ben-Amos (1971) has noted, traditionality is a scholarly, analytical construct, not a cultural fact. The antiquity of material can be established only after extensive research, and the tellers themselves are ignorant of it — just as the typical contemporary speaker of English is unaware of the fact that *pater- and father are cognates. Since Saussure, we recognize that a language can and should be analyzed without reference to its previous states. Just as in the case of language, not all variability in folkloristic material is historical. Bogatyrev and Jakobson basically have synchronic variability (due to the varying exécution of individuals on different occasions) in mind when they suggest that the langue-parole distinction is applicable to folklore. They state (904-5): 'In der Folklore ist das Verhältnis zwischen dem Kunstwerk einerseits und seiner Objektivierung, d.h. den sogenannten Varianten dieses Werkes beim Vortragen durch verschiedene Personen andererseits vollkommen analog dem Verhältnis zwischen der langue und der parole. Gleich der langue ist das Folklore-Werk ausserpersönlich und führt nur eine potenzielle Existenz, es ist nur ein Komplex bestimmter Normen und Impulse, ein Canevas aktueller Tradition, die die Vortragenden durch die Verzierungen des individuellen Schaffens beleben, gleich wie es die Erzeuger der Parole gegenüber der Langue tun.' It should be noted that Bogatyrev and Jakobson do not mean exactly what Saussure meant by exécution. That is, the folkloristic 'variations' they refer to (the analogue to parole in folklore) are not simply a matter of variation in the phonetic actualization of the linguistic signs making up a given work. Nor do they primarily have in mind the linguistic variations that can be expected in the case of free-phrase, as opposed to fixed-phrase, folkloristic genres. Rather, they presumably have in mind what some would term 'folkloristic' as opposed to 'linguistic' variation: differences in nomenclature and attributes of dramatis personae, in the presence or absence of certain plot episodes, the specific nature of certain generic actions, etc. Since Bogatyrev and Jakobson are applying the langue-parole distinction in an analogous, not a literal, fashion, such differences in themselves are not necessarily illegitimate. But there are other objections to their proposal. When they equate an invariant tale with langue, and its various manifestations with parole, they seem to imply a conception of langue as a mere collection of entities (different works of folklore). It is only later in the course of their argument that they speak in terms of structural laws. This matches Saussure's own ambivalence as to whether the sentence belongs in langue or parole. At best, the notion of 'variant' represents a downgrading of the notion of creativity in verbal art. Sentences can be expanded in various ways — through embed-

678

WILLIAM O. HENDRICKS

dings, etc. — and this is seen by linguists as a manifestation of 'creativity'. Folktales are subject to an analogous expansion, but this is regarded by folklorists as yielding merely 'variants'. Bogatyrev and Jakobson, rather than regarding folklore in terms of synchronic identity in the narrow sense (father is identical to father), should have done so more in terms of relations among contemporaneous forms belonging to the same system (father, fathers, fatherly). A more serious weakness of Bogatyrev and Jakobson's proposal is that they have not in fact succeeded in clearly separating the synchronic from the diachronic. Consider, for example, this remark: 'In wiefern diese individuellen Neuerungen in der Sprache (bezw. in der Folklore) den Forderungen der Gemeinschaft entsprechen und die gesetzmässige Evolution der langue (resp. der Folklore) antizipieren, insofern werden sie sozialisiert und bilden Tatsachen der langue (resp. Elemente des Folklore-Werkes)' (905). Elsewhere a distinction is made between 'Sprachveränderungen' on the one hand and 'individuellen Sprechfehlern' on the other. That is, most of the individual contributions of the tale teller are regarded as comparable to speech lapses. Only if they become socialized do they form part of the work of folklore itself. Such an appeal to 'socialization' introduces the time dimension and is another way of asserting the traditionality of folklore.14 This tacit diachronic bias lies behind Bogatyrev and Jakobson's deemphasis of the role of the individual in folklore, at the expense of the social. For instance, with regard to rite, they state that 'wenn auch im Keime dieses oder jenes Ritus eine individuelle Aeusserung lag, der Weg von dieser bis zum Ritus ebenso weit ist, als der Weg von der individuellen Deformation der Rede bis zur grammatikalischen Sprachmutation' (904). Although Bogatyrev and Jakobson's discussion of the 'preventive censorship of the community' parallels linguists' discussions of the mechanism of linguistic change, one can talk about such a notion in a totally synchronic context. That is, rather than regarding it in terms of changes in the system — changes from a status of impossible to possible — one can regard it in terms of a distinction between the occurring and the possible but non-occurring. Consider in this regard the distinction between, in English, the impossibility of /bnik/, the occurrence of /brik/, and the non-occurrence, but theoretical possibility of /blik/. One can talk of 'censorship' as a matter of ruling out the logically possible. The impossible, in terms of the unconscious system intuitively known by the community, needs no special sanction against it. A. J. Greimas (1967), in a recent study of several Lithuanian variants of a European folktale, posited as one logical possibility the existence of a variant in which a priest has a son. Such a possibility would be rejected by a 14 Hymes (1968 : 175) has asserted that 'the nature of the usual emphasis in folklore research upon the traditional has cost heavily'. He cites the case of a collector who discovered that some songs he had collected had been composed by the informant herself. The collector destroyed the tapes because the songs were 'non-traditional', but as Hymes notes, 'What he destroyed was from the standpoint of a structural ethnography the most valuable portion of his work: spontaneous evidence of the productivity of the rules of the genre' (176).

LINGUISTICS AND FOLKLORISTICS

679

devout, rural Catholic community, so in fact this variant does not occur. This may be referred to as 'preventive censorship of the community' and it is of a completely synchronic nature, not involving changes in the system — only a matter of the utilization, or non-utilization, of the full potentialities of the system. One final point may be made with regard to Bogatyrev and Jakobson's observation (911) that a gradation exists between the extremes of, on the one hand, a mutual exclusion of 'producer' and 'consumer' of folklore, and on the other, a complete overlap of the two. This distinction between producer and consumer may not be of crucial importance if one grants an extension of the Chomskian notions of performance and competence along lines implied by preceding considerations. The generative rules of a language, which attempt to capture the underlying competence of every normal native speaker of a language, are not solely rules for the speaker, but also for the hearer. That is, the same set of rules are assumed to have equal relevance for production and reception of language. Analogously, it may be that the asymmetry between producers and receivers of narratives is not of crucial importance. While not everyone can gain recognition as an excellent story teller, everyone can at least respond to narrative discourse — though perhaps only after certain (informal) learning; the acquisition and development of 'narrative competence' is an almost completely unexplored area of research. Even the 'non-producers', with proding, may be able to produce minimal narratives (cf. in this regard the earlier cited work of Labov and Waletzky 1 9 6 7 ) . 1 5

REFERENCES

and ALANDUNDES. 1 9 6 4 . Proverbs and the ethnography of speaking folklore. AmA 6 6 : 6 / 2 . 7 0 - 8 5 . BEN-AMOS, DAN. 1 9 7 1 . Toward a definition of folklore in context. JAF 8 4 . 3 - 1 5 . BOGATYREV, P . , and ROMAN JAKOBSON. 1 9 2 9 . Die Folklore als eine besondere Form des Schaffens. Donum natalicium Schrijnen, 9 0 0 - 1 3 . Nijmegen-Utrecht, Dekker. BRUNVAND, JAN HAROLD. 1 9 6 8 . The study of American folklore: An introduction. New York, W. W. Norton & Co. BUCHLER, IRA R . , and HENRY A. SELBY. 1 9 6 8 . A formal study of myth. Center for Intercultural Studies in Folklore and Oral History, monograph series, 1. Austin, Univ. of Texas. CHOMSKY, NOAM. 1 9 6 5 . Aspects of the theory of syntax. Cambridge, Mass., MIT Press. . 1968. Language and mind. New York, Harcourt, Brace, & World. COLBY, BENJAMIN N . 1 9 6 3 . Comment on Fischer 1 9 6 3 . CAnthr 4 . 2 7 5 . DROBIN, ULF. 1 9 6 9 . A review of structuralism. Temenos 2 . 2 0 3 - 1 2 . AREWA, E . OJO,

15

The research reported in this paper was supported by an ACLS Study Fellowship.

680

WILLIAM O. HENDRICKS

DUNDES, ALAN.

1962.

From etic to ernie units in the structural study of folktales.

JAF 75.95-105.

. 1963. Comment on Fischer 1963. CAnthr 4.276-77. . 1964. The morphology of North American Indian folktales. (FFC, LXXXI.) Helsinki: Suomalainen Tiedeakatemia. . 1969. The devolutionary premise in folklore theory. Journal of the Folklore Institute 6.5-19. . MS. On the structure of the proverb. FISCHER, JOHN. 1960. Sequence and structure in folktales. Men and cultures, ed. by Anthony F. C. Wallace, 442-6. Philadelphia, Univ. of Pennsylvania Press. . 1963. The sociopsychological analysis of folktales. CAnthr 4.235-95. GODEL, ROBERT. 1966. F. de Saussure's theory of language. CTL 3.479-94. GREIMAS, A. J. 1 9 6 7 . La structure des actants du récit. Word 2 3 . 2 2 1 - 3 8 ( = Linguistic studies presented to André Martinet, part one, ed. by Alphonse Juilland.) HENDRICKS, WILLIAM O. 1 9 6 7 . On the notion 'beyond the sentence'. Linguistics 37.12-51.

. 1970. Folklore and the structural analysis of literary texts. L&S 3.83-121. . MS. Linguistics and the structural analysis of literary texts. HJELMSLEV, LOUIS. 1 9 6 3 . Prolegomena to a theory of language, trans. Francis J. Whitfield. 2nd ed. Madison, University of Wisconsin Press. HYMES, DELL. 1962. Review of Indian tales of North America, ed. by Tristram P. Coffin. AmA 64.676-9. . 1964. Introduction: Toward ethnographies of communication. AmA 6 6 : 6 / 2.1-34. . 1968. The 'wife' who 'goes out' like a man: reinterpretation of a Clackamas Chinook myth. SocScil 7.173-99. JACOBS, MELVILLE. 1960. Thoughts on methodology for the comprehension of an oral literature. Men and cultures, ed. by Anthony F. C. Wallace, 123-9. Philadelphia, University of Pennsylvania Press. . 1963. Comment on Fischer 1963. CAnthr 4.277-9. . 1966. A look ahead in oral literature research. JAF 79.413-27. JAKOBSON, ROMAN. 1 9 4 5 . On Russian fairy tales. Russian fairy tales, tr. by Norbert Guterman, 6 3 1 - 5 6 . New York, Pantheon. JESPERSEN, OTTO. 1922. Language: Its nature, development, and origin. New York, Macmillan. LABOV, WILLIAM, and JOSHUA WALETZKY. 1 9 6 7 . Narrative analysis: Oral versions of personal experience. Essays on the verbal and visual arts, ed. by June Helm, 1 2 - 4 4 . (= Proceedings of the 1 9 6 6 annual spring meeting of the American Ethnological Society.) Seattle, Univ. of Washington Press. LAMB, S. M . 1 9 6 6 . Linguistic structure and the production and decoding of dis-

LINGUISTICS AND FOLKLORISTICS

681

course. Speech, language, and communication, ed. by Edward C. Carterette, 173-99 ( = UCLA Forum in Medical Sciences; Brain Function, 3). Berkeley & Los Angeles, University of California Press. LÉVI-STRAUSS, CLAUDE. 1955. The structural study of myth. JAF 78.428^4. (Reprinted in Lévi-Strauss 1963.) . 1960. L'Analyse morphologique des contes russes. IJSLP 3.122-49. . 1963. Structural anthropology, tr. by Claire Jacobson and Brooke G. Schoepf. New York, Basic Books. . 1967. The story of Asdiwal. The structural study of myth and totemism, ed. by Edmund Leach, 1-47. London, Tavistock Publications. (First published: 1958.) MATTHEWS, P. H. 1967. Review of Chomsky 1965. JL 3.119-52. MELETINSKIJ, ELEASAR. 1 9 6 9 . Zur strukturell-typologischen Erforschung des Volksmärchens. DJbVk 1 5 . 1 - 3 0 . NATHHORST, BERTEL. 1 9 6 9 . Formal or structural studies of traditional tales. Stockholm Studies in Comparative Religion, 9. OLRIK, AXEL. 1 9 0 9 . Epische Gesetze der Volksdichtung. ZDA 5 1 . 1 - 1 2 . (Eng. trans, in The study of folklore, ed. by Alan Dundes. Englewood Cliffs, N. J., Prentice-Hall.) PIAGET, JEAN. 1 9 6 8 . Le structuralisme. Paris, Presses Universitaires de France. POP, MIHAI. 1968. Der formelhafte Charakter der Volksdichtung. DJbVk 14.1-15. PROPP, VLADIMIR. 1 9 6 8 . The morphology of the folktale. 2nd ed. Austin, Univ. of Texas Press. SAPIR, EDWARD. 1963. The unconscious patterning of behavior in society. Selected writings of Edward Sapir, ed. by David Mandelbaum, 544-59. Berkeley & Los Angeles, Univ. of California Press. SEBEOK, THOMAS A . 1 9 5 3 . The structure and content of Cheremis charms, part one. Anthropos 4 8 . 3 6 9 - 8 8 . SPERBER, DAN. 1 9 6 8 . Le structuralisme en anthropologie. Qu'est-ce que le structuralisme?, 1 6 7 - 2 3 8 . Paris, Seuil. SYDOW, C. W. VON. 1948. Folk-tale studies and philology: Some points of view. Selected papers on folklore, 189-219. Copenhagen, Rosenkilde & Bagger. THOMPSON, STITH, ed. 1 9 5 3 . Four symposia on folklore. Indiana University Publications, Folklore Series, 8. Bloomington. VOIGT, VILMOS. 1969. Towards balancing of folklore structuralism. AEH 18.247-55. VYGOTSKY, L. 1962. Thought and language. Cambridge, Mass., MIT Press.

FOLK POETRY: G E N E R A L PROBLEMS

V. N. TOPOROV

1. Some general problems which are encountered in the research of folk poetry and which find reflection in the most authoritative and interesting recent works will be considered here. To begin with — several limitations and explanations. By folk poetry in this case is understood the aggregate of metrically and/or phonically organized texts of which the actual sphere of function is the oral tradition. Included here are various types of oral poetic works — from the epos, ballads, and songs to so-called minor forms: collocations, word formulas, proverbs, sayings, embellishments, riddles, praise names,1 expletives, magic sayings, etc., if they display features of metrical or phonic organization. Since in various different traditions (and sometimes in the same one) these minor forms are represented as sometimes 'poetic' and sometimes 'non-poetic' texts, in many cases in treating these forms one must encroach upon the limits of the 'non-poetic' texts. The same should unavoidably take place in treating those major poetic forms which share several characteristics essential for their analysis (e.g. motif, plot) with other folklore genres (e.g. narrative genres). Finally, in considering the research in the field of folk poetry, one must as far as necessary turn also to those works which, strictly speaking, do not relate to this topic but are important in that they are the ones which deal with the most urgent theoretical problems of folklore and, naturally, stimulate the development of corresponding methods in the area of folk poetry study as well (cf. the definitive role of studies of the folktale in the last decade and a half, which coincides with the appearance of the first translation of the well-known monograph by V. Ja. Propp, Morphology of the folktale, which appeared in Russian already in 1928). It is necessary also to recall that the author of this survey does not pretend to a consideration of the numerous works that is in any way complete and uniform; for him what is important is a definition of some tendencies in these works, which signify the entrance of folkloristics as a whole into a new stage of development which may conditionally be called STRUCTURAL-TYPOLOGICAL. In conjunction with this it should be kept in mind that these tendencies in folkloristics were induced (or, at least, are paralleled) by corresponding LINGUISTIC ideas (and sometimes by ideas born in studies of cultural anthropology and merging with ideas of structural linguistics). In this manner, the pages which follow give evidence of the viewpoint of 1

Cf. Hausa kirari, Yoruba orile, Zulu isibongo, etc.

684

V. N. TOPOROV

contemporary structural linguistics on the state of affairs and further perspectives in the field of folklore. The selection of this view is not accidental; it has been determined both by the general role which linguistics has always played in folklore studies and by the central place which contemporary linguistics occupies in a broad range of the humanities. If one speaks of the role of linguistic theories and method in the development of folkloristics in the past, it is sufficient to mention the fact of the influence of comparative-historical linguistics and the corresponding method, which led to the creation of the comparative-historical school in folklore studies (cf. the same with reference to the study of mythology), which received its classical expression at first in the German folklore tradition (the Grimm brothers, Th. Benfey, W. Mannhardt, J. Bolte, G. Polivka, and others), and later in the Finnish 'historical-geographical' school (A. Aarne, K. Krohn, W. Anderson; cf. also the works of S. Thompson, A. Taylor, and others). In this case it is not necessary to speak of the achievements of this school in the past and present. In the broad historical perspective the superiority of the Finnish folklore school is indisputable, but the possibilities for further progress are far from exhausted. But more important here is an indication of those imperfections in the initial principles and concrete research methodology which evoke an ever-growing criticism of the Finnish school.2 The basic shortcoming of the 'historical-geographical' approach is, probably, that the analogy between the comparative-historical approach in linguistics and the approach in folklore set up in correspondence with it was not beyond reproach. Having adopted the enthusiasm for comparative analysis of facts taken in their historical (temporal) and spatial aspects, the comparative school in folklore was forced to postulate so-called Ur-types, composed at a certain time and in a certain place and giving in the course of further evolution the numerous actually-attested variants (for example, of folktales). Reconstruction of the Ur-type was the result of influence by such concepts as Ursprache in comparative-historical linguistics (cf. indogermanische Ursprache), but without corresponding reasons. The reconstruction of the indogermanische Ursprache (or the Ursprache of other language families) was preceded by the establishment of a SYSTEM of correspondences among the individual languages GENETICALLY connected with each other. Each concrete language fact was considered the result of a regular transformation of some Ur-fact. Consequently, from a rather early time comparative-historical linguistics postulated (clearly or unclearly) its object as a SYSTEM of correspondences. In any case, this system of correspondences was always the framework within which research was carried on. And Neogrammarian atomism, about which so much has been written and most often with censure, was controlled by a system of correspondences. In this sense comparative-historical linguistics was the first area in linguistics to deserve the name exact science and to be amenable to formalization of a mathematical type. In comparative folkloristics a theory of the Ur-type, strictly speaking, could not be !

On criticism of the Finnish school in contemporary folkloristics, see Dorson 1963.

FOLK POETRY: GENERAL PROBLEMS

685

proven, and polygenetic conceptions could not be unreservedly repudiated. The concept of the Ur-type arose here not as a result of the establishment of a system of correspondences among various historically attested reflections of this Ur-type, but as a result of a manifest tendency toward contraction of the circle of local variants of a given folklore form concomitant with deeper delving into the past. Reconstruction of the Ur-type on this basis was a matter of not quite correct extrapolation from this tendency. This same tendency was delimited not so much by the essence of the development of the given cultural pattern as it was by the fact of the loss of the old information in the course of historical development. In this sense the Ur-type is the minimal bundle of information which can be defined by the investigator.3 Any retrospective study must take into account this natural loss of information in time. Theoretically, the concept of Ur-type could be saved if one could construct a system of regular correspondences among the reflexes of some folklore fact (e.g. a plot, group of motifs, chain of functions, etc.) in various traditions. Such a system would not only help define the real possibilities for reconstruction of the Ur-type but also (more importantly) would permit drawing definite conclusions about those elements or positions in the system which were lost (as in comparative-historical linguistics). Conclusions of this type are, strictly speaking, the best proof for the correctness of the postulated system. However, one should not yet hope for the possibility of defining a similar system of correspondences in folklore (even with application to particular topics). Circumstances of various kinds hinder this possibility: in the first place, the fact that a MORPHOLOGICAL analysis of folklore genres has not been made (even the folktale, which was fortunate in this respect, is still far from being analyzed in a sufficiently reliable manner); in the second place, the lack of clarity with respect to WHICH elements preserve the most reliable GENETIC information (in the practice of the Finnish school and related trends, 'cultural' indications played in this respect a larger role than facts of the morphology of the given genre — cf. the role of proper names, place names, 'cultural' objects, indications of an alien tradition, etc.); in the third place, the lack of an explanation for the correlation between genetic and borrowed elements (in comparative-historical linguistics an ability to differentiate native and borrowed elements, on the one hand, and the presence of NOT TOO LARGE a quantity of borrowings, is supposed; in the destruction of the last condition possibilities for application of the comparative-historical method are sharply curtailed and in an extreme case may be brought to a minimum);4 in the fourth place, the absence of ' All these considerations are not directed against the concept of Ur-type as such in folklore but have in mind only the procedure of revealing concrete Ur-types. 4 It is not accidental that many researchers object to the principles and practice of the Finnish school, referring first of all to the fact that its adherents do not take into account (or cannot take into account if they would want to) the fact of the cross-influences (moreover multiple influences) of some traditions on others and of literary versions on oral versions. Cf., e.g., Wesselski (1925, 1931) and Kiefer (1947); cf. also, on the theory of oikotypes, Sydow (1948: especially "Geography and folk-tale oikotypes", pp. 44-59).

686

V. N. TOPOROV

at least some relable typology of folklore forms, the result of which is the inability to place a boundary between accidental (bearing a probable character) and regular (reflecting an ontological basis). Therefore, at the present time perhaps the only possibility foreseen for establishing correspondences in application to folklore, considered on a comparative basis, is to search for cases when the folklore form has been determined by more general 'cultural patterns'. 5 But in such a case (even if it is supposed that one can establish a sufficiently strong and untrivial series of correspondences which permit prediction of rules of cross-cultural translation of folklore forms), the researcher must go beyond the framework of genetic comparison into the broader field of typological comparisons and construct correspondences not so much for folklore forms in various traditions as for various cultural traditions on the whole. 6 Consequently, even with such an approach the 'historical-geographical' school in folklore studies would differ significantly from comparative-historical linguistics, which considered compared units as non-motivated, inherent entities. 7 Thus, by the present time, the living connection between comparative folklore studies and comparative-historical linguistics may be considered exhausted. The foundations have been defined and the limits established. The lessons have been learned, the most important of which is the judicious narrowing of the depth of the reconstruction, the transfer of the basic task of research from a reconstruction of the Ur-type to an establishment of the relationships of variants to each other (their affiliation) within the limits of a given historical period in a given area for which cultural ties are indubitable. The best works in this direction are exactly those which go beyond these limitations (cf. R. Menendez Pidal, A. Heusler, Th. Frings, J. de Vries, A. Taylor, H. O. Nygard, L. Vargyas, J. Kemppinen, B. Vrabie, among others). In the postwar years the tie of folklore studies with linguistics took on new forms. This is a matter of the ever-increasing influence on folklore of that wide complex of ideas which is usually called structural linguistics. At the present time this is undoubtedly the most powerful and fruitful influence on the scholarly study of folklore, although the practical results of this influence are only in the beginning stage. In the effect of structural linguistics on folklore (as, by the way, also on other fields in the humanities), two aspects must be distinguished. The FIRST of them is defined by the fact that structural linguistics is coming forth now as a pattern for 5

Many American folklorists of the cultural-anthropological school have directed attention to this possibility. Cf., in particular, also the practical observations on this, such as 'the cultural determination' of content changes the given folklore form in transition from one tradition to another (e.g. cf. the fate of the European folktale in America or Africa). A. Dundes (1963) analyzed in this manner a version of Zuni A T 121 (cf. also in part Dundes 1964). 6 In such a case it is not excluded that these correspondences will form a system applicable only to the suprafolklore ('cultural') level. 7 Here there is no sense in considering reproaches against the Finnish school for those shortcomings which are shared by all folklore studies as a whole and which will be eliminated with the development of a general theory of folklore and with successes in structural analysis of concrete forms.

FOLK POETRY: GENERAL PROBLEMS

687

other social sciences. It is, on the one hand, the most exact of the social sciences and, therefore, is amenable to the introduction of mathematical (not, moreover, only quantitative) research methods and to the application of computer systems;8 on the other hand, linguistics is the only one of the 'exact' sciences which has to do with meaning in the sense which, with corresponding corrections and clarifications, is characteristic for all the social sciences. Hence the mediating role between the social and the exact sciences played by structural linguistics. Just as linguistics itself solves its internal problems by turning to models worked out in mathematics, a whole series of disciplines in the humanities is beginning to use successfully models and methods borrowed from structural linguistics. The newest works in the field of folklore, poetics, structural study of literature, mythology, etc., reflect to differing degrees the productivity and potential of this influence. The well-known universality of structural linguistics among the social sciences promotes the fact that many ideas formed OUTSIDE linguistics are now accepted (or are beginning to be accepted) through the prism of structural linguistic concepts (cf. cultural anthropology, the theory of cultural patterns, content analysis, theoretical informational concepts of the code and message, etc.).9 The SECOND aspect in which expansion of the ideas of structural linguistics is realized is defined by the fact that in structural linguistics a whole series of technical methods for analysis has been worked out (identification of the units, their description by means of distinctive features and distribution, segmentation of the text, establishment of relationships of elements in the paradigm and syntagm, the rules for transformations and generative processes, operations of analysis in the field of semantics, etc.) which can relatively easily be transferred to contiguous disciplines where they contribute new and valuable results. The same applies to the role which some general linguistic concepts are starting to play, beginning with such concepts as langue — parole, emic level — etic level, code — message, synchrony — diachrony, etc.10 Thus, the development of the conceptual and operational apparatus of structural linguistics and the fact that now it is coming forward as the leading semiotic discipline, to a significant degree defining the general organizational principles of many 8

Hence, the ever-growing role of the experiment in linguistics and the attempts to extend experimental studies in many related fields. With reference to folklore, cf., of the old works: Kònig (1929), and of the newer ones: Anderson (1951, 1956) and Katona (1969 : 398ff.), etc. 9 What has been said in no way contradicts the fact that contemporary folkloristics (first of all Anglo-American) is indebted for much to N e w Criticism (I. A. Richards, W. Empson, M. Bodkin, et al.), theories of patterns of culture (R. Benedict, H. Robert, et al.), Gestalt theory (W. Kòhler, K. Koffka, et al.), as, by the way, also to American linguistic tradition in its 'mentalistic' (E. Sapir) and 'mechanistic' (L. Bloomfield) branches. On the role of F. Boas, see Jacobs (1959a: 127). 10 With reference to folklore (including mythological texts) these concepts have been fruitfully analyzed in many works. Cf. Lévi-Strauss (1958: Introduction: histoire et ethnologie, L'analyse structurale en linguistique et en anthropologie, Linguistique et anthropologic, La structure des mythes, La notion de structure en ethnologie, Structure et dialectique), and 1960; Pike 1954-55; Dundes 1962; Sebeok 1959a; etc. It is necessary to give the predecessors of modern structural folklore studies their due. Of the early works, besides V. Ja. Propp's monograph (1968), cf. Bogatyrev and Jakobson (1929, 1931), etc.

688

V. N. TOPOROV

other disciplines, promotes the creation of new conditions for a more profound effect by linguistics on other social sciences.11 Of course, this effect is especially perceptible when it is a matter of studying systems of signs at the basis of which natural language lies. In this respect folkloristics occupies an exceptionally privileged position since it has to do with analysis of relatively SIMPLE TEXTS in natural language (by the latter may tentatively be understood such a class of texts where a small quantity of extralinguistic structures for which rules are easily formulated for their deduction from linguistic structures are constructed over strictly linguistic structures). In this sense, opposed to folklore stands structural poetics, which has to do with essentially more COMPLEX texts (belles-lettres), for the analysis of the higher levels of which a sufficiently reliable procedure cannot yet be proposed. At the same time, if models worked out in an analysis of folklore texts turn out to be, as a rule, too weak and trivial or too abstract as applied to a study of literary texts,12 then the methods of structural analysis of belles-lettres usually may be used effectively enough in folklore studies (cf. the area of metrics, rhyme, sound organization, 'grammar of poetry', composition, style, content structure, etc.). At the same time, the tie of structural linguistics with folklore is not unilateral. Linguistics, accepting such a responsible role in the circle of humanities disciplines, cannot help but seek for itself new fields for the application of its energies and for experimentation. In a certain sense (and this is not an exaggeration), the limits of these fields are defined by the possibility of application of the concept of natural language, in the first place, and LANGUAGE in its SEMIOTIC aspect (as any system of signs plus its interpretation), in the second.13 In this sense folklore finds itself within the sphere of interest of linguistics.14 Increased attention to linguistic analysis of semantics should connect linguistics with folklore studies even more inasmuch as 11 Therefore, A. Dundes is hardly correct when he shows in his works too narrow an understanding of what is called linguistic structure. From this comes the factual denial of the connection between units of the linguistic and supra-linguistic levels and confirmation that the role of linguistic data consists of reinforcement of folklore structure. See Dundes 1964 : 49. 12 In some measure this error is made in the works of Souriau (1950) and Brémond (1964, 1966a and b). At the same time these works are very interesting as an approach to a calculus of situations and a construction of their general syntax. On the correlation of structural analysis in folkloristics and in literary criticism, see Waugh (1966). 13 Cf. Morris (1938, 1946), not to mention C. S. Peirce. 14 The old attempts to find a correspondence between the genre features of 'simple' folklore forms and their linguistic equivalents are in this sense interesting. Cf. Jolles (1929), where the optative is correlated to the folktale, the imperative to the legend, the interrogative form to the myth (of the newer works, cf. Bausinger 1968); cf. the idea, asserted from the times of Herder, that every word is a work of art in miniature and, vice versa, that every poetic genre is a play on a certain grammatical category (cf. the connection of the lyric, epos, and drama with the subject, object, and verb, or with 1st, 3rd, and 2nd persons, or with sound, word, and sentence), cf. the works of J. Petersen, E. Staiger, and others. Besides this cf. the discussion by Lévi-Strauss of the fact that the myth is a phenomenon of language of a higher level than the phoneme, morpheme, and sememe, but mythemes are units coordinated in one way or another with the sentence; on the relation of the latter to the opposition langue-parole, see Benveniste ( 1 9 6 4 : 2 7 3 - 5 ) (according to Lévi-Strauss, the myth belongs to both langue and parole).

FOLK POETRY: GENERAL PROBLEMS

689

folklore texts are characterized — with respect to semantics — by a whole series of such limitations which essentially decrease the difficulty of semantic description and, moreover, very many cases permit experiments which are completely excluded in literary poetic texts. Consequently, also in this respect folklore texts fix the situation of maximal similarity of purely linguistic structures with elementary esthetic ones. Finally, one must not lose sight of the fact that it is precisely folklore texts which permit recording with the most precision their determinability by general cultural patterns. An explicit or intuitive understanding of these facts explains to a significant degree why in recent years so many specialists in the field of structural linguistics have devoted themselves to studies of folklore. 2. As was indicated above, within so-called structural folkloristics15 the leading area must be considered the study of the folktale.16 Strictly speaking, only here are attempts made to establish 'morphological' laws17 and the practice of using models begins to be instilled. In another article in this volume, problems of structural research of the folktale are treated specially. Therefore, it is important here only to enumerate these problems which, having arisen on the material of folktale analysis, have important meaning both for the study of other folklore genres, in particular poetic, and especially for those which have to do with motifs and their sequences. Among these problems it is proper to name 1) principles of cataloguing and indexing material, internal segmentation of the whole corpus of texts (typology of the genre), cross-genre relationships; 2) general methodological principles: differentiation of the synchronic and diachronic approach with predominant attention to the former, distinguishing syntagmatics and paradigmatics, introduction of the concepts of invariant and variant, constant and variable; 3) principles of formal analysis (morphology and syntax): problems of segmentation of the text, isolation of units of description (motif, function, motifeme, etc.), their distribution, 'heroes' (the structural model of characters), transformations (and the connection between discursive and structural isotopes, i.e. of the diachronic aspect with deep structures), models of the generative type; 4) principles of cultural-semantic interpretation: semantic operators, semantic grids, basic oppositions and mediation, the connection 15 Of the general works devoted to this topic, besides those mentioned above, see Sebeok and Ingemann 1956, Kongas and Maranda 1962, Rohan-Csermak 1965, Waugh 1966, Voigt 1969, Jacobs 1966, etc., and more broadly Viet 1965, Tendances principales de la recherche dans les sciences sociales et humaines. Première partie. Paris-La Haye 1970, etc. 18 This was made possible by most varied circumstances, and among them a special place belongs to the work carried out long ago on collecting folktales, their cataloguing, indexing, and classification, and afterwards also their comparison. Of the over-all works, cf. Aarne and Thompson 1964; Thompson 1938,1946,1955,1955-58,1960; Thomspon and Balys 1958; Thompson and Roberts 1960; etc. There is no doubt that this huge work made possible the appearance of Propp's research and the works which followed him. At the same time the tale's significant (and sometimes even full) independence of ritual and the presence in it of sequences of rather easily identifiable units (at first of motifs, then of functions, etc.) also were conducive to the success in structural analysis of the folktale. 17 For more on them, cf. the article by Nikiforov (1928).

690

V. N. TOPOROV

with the problem of social behavior, the system of values, cultural-ethnographic codes; 5) principles of formalized recording of the text, various forms of representation of the results of the description, working up the material for its analysis with the aid of computers.18 It is characteristic that a large part of these problems was or remains definitive also for linguistics. The problem of cataloguing and classifying comes up both as a result of the significant increase of texts entering academic circulation and because of the necessity of solving new methodological tasks. The increase in quantity of folklore-poetic texts or their more or less qualified editions in recent decades is so indisputable that even the appearance of numerous empirical systems of classification, which permit even preliminary treatment and cataloguing of these texts, cannot evoke surprise. These empirical classifications, undoubtedly, are useful in solving particular questions. Be this as it may, it is these that define at the present time the leading types of collections of folklore materials and coresponding studies: a given genre within the limits of a certain (narrow or broader) ethno-cultural tradition, on a certain territory, within a certain chronological framework, in the light of inter-cultural influences and interactions, the thematic principle, isolation of a new genre from a certain old one, etc. (the same types are discerned when it is a question of a CUMULATION of genres). The usefulness and expediency of such clasisfications are defined by the fact that with their help relationships between at least two series of facts,19 one of which is directly connected with a folklore text, are explained. Strictly scholarly classification sets for itself narrower limits — it establishes the relationships between two (and more, of course) series of FOLKLORE STRUCTURES and, consequently, supposes the ability to differentiate these two (or more) series in the first place, and to isolate in every one of them the accumulation of facts which form (or are understood as) the structure, in the second place.20 Thereby, as a classification of this type are given ideally both the segmentation of the given folklore tradition into genres and the segmentation of the text of the given genre into a definite structure (since the structure is understood as 'le système relationel latent dans l'objet', if one is to follow Lévi-Strauss), which in turn assumes a knowledge of the units, levels, and categories. With reference to folklore material 18

These problems are worked out concretely in such works as Propp 1968 (cf. also Jakobson 1945, 1966:82-100); Propp 1946; Lévi-Strauss 1958, 1960, 1964-68:1-3; Brémond 1964, 1966a and b; Greimas 1963,1966a and b; Herskovits and Herskovits 1958; Bascom 1965a; Colby 1956, 1966; Jacobs 1959b; Dundes 1962, 1963, 1964; Armstrong 1959; Fischer 1960, 1963, 1966; Kôngas and P. Maranda 1962; E. Maranda 1966; Horalek 1964; Pop 1967; 1968; Ch. Vrabie 1965; BeneS 1966-67; Bausinger 1967; Meletinskij 1966, 1970a and b; Meletinskij et al. 1969, 1971; Nathhorst 1969; Weinrich 1970; etc. On experiments in using computers (or other devices of a more or less similar type) in the study of folklore texts, see Bronson 1949; Sebeok 1957, 1961; Sebeok and Zeps 1958; Colby, Collier, and Postal 1963; P. Maranda 1966; etc. On a broader plane, see the articles collected in Hymes 1965; Stone et al. 1962; etc. 18 'A classification is useful to the degree that it sheds light on the relation between one set of facts and another' (Kluckhohn 1960: 134). 10 On the very concept of structure, see Bastide 1962 and Vogt 1960, in order to avoid naming all the numerous works on this topic.

FOLK POETRY: GENERAL PROBLEMS

691

it is assured that such a segmentation is limited not so much by convenience of description as by the natural organization of the whole corpus of texts and of every separate text.21 The meaning of this segmentation is all the more important as it is fully recognized (especially if one speaks of genres) by the bearers of the living folklore tradition. Still more important is what remains unrecognized or, at least, not accepted as a clear formulation — criteria by which one genre is opposed to others. These criteria, to a certain extent coincident with distinctive features in phonology (and even more so on higher levels of linguistic structure), permit definition of concrete ways of cultural patterning in folklore and, therefore, the definitive categories of any given culture. The parallel with linguistics may be continued. Every language possesses a set of units characteristic to it which linguists strive to describe, beginning from some universal set. The universal set itself, however, is corrected by data from all the separate languages known to us. A similar position is observed in folklore between the universal scheme and the schemes of particular traditions. As the corollary of such a state of affairs, it is necessary to consider such cases when a comparison of what seemed to be the same genre in two different folklore traditions turns out to be incorrect for the same reasons for which in linguistics t in one language may be viewed as a phoneme characterized by voicelessness and in another as a phoneme not characterized by voicelessness. Cf. the evidently less than full correspondence of Russ. skazki, Germ. Marchen, Swed. saga, Nor. eventyr, Fr. conte populaire, etc., not to mention more exotic traditions (related to this fact, of course, is the discussion of the relationship of the folktale to the myth in various folklore systems) or the absence in many traditions of correspondences to the genre of ballad, so popular in Anglo-American culture. As a result of the fact that the genre 'space' of every tradition is constructed in its own way, the problem of the differentiation and structural description of genres takes on a greater or lesser keenness. But in any case the problem of genre should not and cannot be eliminated or dispersed in a general analysis of style and structure. In this sense the polemics of L. Honko (1968) against the neglect of genre analysis in many works by American and English folklorists is very characteristic. First of all, disagreement is expressed with the assertion that genres may be identified only after an analysis of form and structure on which, in fact, the identification of genres is based (Dundes 1965:127).22 There is, of course, no reason to doubt that an analysis of structure presupposes very serious and reliable criteria for differentiating genres, but with this analysis (more accurately, before it) genres may be differen21 In any oral literature, writes M. Jacobs, 'careful analysis of classified portions of the material displays numerous structurings and multiple determinants of varying weights. Patterns enclose patterns. Structurings intersect, contain, and shape one another. Exacting classification and intensive analysis of all aspects of a literature, following their identification, will show the way, if the research is done without minimization of sociocultural factors' ( I 9 6 0 : 1 2 9 ) . 82 Cf. the definition of genre in Wellek and Warren 1956:221): 'Genre should be conceived as a grouping of literary works based, theoretically, upon both outer form (specific meter or structure) and also upon inner form (attitude, tone, purpose — more crudely, subject and audience. . . . ' .

692

V. N. TOPOROV

tiated within a given tradition. In any case the point of departure in a genre analysis is the idea that texts of any oral tradition are not 'a uniform mass as far as its nature and information value are concerned' (Honko 1968:50), 23 despite the attempts of many researchers (especially if it is a question of traditions of 'primitive' peoples) to designate all types of oral works by one or several supra-genre labels.24 The special structure of every genre and the special extra-folklore (general cultural) motivation of it makes it possible for the genre analysis to lead to 'the thorough classification of the whole material of oral tradition' (Honko 1968:51). In a preliminary classification of genres it is appropriate to trust the local native traditions with their particular nomenclature (cf. B. Malinowski's 'Listen to the natives!') and also the analysis of situations in which the given genre gains importance and attains its higher value.25 An exhaustive analysis of genres presupposes use of the following criteria: 1) content, 2) form, 3) style, 4) structure, 5) function, 6) frequency, 7) distribution, 8) age, and 9) origin (Honko 1968:62). V. Ja. Propp, also stressing that the concept of genre is a unit of classification (inasmuch as, by the way, it reflects the action of certain more general categories, on the one hand, and is perhaps broken down into smaller categories, on the other), indicates somewhat different criteria for identifying genre (poetics, daily use, method of performance, relation to music), which, nevertheless, to a significant degree can be reduced to the criteria mentioned above (Propp 1964a and b). In other general works on the classification of folklore genres somewhat different sets of criteria are proposed (Boggs 1949, 1950; D'Aronco 1963). Every one of them is real within the limits of the given tradition and, therefore, cannot be ignored in creating a universal scheme of criteria for differentiating genres. However, one gets the impression that searches for a maximal quantity of genre-distinguishing criteria within the limits of a given tradition frequently lead to redundancy, concealing the structural characteristics of genre and the hierarchy of the criteria. Therefore, for the time being it is appropriate to give preference usually to the more specific but incomparably stricter classifications constructed on the basis of identification of really relevant features. Such a dual scheme for narratives, in which the different meanings (yes and no) of the two pairs of features, fabulous / factual and secular / sacred, permit differentiation of folktale, myth, history, and sacred history, and, opposed to all of them, the legend (or saga) (Littleton 1965 : graph 1); it is interesting that such characteristics as age seem in this case to have been derived from the results of the 23

Cf. also Sydow ( 1 9 4 8 : 6 0 - 1 ) : 'Die verschiedenen Überlieferungsarten unterstehen völlig verschiedenen Gesetzen und ohne Vertrautheit mit dem System der Überlieferung, dem gegenseitigen Verhältnis und den ungleichen Lebensbedingungen der verschiedenen Kategorien usw. kann man zu keinen zuverlässigen Forschungsergebnissen gelangen. . . . ' 24 Cf. Fischer (1963 : 236-7), where the term folktales covers not only tales, legends, and myths, but also almost the whole epic tradition. Otherwise, see Hermann (1958 : 73ff.). 25 Cf. the differentiation of the three genres of narrative prose among the Kiriwina (Malinowski 1922 : 299ff.); or the differentiation of myth and legend among the Nootka Indians (Sapir 1959); etc. Regarding genre nomenclature (extremely diverse) in the European tradition, see B0dker (1966).

FOLK POETRY: GENERAL PROBLEMS

693

binary classification (graph 2, p. 26).26 In part similar principles are also used by W. Bascom when he constructs a quasi-matrix for distinguishing the three types of narrative prose: myth, legend, and folktale. Various meanings of the categories Belief, Time, Place, Attitude, and Principal Characters serve as criteria for differentiation. Within each category two or three meanings are differentiated (the third meaning is in fact zero in its most accepted conception in phonological matrices): cf. Fact — Fiction; Remote Past — Recent Past — Any Time; Different World (other or earlier) — World of Today — Any Place; Sacred — Secular — Sacred or Secular; Human — Non-Human — Human or Non-Human (Bascom 1965a).27 A classification system of oral traditions which give historical content is constructed by J. Vansina (1958:58ff.; 1961:Table III, p. 120),28 who uses for this categories of a less operationalist character. Special significance is accorded those works in which attempts are made to take into account the ENTIRE genre wealth of a given tradition and to give criteria for their differentiation. This last task is more complicated and its solution often remains unclear even when the genres are known. Not long ago a two-fold test was proposed for the solution of this problem with reference to Gilyak folklore. With the help of this test it was possible 'to arrive at a consistent organization of the corpus which, it is hoped, will also accentuate those features of the genres which deserve additional or special attention' (Austerlitz 1961a). This test proposes two axes — music, by which is understood phrases of sung melody (= variation in pitch along a scale) which are capable of varied repetition and which are linearly adjusted to a rhythmical pattern (= restrictions on lengths of sung pulses and of pauses), and response, the verbal participation by one or more members of the audience in the performance of the particular folkloristic work of art. In taking into account some supplemental details, a matrix of the following type is constructed: Response —

—

Music

+

+

tale

riddle

song

epic ditty

Naturally the analysis may continue at the expense of research of typically 'poetical' 26

The researcher of religious texts also lacks the criterion of trustworthiness. Bascom's indication is important that 'myth, legend, and folktale are not proposed as universally recognized categories, but as analytical concepts which can be meaningfully applied cross-culturally even when other systems of "native categories" are locally recognized' (1965a : 5). 28 Cf. also the works devoted to the differentiation of various types of historical prose (Sydow 1948) and first and foremost Memorat and Sage (Honko 1964, Pentikainen 1968). 27

694

V. N. TOPOROV

devices. It is interesting that in the case when there is a description of all genres of a given folklore tradition and within this case a local nomenclature of genres and/or the conditions of their existence are known, it is possible to establish a system of distinctive features of genres which turns out to be fully satisfactory, at least in an operational sense (cf. the many studies or collections describing the genre structure of separate traditions, for example, on African,29 American Indian,30 Oceanic (e.g. Davenport 1953), North Asiatic,31 etc., material). It is characteristic that the exposure of this system of features is easier the more ritualized the genres are, i.e. the more deeply they are included in the general contexts defined by the action of cultural patterns. Exactly under these conditions, when every genre can be represented as a bundle of distinctive features, it is expedient to speak not of a SET of genres of a given tradition but of their SYSTEM, inasmuch as the distinctive features themselves assign the structure of the genre space and the possible features of the hierarchical organization. It is not excluded that in some cases the introduction of QUANTITATIVE characteristics as well will be expedient for defining the place of a given genre in the given tradition (or the distance between two genres in the genre space). In any case it is completely evident that the genre space of the folklore tradition (by the way, this can relate also to literary tradition) should be characterized by such criteria as the degree of its differentiation, the degree of depth (the distance between the most and least specialized genres in terms of distinctive features), the degree of versality (the quantity of genres capable of transmitting a given concept, the problem of synonymy), etc. The representation of the genres of a given tradition as a system not only permits definition of the solving strength of this tradition and the corresponding load of every genre, and, consequently, of details of the functioning of this system, but also creates the conditions for solving the tasks of diachronic folklore studies, allowing one to turn, in particular, to internal reconstruction of the genre space (which to a certain degree recalls analogous tasks in linguistics). Unfortunately, folklore research in the field of genre clearly uses insufficiently the wealth of experience accumulated in this respect by the study of literature. The theoretical importance of the problematics of genre is explained by the fact that genre is based on the structure of over-all human existence (which is reflected in no small way in the structure of every separate text), that because of this they 'may 29

Cf. the numerous works cited in the valuable bibliography by W. Bascom (1964) (cf. also Bascom 1965b); and, in part, Kotljar 1970. 30 Here it is sufficient to refer to both the studies and collections of folklore texts (with faithfulness to local traditions) prepared by F. Boas, E. Sapir, A. L. Kroeber, and their numerous successors. 31 Cf. the folklore materials on Paleo-Asiatic languages (V. G. Bogoraz, V. I. Ioxel'son, L. Ja. Sternberg, E. S. Rubcova), Altaic languages (G. Huth, G. M. Vasilevic, M. G. Voskobojnikov, K. A. Novikova, V. A. Avrorin, E. P. Lebedeva, I. A. Lopatin, B. Rin£en, V. V. Radlov, I. A. Xudjakov, E. K. Pekarskij, G. U. Ergis, N. A. Baskakov et al.), the Ket language (A. P. Dul'zon), Samoyedic languages (M. A. Castren, T. Lehtisalo, Z. N. Kuprijanova, G. D. Verbov, B. O. Dolgix, et al.), Ugric languages (A. Kannisto, S. Patkanov, et al.).

FOLK POETRY: GENERAL PROBLEMS

695

be regarded as institutional imperatives which both coerce and are in turn coerced by the writer'.32 Therefore, in folkloristics, as in the study of literature, it is appropriate to recognize as erroneous attempts to see in the problem of genres something non-essential or relevant only to classification.83 It is possible that the folkloristic theory of genres would receive additional basis with the introduction of some more general concepts of stylistic (conditionally speaking) character, cf. lyrical, epic, dramatic in E. Staiger's conception. The introduction of these concepts permits the establishment of a more direct tie with the system of distinctive features (than in the case when it is a matter of genres as previously given forms) and to indicate more concrete ways of generating the given genre. With such an approach it turns out that to give a linguistic (or formal) characterization is easier for these general concepts than for already prepared genres, each of which as a special type can be deduced from these concepts or considered the result of their interaction. From this comes the more characteristic case when the type occurs in the text not in its pure form. Of course, in folklore the problem of genres takes on a simpler aspect than in belles-lettres, but even in folklore one constantly meets with texts which with respect to style can be considered creolized (cf. the combination of epic and lyric beginnings in the ballad, dramatic and epic in the folktale or in complex types of riddles, lyric and dramatic in many varieties of songs, etc.). This creolization is most fully and concretely manifested on the level of the structural organization of the text taken in its entirety. In this sense it is characteristic that both in folkloristics and in the study of literature the ideas of 'type', 'metamorphoses', 'mobile structure', etc., going back to Goethe (cf. Propp's morphology of the folktale, on the one hand, and the morphological study of literature from the beginning of the '40s34), begin to take on a special meaning. It is understandable that both in folkloristics and in the study of literature (as a consequence of the methodological arrangements described above) the point of view begins to be confirmed that empirical studies in the field of classification of genres cannot lead to working out the concept of TYPE in its modern interpretation,35 in the first place, and that the prob32

See: Pearson 1941:59ff. (cf. Wellek and Warren 1956:235); Donohue 1943; Wehrli 1951; etc. Cf. the discussion on this question: Actes du 3-e Congrès International d'histoire littéraire, Lyon, 1939 (1940:95ff.); Tïeghem 1939; Staiger 1946; etc. 34 Cf. Mûller 1943, 1944; Oppel 1947; and other works in this direction. Cf., however, the reactions: Staiger 1944. 35 Generally speaking, it is a question of transition from the old 'classificating' principle of defining the type to the new principle of 'stepped' (abstufbar) formation of similar concepts. The latter consists of the fact that the given type is described not simply classificationally, but its place in 'feature space' (Merkmalraum) is indicated, each measurement of which is limited by the boundary of one of the features (cf. the analogous picture in physics, where the spatial position of a point is defined by means of 'stepped' coordinate concepts, and in linguistics, where a similar tendency finds reflection of these or other concepts). Such an approach makes possible a more rigorous description of the types with diffuse boundaries and of various types of transitional formations since it presupposes a systemic point of view and permits the introduction of topological methods of research where earlier exclusively metrical methods were used at best. Cf. in addition the pre-war work by C. Y. Hempel and P. Oppenheim (1936: especially pp. 66ff.). 38

696

V. N. TOPOROV

lem of genre classification is a TYPOLOGICAL problem, in the second place. In other words, in folklore, too, there arises the same problem which arose in linguistics rather long ago — about the correlation between particular and general descriptions (cf. the correlations of the set of distinctive features of the phonemes of a given language and the universal set of these features). Perhaps more keenly than others this problem was posed by B. Nathhorst (1968) in his disagreement with the definition of the riddle, universal in idea, by A. Dundes and R. A. Georges (1963). Taking into acount that formally similar texts in different traditions are included in different cultural contexts (for example, the song may enter into ritual and be an object of faith in one culture and be considered something improbable and serving for entertainment in another culture), it is necessary to agree with the general conclusion from this debate that it is better to set aside for the time being the demand for 'universal validity of generally valid definitions and categories'. If one turns to a typology of oral-poetic genres and their more detailed segmentation, it will turn out that the studies of recent years which should be considered the most promising are those in which 'mobile' (in various respects) cases are analyzed. By the latter are meant those examples which presuppose another member of the comparison and a relationship of greater or lesser proximity between the two terms of the comparison. Such 'mobility' is created in the case of the transitivity of a given genre (under conditions of its recent isolation from some other genre or tendency toward identification with it in the near future), the presence of two (or more) interacting cultural areas, of two (or more) chronological cross-sections fixing various stages in the evolution of the genre, changes in the medium where a given genre has currency, etc. Very characteristic in this connection is the predominant interest (if one speaks about oral-poetic folklore) in the ballad and related formations against a background of research of the poetic epos and songs. If with respect to the epos the basic attention of specialists is attracted by (essentially) historical problems, and in the structural aspect predominantly by the topic of formal structure, then with regard to the song (besides works on poetics) their attention is given to a PRACTICAL classification of types, most often serving cataloguing purposes. Since (generally speaking) the ballad occupies an intermediate position between the epos and the song, its indexing and cataloguing is more complex, and, in particular, it presupposes a preliminary analysis in which essential features of one genre are disclosed. In this connection G. List's research (1968), in which an indexing of ballads is set forth by means of the so-called 'dramatic-narrative elements', is indicative. In part, elements of a similar type figured already in the classifications by T. P. Coffin (1950, 1960, and, with M. E. Leach 1961) and G. M. Laws, Jr. (1957, 1964; cf. Abrahams 1966) and even F. J. Child (1882-98), where they, however, were neither consistent nor the sole element on which the classification was based. Therefore, List identifies the five-membered scheme into which enter: 1) exposition, 2) episode, 3) development, crisis, and solution, 4) episode, and 5) secondary crisis and solution. Ballad plots are considered as something which form 'an uneven

FOLK POETRY: GENERAL PROBLEMS

697

and slowly changing continuum which nevertheless show certain polarities'; a comparison is proposed of corresponding types of ballads with similar plots with the aim of defining 'a series of disparate plot gists which will form the principal basis of the system of classification' (List 1968 : 60). The choice, for purposes of classification, of plot gists (instead of the plots themselves) is based on the special role of 3) development, crisis, and solution; in particular, in the transition from variant to variant. Finally, it is proposed that 'the dramatic-narrative elements found without the plot gists be indexed as independent motives'. Be this as it may, the system of motifs in such an approach is limited more strictly than in S. Thompson's MotifIndex. Along with this, List's classification is, of course, much indebted to corresponding works in the area of structural analysis of the folktale. The cited works of T. P. Coffin and G. M. Laws, Jr., bear a more practical character. In this respect they represent a high example of which — for the striking majority of traditions — there are not even any imitations. In connection with the given topic it is more appropriate, however, to turn attention to a description of variations in traditional American ballads. T. P. Coffin (1950:3ff.) points out that 'Forgetting, contamination of one ballad by stanzas from another, use of a cliche to fill in forgotten material, desire for more dramatic effects, tendency to rationalize unbelievable situations, use of localisms, invention of new story matter, misunderstanding in the oral transmission of a phrase, and the adaptation of old words to a modified or new tune, are all means by which the individual singer may change a ballad.' An analysis of all these cases is essential in the respect that it leads us to the problem of a change in the elements of the structure of the genre and the boundaries of the genre in transition from one cultural tradition to another. The same problems arise in an analysis of new genres in one way or another connected with the ballad, of the influence of literary, professional, regional, and other factors (cf. L. Shepard 1962, A. B. Friedman 1961a, P. Brand 1962, N. V. Rosenberg 1967, etc.), and also in an analysis of balladic treatment of historical or contemporary material yielding to more or less exact dating (cf. e.g. Paredes 1958, among others; Wilgus and Montell 1968; Vargyas 1960-61, 1960-62; Putilov 1960, 1965; and many others). A special problem — about the correlation (and dependence) between the text and the melody — also has relevance for the described topic but will not be specially considered here. 38 Many of the enumerated questions with special stress on the diachronic line of development are discussed in studies of the monographic type, becoming more and more popular, which are devoted to the analysis of a certain ballad or some group of them. Such a study as J. Kemppinen's on the ballad 'The maid freed from the gallows' (Child 95), where the author, having analyzed 1,634 variants of the ballad, comes not only to new conclusions of a his-

38

Bronson 1944, 1946, 1959, 1959-62; Boswell 1967; Herzog 1949; etc. Also, the catalogues of melodies are not considered, although successful patterns do exist (U.S.A., England, West Germany, Finland, Lithuania, Hungary, the Ukraine, etc.).

698

V. N. TOPOROV

torical-geographical character, but also proves that as a basis for this text, treated as a historical ballad at one time, was a mythological ballad which arose on a base of old mythological cultures (Kemppinen 1957); of the book of the same author (1954) on the famous ballad 'Lady Isabel and the False Knight' (Child 4) executed in the same spirit. No less known is the thorough study of the ballad of Heer Halewijn, done by H. O. Nygard (1958, 1965). Using the technique developed by A. Taylor (1931) and J. Kemppinen, the author analyzes and catalogues all the variants available to him, defining the national form of the ballad in four areas (the Netherlands and Germany; Scandinavia; France and French Canada; Great Britain). Although the author constantly has in mind a large set of ballads which are compared with this one, his main aim is not this comparison itself and, moreover, not a search for the prototype but 'to contribute toward our fuller discovery of the nature of ballads and the internal dynamic of balladry by the profound and comprehensive study of a single ballad'. Very significant is the recent work by L. Vargyas (1967) on the medieval history of folk ballads. In it are considered not just one ballad, as in the preceding studies, but several cycles (cf. Ch. 1, 'The originally French stratum in Hungarian ballads'; Ch. 2, 'The survival of the heroic epic of the Hungarian conquest period in our ballads'; Ch. 3, 'The origin of the "Walled-up Wife'" [Vargyas I960]), in each of which a consistent stratification of the original heritage and borrowed elements is carried out. But the most valuable result of the study should be considered the establishment of the fact that the ballad in the process of historical development supplanted a heroic epos; evidence of this is given by, among other things, the fact that in the Middle Ages ballads were popular first of all where (and in that environment, predominantly peasant) the heroic epos had already disappeared. Consequently, in these areas and in this social layer in the Middle Ages, it was the ballad which became the universal poetic genre in which that content (or, it is better to say, cumulation of motifs) which was earlier connected with the heroic epos or heroic song could be expressed.37 L. Vargyas himself emphasizes the importance of establishing the direct sequence of genres on the example of an analysis of similar motifs. In this manner, the problem of synchronic research of the genre system in the given tradition is supplemented by its diachronic aspect. The author tries to understand the reasons for the transformation of the genre which led to the rise of the ballad: what is new in it is that when PSYCHOLOGICAL problems and SOCIAL situations appear, they do so as a nucleus of events, as reasons for dramatic collision; in other words, there arise 'problems of man in society, of relations between men, and of their social positions' (1967 :242).88 The model research of Swedish 57

Many comparative-historical studies fix the process of transformation of early heroic songs into the ballad. See Menendez-Pidal 1936; Meier 1941; Seemann 1955; etc. On ties in another direction, cf. Siuts 1962. 38 For a typology of the ballad, Table VIII is of value in Vargyas 1967 (cf. also 1964). On the typology of the ballad in an adjacent (Rumanian) area, see G. Vrabie 1957, 1966, 1970 : 300ff.

FOLK POETRY: GENERAL PROBLEMS

699

ballads (with an indication of their types, sources, with a description of skillingtryck, etc.) was done by B. R. Jonsson (1961, 1967). There are successes in cataloguing and genre specification of the ballad also in those traditions where formerly these problems were found on the periphery of research interests.39 It is not accidental, therefore, that more and more often the question of creating an index of ballad types for a more or less wide and heterogeneous area (R. M. Brednich 1968) is raised. The advancement of such a goal, undoubtedly, is evidence of a large preparatory work carried out in the field of ballad study. The creation of an index of this type would be an event of epochal importance. Moreover, its results would have relevance not only for folklore (in particular, for studies of epic and song) but also for belles-lettres (it is sufficient to remember the role of the ballad in European romanticism). As has already been said, studies in the field of the song genre are less indicative since the song represents those oral-poetic texts where the lyrical beginning (cf. supra) is manifested to the greatest degree, lending to the whole genre a maximal homogeneity. Of course, there are intermediate cases, but they usually have supplementary (often rather limited) means for their identification (cf. the problem of so-called historical songs or 'episch-lyrisches Lied' — Braun 1963; Schmauss 1959) and, therefore, do not give cause, as a rule, for posing theoretical questions relating to the genre space of songs. Naturally under these conditions the problem of segmentation of this space leads to a more or less empirical classification. In many traditions where the song remains a living folklore genre and is motivated by general cultural patterns, songs are segmented in accordance with the extra-folklore phenomena which determine them (the type of work which accompanies singing, seasonal holidays, rituals, etc.), which more or less strictly define also the contents of the corresponding songs (their thematics). From this comes the most widespread type of classification of songs on the basis of their function and thematics (for example, working, mythological, calendar-ceremonial, wedding, christening, family, children's, love, drinking, and other songs, divided further into subgroups according to the motif-thematic principle).40 Along with this are also proposed classification

('Balada populara'); Chifimia 1957; Caracostea 1948; and the numerous analyses of individual ballads (first of all about Manole, Miorija, etc.; cf., for example, Brailoiu 1946; Spitzer 1959; Fochi 1964; Eliade 1970; etc.; on the adjacent Ukrainian area, see Kirdan 1962,1965; Lintur 1963; etc.; on the Slovak area Horak 1956, 1961; studies in the field of 'robber' folklore, cf. Melichercik 1952, 1956; Bogatyrev 1963; Horak and Plicka 1965; etc.; on posing the question of the ballad genre in Scandinavian folklore, see Dal 1962. 38 Cf. Balys 1958 (every song and ballad — there are 472 in all — is marked with a number corresponding to the systematic catalogue prepared by J. Balys); Ciurlionite 1966; Jokimaitiene 1968. Numerous works on the Russian ballad have appeared in recent years. Besides those mentioned, cf. also D. M. BalaSov 1962, 1963, 1966; Putilov 1964; etc. 40 Cf., as an example, Barauskiene 1968; Kazlauskiene 1959, 1967; Jokimaitiene 1967; etc.; if one is to speak of a song tradition as archaic as the Lithuanian one. Cf. also Laforte 1958; Propp 1961; Kolpakova 1962:11-29 ('Voprosy klassifikacii'); and esp. Erdely 1962; etc.

700

V. N. TOPOROV

schemes which are oriented on a very short (of the nature of a word and not of a sentence) description of the theme, reminiscent in some respects of the procedure in 'content analysis' (cf. J. Sadownik 195641). The very appeal to the theme as a principle of classification is not coincidental. Undoubtedly, the THEME in a song is to a certain extent equivalent to the ACTION in the folktale, and the whole question is whether the researchers will succeed in uniformly and economically isolating the themes and their succession and configuration. Only in the case of success for the song as well will an instrument of analysis be obtained which will be as effective as the motif has been or the function for the folktale. Another analogy with the study of the folktale consists of the fact that with reference to songs there is the same problem of identification of variants that appears in connection with the folktale. However, here it is aggravated by the circumstance that the variety in songs is immeasurably richer than in tales, which results in a huge quantity of transitional types, sometimes not yielding to identification. From this comes the task of describing the changes in the text of songs which lead to the appearance of new variants. Different types of changes can become criteria for cataloguing songs (Havlikova 1967), but the analysis of the varieties itself permits researchers to approach a more adequate understanding of the boundaries of song types. However this may be, the problem of genre specification of poetic folklore texts should lead specialists to a recognition of the morphology of these texts with the same fullness and rigor as in the recent studies of the folktale. The establishment of a system of genres within a given tradition (and consequently of the rules of transition from one genre to another) and the successes in the typological study of genres undoubtedly suggested to researchers the goal of cross-genre analysis, understood as the definition of genre structures (transforms) conveying an assigned content. It may be hoped that ideally such analyses may be constructed as calculations from which ensues the possibility of prognostication in folklore and cultural-anthropological research. Cross-genre analysis, the importance of which has already been noted (Dundes 1963), is called upon to play, with respect to the CONTENT of folklore forms, approximately the same role as transformational analysis in linguistics, which permits rigorous description of the meaning by indicating the set of syntactic transforms. In any case, in folklore studies (as also in linguistics) transformational theory again leads the investigator to semantic problems but already on a more formal and, consequently, more easily verifiable basis. By the same token, it seems, the lacuna could be filled between folkloristics of the descriptive school and culture-and-language studies. It should be said that representatives of the historical-geographical school in folklore studies have in their practical studies more than once turned to cross-genre analysis when the fate of a text which a given content, reflected in secondary texts of DIFFERENT genres, inter-

11 A clarification and change of principles of an earlier classification of Polish songs given in a series of works by J. St. Bystron.

FOLK POETRY: GENERAL PROBLEMS

701

ested them.42 However, attention in such cases was directed at the preservation of the content and not at the rules of transformation in the cross-genre movement of the text. In a manner similar to the way cross-genre analysis defines the boundaries of a given genre FROM WITHOUT, i.e. from the point of view of the whole system of genres in the given tradition, the analysis of different variants of texts of a given genre defines the internal boundaries of a given type of texts. The latter leads us to the problem of the invariant in folklore material, i.e. to the second cycle of questions mentioned above. 3. While not dwelling particularly on this series of questions, since existing works on oral-poetic folklore only to a slight extent touch upon these problems, it should be stressed that these problems will undoubtedly become more urgent as more successes are attained in the structural analysis of texts of corresponding genres. The question on the differentiation of synchronic and diachronic aspects of the study of folklore texts is in principle clear, although the simple extrapolation of the relationship of these aspects from linguistics is hardly expedient. The fact of the matter is that the researcher most often has to do with folklore texts which have already gone beyond the limits of their original locale and, consequently, have lost those sufficiently influential or even compulsory motivations from the extra-folklore system (e.g. ritual) which earlier defined the structure and semantic interpretation of the text. It is natural that in this new locale (e.g. not ritual but a sphere of entertainment) the text is subjected to a somewhat different structuralization and different semantic interpretation. Therefore, for texts of this type (with a changed locale) a minimum of two independent analyses must be constructed 43 which in concrete studies are very frequently merged into one heterogeneous description. Along with this a limitation of the research to texts only in their original or only in their secondary locale would be a sensible limitation and would lead to too strong a simplification. Of course, the structural approach to folklore texts is oriented on a synchronic consideration of them. Numerous analyses lead to the isolation of a set of elements and the setting up of the pecularities of their distribution in the text, including a 42

Nonetheless, existing indexes of folktale motifs may be reproached for ignoring other genres in which a given motif is reflected. Strictly speaking, there exist motifs for which the original locale may be defined, but one cannot with certainty establish their present genre ties. Cf. A T 365 (Lenore) in the variant 'the return of the dead brother' in the Balkans (myth, tale, ballad, song, tradition) or some texts of a cumulative type which are reflected in tales, songs, carols, series, riddles, jokes, etc.; cf. Toporov 1963; cf. also cross-genre analysis in folktale and superstition among the Zuni in the mentioned article by Dundes (1963 :128). 43 Cf. with regard to the ballad the differentiation of the diachronic axis 'des relations, qui s'établissent entre la succession des motifs' and the synchronic axis 'des relations, qui s'établissent entre les différents schémas selon lesquels le même motif se répète'; see Eretescu 1966. Besides all this it should be remembered that the differentiation of synchrony and diachrony in poetics does not fully correspond to what is observed with respect to linguistics. E. Stankiewicz (1961) focuses attention on this.

702

V. N. TOPOROV

differentiation of the compulsory (with strict determination) and more or less free sequences. These elements and their combination do not fully coincide with those blocks with which traditional folkloristics operates (tropes, figures, formulas of various types, etc.), although the latter most often can regularly be deduced from the elements obtained in a structural analysis. In this way, studies of folklore texts and, in particular, of oral-poetic ones, repeat the same course taken in works on the folktale. It may be said that for present studies in the field of folklore (besides the tale) one of the most vital tasks is the discovery of INVARIANTS, which amounts in practice to working out the formal procedure for defining the variants of a given text and for reducing them to the invariant. One can speak in this manner of the situation in which the study of the folktale found itself before the appearance of Propp's book. True — the task of finding such invariants in the epos, ballad, song, etc., is incomparably more complicated; there is no guarantee that the invariants which are to be found will, in the first place, turn out to be comparable as to the concreteness-abstractness feature with the FUNCTIONS in the folktale and, in the second place, that they will permit realization of a description of oral-poetic texts which will be as adequate as in the case of the tale. In any case, without preliminary cataloguing of epic, ballad, song, and other texts, the search for invariants can be conducted only gropingly since it always remains unclear what to consider accidental and what typical. Evidently for texts with plots of the oral-poetic tradition, invariants may be thought of as something comparable to the functions of motifemes in folktales (Bowra 1961:178ff.; Chadwick 1969; Meletinskij 1963; etc.), although rules for their combination in concrete texts (syntagmatics), as well as their set itself (paradigmatics) will undoubtedly differ significantly from what is known from the analysis of tales. Unfortunately, studies of concrete texts in this respect are insufficient. For poetic texts of song nature (if it is not a question of .songs with plot) the solution to the problem of the invariant is connected with additional difficulties. Also such a solution is not excluded, at least for many old traditions, that all song texts connected with one definite event (or seasonal, working, family, or other cycles) are found in a variant relation, and it is, therefore, possible to speak of ONE song-invariant. In any case, acceptance of this solution would not be contradicted by the numerous studies on collecting and cataloguing song variants. Quite often it turns out that, within the limits of a given cycle of songs which have a certain coordination in time with some (extra-folklore) event, all possible song variants are possible which can be generated by the available ready-made cliches with a maximal freedom of distribution and that, consequently, there is some continuum which does not yield, strictly speaking, to separation into parts which would be relevant from the point of view of structural analysis.44 For those traditions where this is actually the case, only the relation of such invariant44

In any case, the reduction of all songs of a given cycle to one is much more justified than is the same operation with tales, for which Lévi-Strauss reproached Propp. A similar problem is considered demonstrably in the article by Amzulescu 1966; cf. in part Birgu-Georgescu 1966.

FOLK POETRY: GENERAL PROBLEMS

703

schemes of the song among themselves is significant, each scheme being strictly motivated by extra-textual reality. With this problem of the invariant and identification of the units of the text's structure with reference to folk poetry is directly connected the relationship to the role of the syntagmatic and paradigmatic analysis. The meaning of both in general principle and in the analysis of concrete texts was elucidated more than once (cf. R. Jakobson, Th. A. Sebeok, R. Austerlitz, S. R. Levin, et al., see below). Therefore, it is important to point out here, remembering the controversy between C. Lévi-Strauss (1960) and V. Ja. Propp (1966: Appendix) on this matter,45 that the choice between syntagmatics and paradigmatics essentially depends on the structure of the analyzed material and on the goals which the researcher places before him. With sufficient completeness of material and striving for a maximally full description, an analysis of both syntagmatics and paradigmatics is no doubt necessary. Moreover, as also in language, here the possibility is revealed of coordinating in principle the results of both analyses, having established the connection between syntactic functions and paradigmatic relationships (e.g. Winter 1959). From this, however, it does not follow that the problem of the practical advantages of one or the other approach has no bearing. It is understandable that LéviStrauss's preference for paradigmatics was justified by this interest in the myth, in the mythological system, by the set of logical operations standing behind it, and, finally, by requirements of the nonlinear structural model interpreted in the spirit of Boolean algebra. On the other hand, Propp's syntagmatic emphasis was predetermined by his interest in the morphology of the folktale. With reference to oral-poetic texts, ideally it would be important to have both paradigmatic and syntagmatic descriptions and the rules of their correlation. Probably the differences among texts in both descriptions could turn out to be an important typological criterion for the classification of these texts. Thus, in advance it could be thought that poetic forms with reference to a plot, syntagmatic analysis gives an essentially richer material than with reference to the plotless song with maximal freedom of distribution of its component parts. In exactly the same way paradigmatic analysis of the mythological parts of the epos or of mythological songs in a limited series of corresponding texts rather easily permits discovery of the system (the rules for generating texts from this system can remain partially unknown) which is only partially and unauthentically reconstructed on the material of maximally demy thologized texts. Finally, giving attention to both approaches is essential because it can theoretically be proposed that such cases are present when from THE SAME paradigm texts with TWO DIFFERENT types of syntagmatics can be generated (for many traditions such a case is actually evidenced in myth and folktale). Unfor45

See also the works treating this topic by E. Leach, A. Dundes, B. Nathhorst, H. Weinrich, E. M. Meletinskij, et al. For general results, see Meletinskij 1969, 1970c; etc. The attempt was undertaken by A. J. Greimas (1966b, 1970) to reconcile the syntagmatic and paradigmatic approaches.

704

V. N. TOPOROV

tunately, in concrete studies of folk poetry the problem of syntagmatic and paradigmatic approaches either is not posed in general or is reduced to a description of the purely linguistic elements of the text (frequently even without indication of how they are used on supra-linguistic levels). This has evoked the criticism of many specialists. As concerns handling this problem in connection with the general topic of genre typology (in particular, oral-poetic genres), the whole basic work lies ahead. Although future results remain less than clear, their importance, as also their significance for the research of more complex texts (for example, of individual poetic texts) and for an elucidation of the relations between linguistic and supralinguistic levels, is obvious. As a conclusion to this section it is appropriate to point out that the connection of syntagmatics with paradigmatics cannot be excluded because every poetic work generates its own code (cf. the generation of paradigms in syntagms and of syntagms in turn by paradigms); 46 in other words, 'the poetic function projects the principle of equivalence from the axis of selection onto the axis of combination' (Jakobson 1960:358). 4. Turning to the third series of problems (cf. p. 689), which arise at the intersection of structural linguistics and folklore studies, it is appropriate to say a few words about the problem of poetic language which supposes that every poetic work is in a special way organized language. In its most conclusive aspect, this conception, which takes its beginning from the Russian formalists,47 was developed by R. Jakobson and further by his successors. Proceeding from a scheme of components, obligatorily present in each act of verbal communication: Content Message Addresser

Addressee Contact Code,

a system of the functions of language is constructed which is isomorphic to this scheme: Referential Poetic Emotive Conative Phatic Metalingual. From this it follows that the poetic function is formed by the orientation to the message for its own sake, and poetics is defined as that portion of LINGUISTICS which studies the POETIC function in its connection with other language functions 46 47

See Levin 1962 : 41; cf. the analogous thoughts of P. Valéry. Cf. especially Tynjanov 1924, on the unity and closeness of the verse sequence.

FOLK POETRY: GENERAL PROBLEMS

705

(Jakobson 1 9 6 0 : 3 5 3 - 9 ) . Other studies in this direction are constructed either as a more detailed definition of language structures in poetry and an explanation of what exactly comprises the special organizational principle of the poetic text (in connection with which at least two series of problems arise: correspondence of the syntagm and paradigm and the topic of EQUIVALENCY with respect to both of them) 4 8 or, on the other hand, as a sort of establishment of the differences between poetic and non-poetic language (including the language of science). 4 9 In a somewhat different light this same series was elucidated in two other directions, likewise yielding — directly or indirectly — to the influence of linguistic ideas. I have in mind the development of de Saussure's ideas in stylistics of the linguistic character (cf. L. Spitzer 1948, 1949) and of those of the American linguistic and semiotic tradition in I. A . Richards' Principles of literary criticism and the corresponding school. Finally, the tendency to consider the literary work as 'ein SPRACHLICHES Kunstwerk', as a self-contained linguistic work in various affiliations of Husserlian phenomenology in literary studies (E. Staiger, W. Kayser [ 1 9 4 8 , 1958], et al.).6« In this way, different trends in literary studies more and more agree that in a 48

Levin 1962 (cf. on couplings, structures which are characteristic of poetic language and which function in the direction of unifying the text in which they appear; on paradigms and positions and, in connection with them, on the two types of equivalency and their relations to poetic language); Levin 1963 = "Statistische und determinierte Abweichung in poetischer Sprache", Mathematik und Dichtung: Versuche zur Frage einer exakten Literaturwissenschaft, pp. 33-48 (München, 1965) [cf. in the same collection several other articles which are interesting from the point of view of this problem: H. Kreuzer, "Mathematik und Dichtung", 9-20; M. Bierwisch, "Poetik und Linguistik", 49-66; K. Baumgärtner, "Formale Erklärung poetischer Texte", 67-84; W. Fuchs and J. Lauter, "Mathematische Analyse des literarischen Stils", 107-22; N. Ulrich, "Uber ein mathematisches Modell zur Bestimmung literarischer Stillkomponenten", 185-92; L. Dolezel, "Zur statistischen Theorie der Dichtersprache", 275-94]; Kurylowicz 1947; Sebeok 1956:436 (on the periodicity of relevant units and the rules of their organization which lead to the creation of a rhythm of poetic speech as distinguished from prose), etc.; Stankiewicz 1960a and (on emotive function) 1960b, 1961 (the problem of the linguistic code in a poetic text, the differences between poetic discourse and everyday casual language or prose and the distinctive features of the former, the multiplicity of functions of the poetic message, etc.); Saporta 1960 ('1. The application of linguistics to poetry must assume that poetry is language and disregard whatever else poetry may be; 2. Syntactic statements, that is distributional statements, are to be explored before semantic, if only because they seem to afford the desired degree of precision; 3. Stylistics is in some way dependent on linguistics, since style cannot be clearly defined without reference to grammar, but whereas the aim of grammatical analysis is essentially predictive, the aim of stylistic analysis is primarily classificatory; 4. Every message may be said to deviate from a norm in two ways which are independent in that one or the other or both may be present. The two ways involve the elimination of certain restrictions and the introduction of new ones'); L. Dolezel 1966; etc. Cf. in part the studies which have style as the main goal: "Stil und Formprobleme in der Literatur", Vorträge des VII. Kongresses der Internationalen Vereinigung für moderne Sprachen und Literaturen (Heidelberg 1959); Riffaterre 1959; Dolezel and Hauzenblas 1961; Mayenowa 1949; a number of works by J. Mukarovsk^, V. V. Vinogradov, et al. 49 On the problem of scientific language versus poetic language, cf. S. Marcus 1965 (cf. his monograph on mathematical poetics: Bucurejti, 1970); and in part, Savory 1967. 50 Undoubtedly, also in other schools of literature the conviction is gradually being formed that the solution to a number of important problems (first of all stylistic ones) must be sought in the direction indicated by linguistics; cf. Budzyk 1956.

706

V. N. TOPOROV

literary (in this case, poetic) work there exists some important aspect (besides possible others) which may and should be studied by linguistic methods. This aspect is poetic language as a means of organizing the text such that (if the traditional concepts are used) 'form' and 'content' are found in essentially fuller fusion than in nonpoetic speech.51 For the time being, evidently, one can limit oneself to this very general and incomplete definition of poetic language. It should be immediately stipulated that basically works on poetic language have been constructed not on folklore material but on the basis of literary texts where it was necessary to deal with a whole series of circumstances which seriously complicated the task of the researcher. Because of this it is precisely folklore oral-poetic texts (as, undoubtedly, a simpler category of cases) which ideally correspond to notions of the character of poetic language and can further serve as a model for the analysis of poetic texts of greater complexity. In research of oral-poetic folklore, first and foremost of the epos,62 a large role was played and continues to be played by works devoted to formulaic constructions which were begun by M. Parry (1928, 1929, 1930-32, 1933, and, with Lord, 195354) and continued by A. B. Lord (1948, 1951, 1956, 1960, etc.) and their successors (cf. Skendi 1954; ch. X, "Verse and formulas"; Emeneau 1958). Despite the attempt to go beyond the limits of the epos, studies of this type cannot pretend to a universal approach to folklore in all the heterogeneity of its genres. Besides this, an analysis of formulas on the various levels of an oral epic work despite all its value neither leads directly nor in the shortest manner (or even in general) to a segmentation of the text which would reveal the structure of the text in its entirety, nor permits construction of the corresponding models. From the point of view of structural segmentation of the text, the formulas can be accidental conglomerates of units. Therefore, reproaches addressed at this trend in research are fully understandable, in particular from the direction of structural folkloristics (Dorson 1963: 109ff.) Nevertheless, the resources of 'formulaic' analysis are not exhausted. And the results obtained in the best works of this school have increasingly more direct relations to an analysis of the STRUCTURE of oral epic texts. It is important to note that such a change is not only the result of beginning influence by the methods and ideas of structural folklore studies but is explained also by the further perfection of method and expansion of levels to which the 'formulaic' approach is 51

' . . . Our final view, implicit in our whole narrative and in whatever moments of argument we may have allowed ourselves, has been that "form" in fact embraces and penetrates "message" in a way that constitutes a deeper and more substantial meaning than either abstract message or separable ornament. In both the scientific or abstract dimension and in the practical or rhetorical dimension there is both message and the means of conveying message, but the poetic dimension is just that dramatically unified meaning which is coterminous with form' (Wimsatt and Brooks 1957 :748). On the unsatisfactoriness of the very division into 'form' and 'content' with respect to poetic texts, see Weitz 1950: ch. 3; Sebeok 1959b : 393; Ingarden 1939 : 51ff. 52 Cf., however, the attempts to broaden the sphere of application of the 'formulaic' method beyond the limits of the epic itself (e.g. to the ballad or the folktale). See Jones 1961 (cf. Friedman 1961b) or t-ord 1962 : xiii-xxix; etc.

FOLK POETRY: GENERAL PROBLEMS

707

applied.53 It seems that further perspectives of this approach unfold in connection with recognizing the importance of the following premises to which the particular attention of specialists has recently been fixed: explanation of the tie between the length of the completed formula and the internal segmentation of the verse into parts by word-breaks which are elements of its structure (consequently, by means of the formula a relationship is established here between the rhythmical level of the verse and the units of the scheme of expression54); the problem of the position of the formula in the verse and of words in the formula (and sometimes even broader — the position of the verse in a larger fragment, on one hand, and the position of morphological and syntactical forms within identical contexts, on the other) from which the problem of the elements' distribution arises and, further, the problem of identity and difference which presupposes resorting to paradigmatics; the relationship between formulas (and via them also between structural units of the texts) in the linear sequence of the poetic text and methods of syntagmatic coordination of formulas in this sequence; a dictionary of the formulas arranged to correspond to the basic semantic fields of epic works.55 With such an extension of problematics, the analysis of epic formulas leads the researcher to posit the task of synthesizing the text, or, more precisely, to the task of constructing models of synthesis of corresponding texts (models of generation, cf. works on generative poetics in other schools: M. Halle, S. R. Levin, M. Bierwisch, K. Baumgartner, et al.). Therefore, one may think that further studies of the 'formulaic' structure of the epos and of other oral-poetic genres (naturally, this can be a question also of tales, charms, proverbs, riddles, etc.) do not lose their urgency; more than this they are probably most closely tied with genuinely structural studies in folklore. Independent of the described evolution of this trend in the research of folklore texts, it preserves its importance for the compilation of dictionaries of formulas within a given genre, tradition, etc. Such dictionaries would fix a level intermediary between the lexical 53

By formula is usually understood 'a group of words which is regularly employed under the same metrical conditions to express a given essential idea' (Parry 1930:80, Lord 1960:4). Bsides the works of M. Parry and A. B. Lord indicated, cf. the newer works by Kirk 1962: Pt II. §§1-2, 1966a (the problem of the correlation of rhythmical units and units of meaning, the analysis of enjambement, the relationship of the length of the formula or the word to their fixation in verse, the controversy with H. Frankel [1926]); Xarkins 1963 (against the concept of formulas as a passive principle of organizing the text, about the 'grammar' of word formulas, and so on); etc. 55 Cf., in addition to that mentioned, Kirk 1966b; Russo 1966 : 217ff.; Lord 1960:30ff., 65ff.; etc. (especially important is the inclination of formulas to certain semantic bundles, the substitution depending on meter, the determinability of one part of the verse by the other, the role of formulas in the creation of poetic grammar, the role of formulas in the construction of plot sequences, etc.); see also Lord 1956 (the connection of key topics with the choice of sounds, the role of sound chains in the composition of the text — from sound to word, from word to sentence, from sentence to motif, and so on); etc.; Pop 1968 (questions of modeling and composition); Ionescu 1966 (the role of parallelisms based on similar syntagms in the composition); Rychner 1955 (the compositional role of the line or laisse, the relationship to formulas, the set of formulas for designation of the basic actions, a dictionary of motifs, the idea of full cliche-ing of the whole genre); Peukert 1960, 1961; Pollok 1964; Evgen'eva 1963; Ozols 1961; etc. 54

708

V. N. TOPOROV

and the motif-forming, connecting linguistic structure to supra-linguistic structures with plots. This is not to mention the significance which such dictionaries would have for typological studies. Along with this an inventory and classification of formulas would be an essential contribution to the theory for the creation of the most stable codes, transmission of texts in time, and, finally, to the problem of the correlation of the structure of a text and the structure of memory.56 In a certain sense, but with essential limitations, this aim (the compilation of dictionaries of formulas, studies of a statistical nature of the vocabulary and text, etc.) call to mind the task, the solution of which is assigned to content analysis, in particular to that part of it which is called contingency analysis.57 The first results of applying content analysis to the research of folklore texts are already known (Armstrong 1959; Saporta and Sebeok 1959; Sebeok and Ingemann 1956:265ff.), as also are some perspectives which are opening. It is necessary to think that connecting content analysis with linguistic problematics will help it become something greater than it is row. In any case, it is hoped that the results of applying content analysis will have importance not only for content but also will clarify corresponding structures of the aspect of expression. 5. One of the key concepts shared by all who try to apply structural methods to an analysis of folklore texts is that of LEVEL. With regard to poetic texts this concept attains a special supplementary meaning (see below). Of course, the concept of level, aspect, layer, etc., was applied in practice earlier to the research of poetic texts.58 The contribution of the structuralist approach to the concept of level consists of the fact that, in the first place, within each level are identified units which form a SYSTEM and, in the second place, these levels build on one another in such a way that they form a series of interconnected, hierarchically-organized structures (in other words, in connection with levels, horizontal and vertical axes are differentiated). Among the consequences of such a concept must be considered the theoretical possibility of transition from one level to another since units of the given level are synthesized into units of the following (higher) level, etc. Since units of the lower levels are linguistic units, it becomes possible through consistent synthesization (or, in other terms, re-coding) to pass from solely linguistic units to units of higher supra-linguistic levels. The structured analysis of folklore is based squarely 66

Cf. the series of ideas connected with this topic in such works as Meyerson 1956; Vernant 1969; etc. This same topic comes up in practice in analysis of the changes which a text undergoes in the process of its reproduction in time and space. Cf. Lord 1960:123ff.; in part cf. Uxov 1970; etc. 57 Cf. the definition: 'Content analysis is a research technique for the objective systematic, and quantitative description of the manifest content of communication* (Berelson 1952:18). With the aid of content analysis some specific conclusions are obtained about some aspects of the goal-oriented behavior of the author of a text. Quantitative content analysis is a statistical technique for obtaining information about content variables (George 1959). 58 Cf. the scheme by J. Petersen, W. Kayser, R. Wellek, and A. Warren, et al., if one is to speak about the study of literature.

FOLK POETRY: GENERAL PROBLEMS

709

on the supposition that 'all the products of verbal behavior may, in the last analysis, be reduced to terms of symbolic logic' (Sebeok and Ingemann 1956:261-2) and that modern structural linguistics may be considered one of the varieties of discrete mathematics. Sebeok, paying ample attention to the concept of level in his concrete analysis of Cheremis oral-poetic folklore texts, usually proposes (for the cases indicated) to identify two levels: 1) glottic and 2) metric, taken in two aspects: a) sound and b) meaning. From this come four groups of topics isolated in an analysis of a poetic text: I. phonemes, distinctive and configurational features; II. morphemes; III. rhythm and meter; IV. images, metaphors, symbols; etc. (1956, 1959b, 1960). The system of levels, similar in principle, treated from the standpoint of the analyzing description (the opposite of the synthesizing model which ideally generates texts from the higher levels to lower ones, i.e. consistently transforms extralinguistic content into chains of symbols of increasingly lower levels) was proposed by V. V. Ivanov and the author (Ivanov and Toporov 1966, Toporov 1966). In this latter attempt there is proposed a scheme of ascent in levels — right up to level of motifs, plots, and functions, and a system of notation is introduced, as with Sebeok for each level. On concrete descriptions of the organization of various levels of oralpoetic texts, cf. below. A somewhat different approach to the same problem of levels is connected with the ideas of Husserl and is reflected in the works of Husserl himself as well as in those of his successors.59 Among such approaches, the one more well-known than others is that associated with the name of R. Ingarden, who proposed a theory of 'two measurements' of a work of art. In one of these measurements, one must deal with the sequence of phases which replace each other — of parts of the work, and in the second with a set of jointly-occurring heterogeneous components, or 'layers'. Among the latter belong: one connected with sound; connected with meaning; connected with subject; connected with form (having in mind the form in which the portrayed object is presented). It should be noted that the play in a poetic text of these two measurements — multiphasedness and multilayeredness — independent of studies in the field of phenomenology, was stressed not only by R. Ingarden but also in works on structural folkloristics (and, more extensively, on the structural theory of the text per se in general).60 The definitive level of oral-poetic folklore texts is METRICAL. The linguistic foundations for analyzing this level were laid down in the decade after the appearance in 1933 of N. Trubetzkoy's article "Zur Struktur der mordwinischen Lieder", first of all in the works of R. Jakobson (with Lotz, 1941) and J. Lotz (1942, 1954, 1956, 1960). In these works it was demonstrated that the linguistic analysis of 59

See, among others, Spet 1922-23 (cf. especially 'structure of the word in usum aestheticae' and the quasi-mathematical formula of the esthetical perception of the word); and Ingarden 1931, 1947, 1957, 1961. 60 Partly similar constructions, but in coordination with linguistic structures, are advanced in the article by Gorny 1961.

710

V. N. TOPOROV

metrics exhausts its description (especially if the widespread variations in linguistic analysis are taken into consideration), concentrating on the formal features of the material defined by a set of linguistic units and their arrangement (e.g. syllable, word, member, sentence, segment, verse, and utterance). In the same way, prerequisites were made for constructing a typology of metrical systems and — as a special case — for the reconstruction of lost metrical schemes, known only from their relatively late reflections in historically attested traditions.61 A significant part of these works was based on the material of folklore poetic traditions. Therefore, the model research of metrics of the Ob-Ugrian texts by R. Austerlitz (1958) was a further development of an already-formed tradition. Austerlitz (1958:32) found the initial methodological principles of his analysis in Lotz's formulation (1954:376, fn. 22): 'In certain cultures there are texts which are characterized by a numerical regularity of the phonic material occurring in various semiotic frames (words, phrases, sentences). In the earlier article classes of notions were distinguished: phonic, semiotic, and metric. However, the metric constituents such as hemistiche, verse, strophe, etc., are not fundamental; they are definable superstructures imposed on the semiotic-phonic stratum and should be treated as descriptive categories satisfying the structuring of the entire linguistic material in numerically semiotic frames. The basic regularity refers generally to the number of syllabic pulses in such frames. Further metric differentiation depends on the use of the prosodic features: stress, pitch and duration.' In the analysis of metrics by Austerlitz, besides many other achievements, the indications of the connection between the units of the metrical level and grammatical units are valuable (cf. IV, 'Grammatical notions relevant to metrics'). Thus, in the book all those grammatical concepts are described which are essential for metrical analysis; metrical-grammatical correlations are established (morpheme—the parallelism is the etymological figure; word — repetitive frames, terraces; phrase — various types of lines; sentence — structures ending in finite verb); sphere of application of quantitative (syllables) and qualitative (morphemes, word classes, and syntactical constituents) analysis are defined; and the very important dichotomy verb-noun is emphasized. In Austerlitz's opinion, metrical analysis includes all the procedures which are connected with the segmentation of the texts into constituent units, the description and classification of these segments, the description of their internal structure and the application to them of quantitative criteria. Special attention is given to the procedure of segmentation, if one starts from the complete text, beginning with the division of sentences along major syntactical boundaries, which most often are formulated by finite forms of the verb or their equivalents (the intervals between sentence boundaries form sub-segments). As a result of successive segmentation, an inventory of units which repeat in the text and are used as metrical devices is 81

Cf., first of all, Jakobson 1952, not to mention his early works on Czech, Serbo-Croatian, and Russian verse.

FOLK POETRY: GENERAL PROBLEMS

711

obtained. By the same method an internal organization of the text is revealed,62 the units of which can receive further interpretation as to content.83 Although the study of Ob-Ugrian metrics remains the most complete and valuable model in this field, in recent years there have appeared more and more works which demonstrate the possibilities of the structural-linguistic approach to this level in the analysis of folklore texts.64 The basic achievements of recent years in the research of the PHONOLOGICAL level consist of finding a very convenient unit of this level — the distinctive feature (along with the phoneme). Treatment of a poetic text on the level of distinctive features permits attainment of much more subtle results than those gained by operating strictly with phonemes.65 First of all, this refers to the analysis of the distribution of phonic elements to which the statistical approach can now be applied with greater effectiveness (cf. for example, the definition of tone balance proposed by L. G. Jones [1965; cf. Lynch 1953], or the quantitative data about non-random distribution of initial phonemes,66 or the definition of the degree of sound organization [similarity] of the verse in relation to the given feature [Toporov 1966: 80ff.], and other indexes). Along with this the use of the distinctive features of phonemes in a description of a poetic text affords the possibility of uniting into a whole all types of sound reduplications and of giving their detailed classification (cf., with respect to terminology, Masson 1961). There is no need to speak in more detail 68

Cf. Chart 1 (Austerlitz 1958:40), where a general scheme for the identification of lines is given: Total Non-Verbal

Verbal

Isolated

Anacrustic

Cohesive

Orphan

Parallel

Terrace

Introductory Exclamatory (cf. also the tales on pp. 81, 86, 88, 109, etc.). Cf. the idea, emphasized by some scholars, of the interdependence between rhythm and thematics ('expressive halos' fixed for some metrical strtuctures) (Taranovski 1963). 64 Cf. Trost 1961; partly in this connection the observations are interesting in Stokmar 1952; Pope 1966; and a number of works by W. P. Lehmann, S. Chatman, et al. 65 For examples of the use of distinctive features in analyzing poetic folklore texts, see Sebeok 1956 : 431-2, 1960:225; Toporov 1966 : 78-83; in connection with literary texts, cf. the same in the article by Taranovski 1965. It is appropriate to note that the very choice of features is somewhat different from the similar procedure in the matrix of the identification of phonemes and permits the use of redundant features. 86 See Sebeok and Zeps 1959. The authors proceed from B. F. Skinner's principle of formal strengthening and from structural-linguistic ideas in an analysis of alliteration phenomena (Skinner 1939, 1941, 1957). Of other works on alliteration, cf. also Vargyas 1955 and Cizevsky 1947. 63

712

V. N. TOPOROV

about the manner in which these studies are carricd out in identifying the units of the phonological level and their distribution. It is more important to stress that in recent works in studying the distribution of phonic elements scholars more frequently attempt an interpretation in 'meaning' (conditionally speaking) of the results of the distribution of phonetic sequences, thereby continuing the brilliant experiments of F. de Saussure (1964, 1967, 1968, 1969, 1970, 1971a, 1971b), E. Sapir (1929), and some other specialists (not to mention the work of E. Sievers). It is true that these works most often are not directly connected with an analysis of folklore texts.87 In folklore texts the repetition of similar phonetic configurations is most often semanticized by their being found under identical grammatical conditions (identical morphological categories, most often in identical syntactical constructions), by matching to the given key word (more often a noun) words grammatically connected with it which in respect to sign are nearer to the key word, or, finally, by abstract symbolization.68 There also exists such a method of semanticization that indicates the fact that the given text is qualified as folklore (or, with even more specialization, as a song, etc.), cf. some characteristic changes of sounds or rules of their distribution which are not encountered at all in normal language or, at least, not in juxtaposed words (Bogatyrev 1962, Jakobson 1962).69 A special role is played by the realization of phonetic configurations as a means of answering (true or false) in riddles in many traditions.70 Semanticization of the elements of the phonological level, no matter how conditional it is, is deeply justified and in a folklore oral-poetic text most often may be explained reasonably (for a literary text it is significantly more difficult to do this). If, paraphrasing W. H. Auden, in good verse 'the element of chance in language' ought not 'to predominate over the element of fate and choice',71 then these elements of fate and choice should be reflected on other levels and/or in another place in the same text. Therefore, it is appropriate to welcome all attempts to find these reflections and, consequently, the connections between the units of the phonic level and the other levels. Cf. the above-mentioned consideration by A. B. Lord as to how definite clusters of sounds assist transition from one episode to another in the plot; or the observation by J. Thompson that the relationships between the structure of sound of the English language, the structure of sound of the metrical pattern, and the structure of sound of the line of verse may be described • 7 Cf. at least de Groot 1957; Chastaing 1958, 1962; Karlgren 1963; Fonagy 1963; etc. Cf., moreover, the valuable work on Lithuanian folklore, D. Sauka 1970 : 222ff. 68 E.g. if the difference between the low level of tonality and the neutral (or high) level is clearly expressed in the given fragment, the low level may be interpreted as an indication of some abstract meaning selected according to additional criteria. 89 Cf. analogous phenomena in the Indian tradition, Jacobs 1959b: 7. 70 Cf. the type of riddles in which the phonic configuration in the question is analogous to configurations in the anticipated answer (cf.: Kakasangete I kohla pwutangko pakpaker? Tipaker. 'I saw two boys boxing. What is it? Tobacco' (Blackburn 1964; Polivanov 1918). It is interesting that this type of riddle in many traditions is opposed by another type in which the question and answer are coordinated in meaning. By the same token riddles occur in the role of exercises for training in mastery of the rules of synonymy and homonymy in the given language. 71 Thomas A. Sebeok recalled this statement in one of his works.

FOLK POETRY: GENERAL PROBLEMS

713

in the following manner: 'the metrical pattern imitates the structure of sound of the language; the line of the poem imitates the metrical pattern; and, therefore, the line of the poem imitates the structure of sound of the language' (1961); or those dependencies between sound and meaning which D. H. Hymes (I960) 72 establishes, etc. Although many theoretical prerequisites and basic parts of the apparatus of description of the phonological level of folklore poetic texts are known, practical analyses are quite far from completion and most frequently are inconsequential. Therefore, an exhaustive description of this level with an indication of its connections with elements of the other levels and of basic directions in its semanticization remains among the most urgent desiderata of folklore studies. It is not excluded that basic difficulties along the route of solving this problem will bear a purely technical character (a method of presenting the results). Without the attainment in this field of a clear picture, fruitful development of a typology of folklore genres constructed on phonological criteria is impossible. In conclusion to this section, it is worthwhile to direct attention to the problem of the isomorphism of configurations of phonic elements and configurations of elements of the other levels — right up to the highest.78 In any case, oral folk poetry is one of the most crucial areas in which one of the basic contemporary problems — sound and meaning — should undergo experimentation. 6. The identification of the MORPHOLOGICAL level in the research of poetical folklore texts must not be rated just as a contribution to linguistic theories and concepts or as an attempt to describe exhaustively the given text according to different — essential or non-essential — criteria. The reason for selection of this level lies elsewhere. The poetic text (if it is not doggerel) is that locale in which GRAMMATICAL meanings, if they do not lose (in any case, not fully) their compulsoriness or necessity, are, nevertheless, maximally realized and introduced into the sphere of the recognized. They become capable of supplementary (already noncompulsory) semanticization, presupposing a sort of possibility of selection between them (cf. extreme cases when into the poetic text is incorporated a whole paradigm of grammatical forms, each of which serves strictly poetic aims).74 It is completely 72

Cf. on the use of phonic means in the language of (lyrical) poetry, Langer 1953 :258-9. The concept of isomorphism presupposes the concept of position, in this case — of the relevant place in the verse which is taken by the elements noted of various levels; this can be observed either in the whole body of texts of a given genre (cf., e.g., the system of rhymes including, naturally, also internal rhymes, in the Lithuanian ballad, where the rhyme of every relevant position is assumed with any other), or even within the limits of one text: cf. the rhyme scheme in Burmese poetic texts with strict rhyme patterning (including here riddles) (see Maung and Dundes 1964): X - - X - X - X 74 Cf. the set of case forms of the name of the god Agni in the Rig-Veda (I, 1); cf. also for Vsas and some other gods. T o this there correspond sound configurations indicating the name of the deity to which the hymn is devoted (cf. Vac in X, 125). 73

714

V. N . TOPOROV

evident that R. Jakobson was correct in pointing out, in his remarkable article on poetry and grammar, that 'of all fields of speech activity it is precisely poetic creativity which endows "language fictions" with the most significance' (1961:399). And although the poets themselves in their creativity, and sometimes also in making excursions into the field of theory (cf. Gerard Manley Hopkins), testify to the exclusive importance of grammatical figures in the poetic text; although researchers of folklore poetry also have pointed out that in it an intensified play on the keyboard of grammatical forms and meanings take place, the customary approach to the morphology of the poetic text has remained and continues to remain such that the research goal consists of seeking out archaisms or other dialectal deviations from the modern literary language. It is natural that in such an approach not only are they not interested, as a rule, in those grammatical facts which are not rarities, but even the rarities themselves are not studied from the point of view of their use in attaining a poetic effect. Morever, for some, poetic language itself is 'a kind of contre-grammaire, an unruly appendix to non-poetic language' (Stankiewicz 1961: 12). Of course, study of deviations in folklore poetic texts is important, but this importance is explained rather by the very fact that even the deviations are nothing other than the appearance of a CONSCIOUS relationship to grammar in the poetic text.75 Under these conditions it is proper to relate positively to those works in which attempts are made to describe poetical grammar (in folklore or literary texts) in more or less FULL scope (Ozols 1961; Evgen'eva 1963; Davie 1955; Berry 1958), or, on the other hand, the POETIC usage of INDIVIDUAL grammatical forms (Aleksynas 1968), or, finally, a COMPLETE description of the grammar of a GIVEN text (or texts) with an indication of how it is used for ESTHETIC aims.76 Here it is appropriate, of course, to recall that the reliability of conclusions in this area depends significantly on the quantity of the texts analyzed. A detailed study of the morphological level (with an introduction of statistical data) will undoubtedly permit differentiation of oral-poetic genres (or even smaller types) from each other. Going further, researchers may be confronted with the task of differentiating 'local' grammatical models which coexist in the confines of one text (or one group of texts).77 Having solved these problems, it is appropriate to turn to the problematics of a

75

Cf., e.g., omission of the article, confusion of past tense forms in the indicative and indirect moods in Bulgarian songs, etc. (see Andrejcin 1961). 78 Cf. the already cited articles of Sebeok devoted to an analysis of individual Cheremis poetic texts. 77 Cf. the cited article of the author where it was shown that within the limits of Lithuanian ballad texts, two grammatical models may be divided out. The first (conditionally, 'song') is characterized by a definite predominance of forms of the 1st and 2nd person in the verb, of personal and possessive pronouns; frequent tenseless use of personal verb forms; an abundance of 'directed' forms (imperative, vocative, interjections of a stimulating type), etc. A second model (conditionally, 'epic') is characterized, on the contrary, by a strong predominance of forms of the 3rd person in the verb; the use of tense forms of the verb with their rather precise differentiation according to themes; the reduction of 'directed' forms, etc.

FOLK POETRY: GENERAL PROBLEMS

715

morphological typology of the genres within a given tradition or in various traditions. It is unnecessary to speak here of the technique of morphological analysis of oral-poetic texts (identification of the units, distribution of the forms; analysis of the environments in which different forms may be maximally equated or similar forms may be maximally differentiated, cf. the problem of equivalence and difference of grammatical elements in poetic texts78), since it is borrowed from structural linguistics (most often of a descriptive trend) and is used in many folkloristic works already named. On the other hand, it is expedient to point out here the problem of the connection between units of the morphological level and other levels, both lower and higher, since the morphological codification of the text (as an intermediate one) most easily permits combination or summarization of the phonological and syntactic content of the text. Generally speaking, it may be admitted that these inter-level ties are realized with the aid of repeating elements which occur simultaneously as grammatical forms of their components. A special case is rhyme. Its grammaticalness is not a sign of primitivism but a realization of the feature of economy. Rhyme is that area where grammar and meaning meet with sound in a most intimate manner 71 and where norms of equivalence are established for the given genre and tradition (thus, if A, B, C, D . . . N are different words with identical grammatical meaning, and k, 1, m, n . . . are morphs conveying this meaning, then k, 1, m, n . . . may in the given tradition be interpreted as equivalent rhymes regardless of their sound similarity; an evidently opposite example which confirms the same premise is the absence in poetic texts of many traditions of rhymes in which words belonging to different grammatical categories participate [in the case that an attempt to find such rhymes in language would not present difficulties]). From this results the theoretical grammaticalness of rhyme in oral-poetic texts. Other examples of the correlation of phonological and metrical elements with grammatical ones have already been noted above (R. Austerlitz, et al.). Also the means for connecting the morphological level with syntactic structures, with the level of elementary images, or even with the level of (the simplest) motifs may be pointed out. The distribution of identical grammatical meanings in the text may form the most varied configurations (they may conditionally be classified according to the same schemes of sound reduplications which were proposed by O. M. Brik) — from stable, symmetrical, or correctly alternating ones to dynamic, contrasting ones or TS

Cf. on the two types of equivalency in connection with the paradigm Levin 1962:29ff. (in the same source, cf. their usage in poetic texts). 79 The dependence of these components was formulated by a poet: Meaning is vanity, and the word — only noise, While phonetics is the servant of the seraphim. (O. Mandel'Stam) On rhyme especially, cf. another citation from the same poet, already used by R. Jakobson: And the bell of verbal endings Shows me the way in the distance

716

V. N. TOPOROV

those which are built on a principle of ascent or descent. Such configurations of grammatical meanings define in a natural way some isomorphic configurations of syntactic structures which in turn presuppose to some extent or other similar constructions also on the level of motifs (similar to the way in which rhythm or melody can implicate some content structure although in a very general and approximate way). Therefore, it is not accidental that in recent years the exceptional interest of folklorists and linguists is evoked by the problem of parallelism in oral-poetic texts.80 With the aid of parallelism, isomorphic and heteromorphic elements of all levels are explained and, so to say, X-rayed, the movement of the action in the text is realized, and, finally, the verses, lines, and even larger compositional blocks are bound together. Finally, the analysis of elements of the morphological level in many cases permits definition of the 'regularity' of the text which is given by the quantity of 'voices' which participate in its creation and sometimes even assists in reconstructing the archetype of the given text (thus, the dialogic or — more extensively — dramatic form of a series of balladic, song, or other texts may be ascertained from their present monologic narrative structure, cf. M. Jacobs's discussions of the fact that, in the case of the Indians described by him, the story reminds one more of a stage play, and its participants actors, rather than, respectively, a story and characters.)81 The most important goals in the research of the morphological level in oral-poetic texts in the near future evidently consist of an inventory of the grammatical elements and schemes of their distribution (with an indication if not of the permissible rules then of the active prohibitions), in the first place, and the establishment of their role in connection with the other levels, in the second place, and a definition of the role of these elements in the composition of the poetic text, in the third place. 7. Research on the SYNTACTIC level in oral-poetic texts was in part already discussed in connection with the consideration of lower levels (first of all, cf. the notes on the connection between the units of these levels and elements of the syntactic level). It is evident that also for the syntactic level there looms the task of identifying the basic units (immediate constituents) and of establishing the rules for their combination in the poetic text with an indication of those limitations (or, on the contrary, additional possibilities) which differentiate this text from non-poetic ones. As in previous instances, ready solutions are offered here by structural linguistics. 80 Cf., of the older works, Newman and Popper 1918-23; Steinitz 1934; Tschang 1937; Austerlitz 1958, 1961b; Poppe 1958; Gonda 1959; Zirmunskij 1964; Ionescu 1966; etc. To a certain extent the extensive literature devoted to the form and function of beginnings and endings in folklore texts should be added here. 81 A broad excursus into the field studied by M. M. Baxtin is opened up here. In some very general and not fully clear form, a connection is drawn with the ideas of K. Burke about literature as a symbolic action, about works of art as a special strategy of behavior, and about the possibility of dramatized description of the object of research (apparently, K. Burke's 'grammar of motives' is not only a tribute to the metaphorical manner of expression). Cf. Burke 1941, 1945.

FOLK POETRY: GENERAL PROBLEMS

717

In the works of Thomas A. Sebeok, R. Austerlitz, and many folklorists who followed after them, procedures are proposed for syntactic segmentation of the text, beginning with the text as a whole and ending with immediate constituents. In other studies the stress is placed, on the contrary, on procedures of a synthesizing type, and it is indicated how elementary syntactical constructions are obtained from the grammatical classes of words and how the elements of these constructions are developed into larger syntactic blocks up to sentences (Ivanov and Toporov 1966: 103-5). These rules of development may in principle be applied several times and combined with each other by means of a substitution of corresponding elements due to which more complex rules may be obtained from relatively simple ones. A further complication of syntactic structures comes with the assumption of cycles. For cycles, and also for some other rules, quantitative limitations may be established which relate to the number of possible applications of the corresponding rule in the same sentence. Besides limitations of a quantitative character, qualitative limitations also exist — prohibitions which relate to the possible combinations of classes of words, grammemes, and grammatical meanings. As a result of the usage of rules of the type indicated, there arises the possibility of describing a tree of the sentence without reference to the order of the elements in the actual text, which may be considered a projection of this tree on a straight line. The specification of the order of the elements is very important since different methods of ordering these elements can evidently be used for differentiating texts or parts within one text. On the basis of the rules regulating the structure of the simple sentence, a description is possible also of the rules for constructing complex sentences which are obtained either by means of a cyclical combination of simple sentences or by means of the introduction of some combinatory elements and corresponding transformation of the given simple sentences. Naturally, the procedures for analysis mentioned above on the syntactic level not only do not contradict the described synthesized approach, but also are fully compatible with it. It is necessary only to remember that the synthesizing approach in principle leads to the construction of a stronger model of the given phenomenon. The importance of the synthesizing models for study of the syntactic level may be explained by the fact (among others) that, in the first place, the syntactic level most clearly connects linguistic elements with the components of poetic language,82 and that, in the second place, the syntactic level to a much more significant degree motivates (or is connected by) those supra-linguistic levels of the poetic text which are connected with actions and motifs. The special currency of the syntactic level in oral-poetic texts is emphasized by a heightened play of the elements on this level. This play, in particular, appears in test texts existing in various traditions, the aim of which is to attract the attention of the participants 82

Cf. the varied attempts at radical interference in syntax in contemporary poetry, right up to the tendency 'to abnegate syntax' (Davie 1955:99); Cf. Hammond 1961 in which (p. 482) there is a definition of what 'poetic' syntax is: 'Syntax is poetic when grammatically equivalent constituents in connected speech are juxtaposed by coordination of parataxis, or are otherwise prominently accumulated.'

718

V. N . TOPOROV

in the communication to what is being spoken of or to verify whether this attention exists ('phatic' function), and to establish whether the participants in the communication are using the same or a somewhat different code ('metalingual' function). Such test texts actually demonstrate the MAXIMAL possibilities of syntactic operations (up to and including occasional ones or those only theoretically assumed or even those which go beyond the limits of what is permitted in the language) or emphasize the lack of correspondence in syntactic figures which automatically follow one after the other and the structure of higher levels which are connected with them (in particular, plot levels); cf. the syntactically correct texts noticed with the substitution of non-corresponding lexemes (on the pattern of exercises of this type in linguistic literature on transformational grammar). By the way, also in the oral-poetic texts of other than a testing type, a somewhat frequent delay in perception of the aboveindicated lack of correspondence leads to the special effect of alienation, the removal of the automatism of the content level within the framework of the automatism of lower levels. It is of this, in particular, that the significance of poetic usage of redundancy consists on the level of the syntactic organization of the text. Cf. the role of syntactic elements in striking a balance in the text between tensions and releases, etc. Although there are many works on the syntax of oral-poetic texts, on the whole their shortcomings lie in their underestimation of the conscious realization of this level and the play of its elements and also in their disregarding the goal of defining the motivation by the elements of this level of content structures. A dictionary of elementary syntactic constructions and the clichés which comprise these constructions would bring researchers nearer to a solution to the problem of synthesizing real texts, in the first place, and a typology of oral-poetic genres on the syntactic level, in the second. For the time being, in both instances only the first steps have been made, and they are often stimulated by the intuition of the researcher and the logic of objective analysis of oral-poetic material rather than by a clearly-recognized aim. Indebted for its origin to structural linguistics, the recognition of the importance of the TRANSFORMATIONAL level in the study of folklore texts, primarily of poetic texts, must be considered one of the most significant achievements of modern folklore studies. This level poses before the researcher with exceptional clarity the problem of equivalence and is, therefore, already becoming a prerequisite to synthesized procedures in generating the text. With the assistance of this same level, as was noted above, the concept of the IDENTICAL (equivalent) meaning first receives precise operational definition by means of a cumulation of transforms which are formed according to corresponding rules of transformation. 'These transformations with a preservation of meaning are possible when lexical limitations do not hinder them; because of this, in a given concrete word it is necessary to select only some one method or several (but not all) methods. In other words, one can carry out a breakdown of all words according to how they are able to participate in transfor-

FOLK POETRY: GENERAL PROBLEMS

719

mations. The indicated syntactic relationships . . . could be described not only by means of the indicated syntactic transformations of word combinations . . . but also with the aid of transformations, on one hand translating a phrase into one word . . on the other hand, translating one word or phrase into a whole sentence . . . ' (Ivanov and Toporov 1966:108). 83 The presence of syntactic and lexical synonymy and the corresponding wealth of transformations (versatility) make possible the task of generating not one text but MANY variants of a text which express a given meaning. Naturally, the versatility of language creates the possibility of a STYLISTIC choice of one of many methods of expression, and here the researcher already crosses over to the supra-linguistic level of analysis. In some works (unfortunately, there are not yet many of them) not only are the advantages of the application of transformational theory to an analysis of folklore poetry demonstrated (transformational theory allows operating with a reduced set of units — semantic distinctive features; it extends the field of research at the expense of the analysis of complex semantic schemes which are formed by syntagms, phrases, etc.; it introduces internal criteria into semantic analysis; it permits carrying out a parallel analysis on various levels of the poetic text; it gives the possibility of posing the question about the semantic system of the given text, genre, etc.84), but also attempts are undertaken to analyze concrete poetic texts. It is necessary to suppose that the number of such attempts and, mainly, their depth will increase. Ideally, efforts of this type could lead to the creation of transformational grammars of oral-poetic texts and thereby provide typological studies in the field of folkloristics with one more important series of criteria. 8. In order to know the way in which syntactic schemes in an actual text are filled out, it is necessary to have a corresponding dictionary with an indication of the frequencies and valences of the elements of this vocabulary. Such dictionaries, undoubtedly, would help reveal the differences in genres and types of poetic texts. But it is still more important that these dictionaries, taking into account especially their frequency characteristics, would rather unambiguously permit isolation of the SUBJECT level, in the first place, and would lead to MICRO-MOTIFS, in the second place. The series of subjects or themes of a given text, type of texts, or genre is rather easily identified (for this, in particular, computers have already been used more than once,85 cf. above on the method of content analysis). Usually (in reference to oral-poetic texts), the whole lexicon is segmented into local subsets without 83 O n the problem of the use of models in folklore studies, see the materials of the symposium published in Studia Ethnographica, 5, 1969 (Budapest); cf. especially the valuable article by V . Voigt, "Modellalasi kiserletek a folklorisztikaban"; cf. also E. Hankiss, "Kommunikaci6elmeleti modellek es nepdalkutatas", etc. 84 See Golopenjia 1966 (strongly influenced by Katz and Postal 1964). 85 Cf. Colby, Collier, and Postal 1963; Sebeok 1952, 1959b, 1965; Sebeok and Ingemann 1956; Sebeok and Zeps 1961; etc. Without resorting to computer technology, 'themes' are identified in a significant quantity of folklore studies (cf., e.g., Jacobs 1959a and b, 1960, 1966).

720

V. N. TOPOROV

special difficulty. Thus, e.g., in a Lithuanian ballad these CHARACTERS are distinguished: the maiden, the sister, the daughter, united by the features 'female' and 'young'; the young man, the brother, the son, the soldier, joined by the features 'masculine' and 'young'; the mother, the old woman ('female' and 'old'); the father, the grandfather, the old man ('masculine' and 'old'); the mother-in-law, the daughterin-law, the widow (with the additional feature of harmfulness); SUBJECTS connected with the (basic) CHARACTERS: a ring, a wreath, a handkerchief, a horse, a saddle, a sword, etc.; LOCAL-TEMPORAL indications: a forest, a mountain, a road, the sea, the night, the morning, a war, etc.; CONCEPT-SYMBOLS: a rue, a braid, a lily, a snowball tree, a tear, a heart, blood, a word, clover, a cap, etc.; EPITHETS; basic ACTIONS, etc. Using these data, both the series of themes in the texts which use such vocabularies and also, with a certain probability, the series of basic plot schemes may be outlined rather simply. For the Lithuanian ballad it was shown how the plot possibilities arise from an analysis of the vocal vocabularies stipulated by an implication of the type: Ri (M & B)

(R2 (M, B) j/R 3 (B, M) l / R , . R 3 (B),

(Y is the sign of a disjunction), where Ri = 'to love', R 2 = 'to unite', R3 = 'to kill', R4 = 'to cause', M = 'maiden', B = 'boy' (i.e. if a maiden and a boy love each other, then one of the three following outcomes is possible: either the maiden and the boy join together, or the boy kills the maiden, or the boy perishes); in this case a further detailing is possible (Toporov 1966:113ff.). Analogous results may be obtained also for other genres, especially those which are strongly cliched (like charms — Cernov 1965), or, on the other hand, have a clear symbolic semantics (cf. dreams — Sebeok and Ingemann 1956, or wedding songs — L. Sauka 1969). Compiling vocabularies of this type, dividing out local subvocabularies, defining main semantic fields, key concepts, and symbols, and, finally, studying how the possibilities of transition to a higher level of motifs arise from this material — all this should comprise an inevitable stage in folklore studies. Side by side with a consideration of how the micro-motifs are constructed from the elements of the vocabulary using the available elementary syntactic constructions, which advance the action from the initial situation to the final (resultant) one, it is important to keep in sight how the elementary images belonging to a special level of the analysis are constructed. Thus, for example, in taking into account the potential for occurrence of a name N with the epithets Adji, Ad]2, Adj3 . . . Adjn and of the epithet Adj with the names N', N", N'" . .. Nn, it is possible to define several features of the subject (in a poetic text, these features are almost always essential) and to define the object by means of a set of features connected with it. Along with this, many objects can be united into a cumulation characterized by the presence of a general feature (or features). In this way here again there arises the pivotal problem of equivalence and difference transposed onto two axes — the axis of succession and the axis of simultaneously (cf. in connection with them the

FOLK POETRY: GENERAL PROBLEMS

721

linking and nesting operations in the meaning given them by U. Weinreich). The FIRST of them directly leads to the level of MOTIFS, FUNCTIONS, etc.; there is no sense speaking of this here (it is sufficient just to point out the extensive literature on this question in connection with the folktale). One may only express the assurance that for many oral-poetic genres (and for some literary genres, types, or even trends) the identification of such motifs or functions would not present special difficulty. Together with this the compilation of such inventories would open a new stage in the research of one of the most important problems of cultural anthropology and semantics — the way in which extralinguistic content becomes the material of literary works (Rychner 1955 :128ff.). With the SECOND of the indicated axes is connected the problem of constructing more complex images, particularly metaphors (in the broad sense of this term). Recent studies in the area of the structure of the metaphor (particularly the excellent works with a logical analysis of the metaphor and a formalization of its various types by J. Pelc 1961 and Ju. I. Levin 1965, 1969) 86 give the possibility of speaking about the synthesizing mechanism for generating metaphors, one of the most important links in stylistic research of poetic texts. With respect to FOLKLORE poetic texts, where the set of elements and rules participating in the creation of metaphors of a given type is extremely limited, the problem of synthesis (and also of transformation) of metaphors is close to a concrete solution, in principle, and awaits its researchers in connection with individual folklore traditions.87 The quantity of levels of analysis and synthesis of folklore poetic works is not exhausted by what has been cited above. However, they will not be specially considered here, the more so because concrete folklore studies of recent years do not frequently treat them or give interpretation to those which cannot be considered fully coordinated with the results of description of the lower-lying levels. It is true that rather many researchers of poetic folk literature touch on the topic of SYMBOLS (it was treated more than once in concrete studies of Cheremis texts by Sebeok). Along with this, symbols as elements of a definite system in which they can be described using some matrix or set of features are not researched despite the evergrowing quantity of monographs and dictionaries of symbols in various cultural traditions. Therefore, what symbolic space is and the nature of the principles of its hierarchical organization remain, as a rule, unclear. In any case, complex, multi-layered symbolic structures exist which seem to generate their own code and form an infinite series of ever-deepening symbols which attain a certain independence and exceed the limits of the theme of the given text. Of course, such complex 8

' Cf. also the valuable study by Permjakov 1968: Introduction. Here it is appropriate to recall the old work of great value by I. A. Richards (1936). In a completely different aspect, cf. Vianu 1957, 1961; Bratulescu 1966; Ruxandoui 1966; etc. 87 The importance of the problem of the metaphor goes far beyond the bounds of the stylistic level since the metaphor seems to summarize in itself the set of operations of importance for a given culture and which have their special linguistic expression.

722

V. N. TOPOROV

systems of symbols are revealed not in folklore texts but in the most 'intricate' works of belles-lettres or mystical-religious literature, but nonetheless, in building a universal theory of symbolic systems, it is proper to take into account even these extreme cases.88 In exactly the same way the broad circle of SEMANTIC problems which arise in folklore research remains here outside consideration. Already the very stratification of a folklore text into levels, the identification of units on each of the levels, and their configurations within the limits of the given level and of cross-level structures, as was indicated above, defines in a general way some basic principles of the semanticization of the text (cf. identical, different, opposite, repetitive, symmetrical, etc. — Segal 1966). In the works of Lévi-Strauss an attempt is made to identify the basic semantic oppositions of a binary type which form the whole system; to define the sets of 'classifiers' (vegetable, zoomorphic, etc.); to take into account the heterogeneity of the codes in which messages in a given tradition are created. This general network of concepts, in many of its elements indebted for its origin to structural linguistics, is beginning in turn to influence semantic studies in the field of folklore and is inciting researchers to construct systems of semantic oppositions also for more advanced cultural traditions than the ones studied by Lévi-Strauss (see Ivanov and Toporov 1965). In this respect ever greater similarities among the various schools of science which study semantics are discovered (cf. cultural anthropology of the American type, classical French sociology, structural linguistics,89 etc.). New ideas and methods for the linguistic analysis of semantics promise broad perspectives to folklore studies on this level. The transition from an analysis of lexical semantics to an analysis of the semantics of whole texts and their component parts, which is beginning to characterize contemporary linguistics, undoubtedly answers the most pressing demands of folkloristics. It should be noted that on this route there are revealed connections between the semantic segmentation of a text, 88

See the outstanding work by Emrich (1943). Of the literature of recent years in the field of research of the symbol three groups of works should be disttinguished: 1) those devoted to an analysis of the symbols of individual cultural-historical traditions (cf. Daniélou 1961; Hammerschmidt et al. 1962; Ehrich 1959; Golf 1963; Duchesne-Guillemin 1961; Kirfel 1959; Marchal 1956; Gimbutas 1958; Griaule 1952, 1957; Hermann 1961; Musurillo 1961); 2) those devoted to an analysis of individual symbols (basically the main and most all encompassing ones) (the cross, the mandala, the cakra, three, four, the lotus, the egg, the palanquin, an eye, elements, colors, etc., on the one hand, and a more indefinite series of signs, on the other); 3) general works on symbolization and the structure of the symbol (cf., e.g., Werner and Kaplan 1963). For a more complete bibliography, see Lurker 1964-68; Jobes 1962; Hermann 1961; etc. 89 Cf. the remarkable article by Weinreich 1963, and several of the subsequent publications by the same author. Cf. especially the ideas about the meaning relationships among the components of complex signs which are themselves constructed by means of both linking and nesting. Cf. the formulation of the hypothetical universal (p. 134), the significance of which is very great also in an analysis of the semantic level of folklore texts: 'In all languages a combination of signs takes the form of either linking or nesting, and all languages use both patterns in kernel sentences. N o further patterns are introduced by transformations. While the number of levels is not theoretically limited, linking on more than three and nesting on more than four is very rare.'

FOLK POETRY: GENERAL PROBLEMS

723

in particular of epic segmentation, and the problem of memorization, in the first place, and the operational principles for separating out the motifs, in the second place (cf. K. Horalek 1968). It is not coincidental that folklore studies have recently begun to appear which focus attention on the kind of role that is played by semantic categories (e.g. time, space, miracle, etc.) in the plot structure of the text (among them, the oral-poetic text), cf. the interesting works of S. Ju. Nekljudov (1966, 1969).90 It was pointed out above that folklore texts at the present time may be viewed as one of the basic proving grounds where linguistic structural semantics is sharpening its methods. Along with this, semantic research of these texts cannot circumvent a whole series of problems connected with the original locale. For many traditions such structures are renewed (and in many archaic ones they exist even now) within the limits of known genres (folktale, song, riddle, praise words, cant expressions, obscenities, etc.), which propose a DIALOGIC QUESTION-AND-ANSWER form (often with numerical segmentation as a compositional feature: f i r s t . . . second . . . third . . . ) . This form, often realized in literary poetic duels, is not infrequently constructed as a sequential, stepped narrowing of the theme 91 — from maximally wide, cosmological to cultural-historical and narrowly everyday. It is completely evident (and this was shown in many other studies) that such schemes reflect in their sources the literary part of archaic rituals of the New Year type when the world falls in, time stops, all the oppositions which have organized this world are removed, a duel takes place with chaotic forces, and, after the victory over them, the former picture of the world is restored. Undoubtedly, in an analysis of folklore or mythological texts taken in such a locale, the researcher should remember the determinability of the semantics of these texts by extralinguistic (and extra-folkloristic) systems, on the one hand, and the connection of the diachronic aspect with deep structures of the given text, on the other. In this area semantic analyses necessarily coincide with the same trends in science which are occupied with the study of cultural patterns, values, the problem of social behavior, and semiotics in the whole totality of systems which comprise it and are interrelated. 9. In this survey attention was directed particularly to a certain series of ideas and descriptive methods in contemporary folklore studies which takes shape under the influence or in close cooperation with structural linguistics. Therefore, naturally, many valuable studies remain undiscussed if their basic interest is in other spheres. But, be that as it may, the desire here was to stress the OPTIMISTIC viewpoint of the prospects for developing folkloristics and, particularly, that sector of it which studies poetic texts. Now scarcely anyone will decide to dispute seriously and 90

In numerous studies of the epos there also occurs (not infrequently nor fully consciously) the thought about the connection of semantic and plot components. Cf. particularly Bowra 1961; Meletinskij 1968; and a large number of other works. 91 Cf. an analogous method in Russian folk songs (with respect to images) noticed already by B. M. Sokolov.

724

v . N. TOPOROV

argumentatively the role of structural and mathematical92 methods in the study of folklore. REFERENCES and S. THOMPSON. 1 9 6 4 . The types of the folktale. A classification and bibliography. FFC 184. Helsinki. ABRAHAMS, R . D . 1966. Patterns of structure and role relationships in the Child ballads in the United States. J A F 7 9 . 4 4 8 - 6 2 . Actes du 3e Congrès International d'Histoire littéraire, Lyon, 1939.1940. Helicon 2. ALEKSYNAS, K . 1 9 6 8 . Veiksmazodziij samplaikos lietuviij liaudies dainose. Literatura ir kalba 9.354-86. Vilnius. AMZULESCU, A . I. 1966. Observatii introductive la cercetarea cîntecului 'propriu zis'. Revista di etnografie $i folclor 1 1 . 9 7 - 1 0 8 . ANDERSON, W . 1 9 5 1 . Ein volkskundliches Experiment. F F C 1 4 1 . Helsinki. . 1956. Eine neue Arbeit zur experimentalen Volkskunde. FFC 168. Helsinki. ANDREJCIN, L. 1961. O poetike bolgarskix narodnyx pesen. Poetics — Poetyka — Poètika, 5 0 1 - 4 . Warsaw and The Hague. ARMSTRONG, R . P. 1 9 5 9 . Content analysis in folkloristics. Trends in content analysis, 151-70. Urbana. AUSTERLITZ, R . 1958. Ob-Ugric metrics. The metrical structure of Ostyak and Vogul folk-poetry. FFC 174. Helsinki. . 1961a. The identification of folkloristic genres (based on Gilyak materials). Poetics — Poetyka — Poètika, 505-10. Warsaw and The Hague. . 1961b. Parallelismus. Poetics, 439-43. Warsaw. BALA§OV, D. M. 1962. Postanovka voprosa o ballade v russkoj i zapadnoj fol'kloristike. Voprosy literatury i narodnogo tvorcestra. Petrozavodsk. . 1963. Russkaja narodnaja ballada. Leningrad. . 1966. Istorija razvitija zanra russkoj ballady. Petrozavodsk. BALYS, J . 1958. Lithuanian folksongs in America: Narrative songs and ballads. Boston. BARAUSKIENË. 1968. Sisteminis lietuviij liaudies daimj tekstij katalogas. Literatura ir kalba 9 . 3 8 9 - 4 0 5 . Vilnius. BASCOM, W . 1 9 6 4 . Folklore research in Africa. JAF 7 7 . 1 2 - 3 1 . . 1965a. The forms of folklore: Prose narratives. JAF 78.3-20. . 1965b. African folklore and literature. The Africa world, a survey of social research in Africa, ed. by R. A. Lystad. New York.

AARNE, A . ,

M

In connection with the latter there arises the increasingly urgent problem of working out the symbolic metalanguage and a system of recording folklore texts. The first attempts in this field have already been made, and there immediately arises the problem of correlating these systems with each other.

FOLK POETRY: GENERAL PROBLEMS

725

R., ed. 1962. Sens et usages du terme structure dans les sciences humaines et sociales. The Hague. BAUSINGER, H. 1967. Bemerkungen zum Schwank und seinen Formentypen. Fabula 9.118-36. . 1968. Formen der Volkspoesie. Berlin. BENES, B . 1966-67. Lidové vyprâveni na moravskych kopacich (pokus o morfologickou analyzu povërecnych povidek podle systému V. Proppa). Slovâcky Narodopisny Sbornik pro moravskoslovenské pomezi, 41-71. BENVENISTE, E. 1964. Les niveaux de l'analyse linguistique. PICL 9. BERELSON, B . 1952. Content analysis in communication research. Glencoe, 111. BERRY, F. 1958. Poet's grammar. London. BÎRGU-GEORGESCU, L. 1966. Contaminarea — un procedeu artistic de compozi^ie. Studii de poeticä si stilisticä, 69-80. Bucuresti. BLACKBURN, T . 1 9 6 4 . Some examples of Ponapean folklore. J A F 8 0 . 2 4 7 - 5 4 . B0DKER, L. 1 9 6 6 . Folk literature (Germanic). International dictionary of regional European ethnology and folklore 2. Copenhagen. BOGATYREV, P. G . 1 9 6 2 . O jazyke slavjanskix narodnyx pesen v ego otnosenii k dialektnoj reci. VJa 3 . 7 5 - 8 6 . . 1963. Slovackie epiceskie rasskazy i mifo-epicskie pesni. Moskva. BOGATYREV, P . , and R . JAKOBSON. 1 9 2 9 . Die Folklore als eine besondere Form des Schaffens. Donum Natalicum Schrijnen. Nijmegen & Utrecht. . 1931. K problème razmezevanija fol'kloristiki i literaturovedenija. Lud Slowiafiski 2. BOGGS, R . S. 1 9 4 9 . Folklore classification. SFQ. . 1950. Types and classification of folklore. Standard dictionary of folklore, mythology and legend 2.1138-47. New York. BOSWELL, G . 1 9 6 7 . Reciprocal controls exerted by ballad texts and tunes. J A F BASTIDE,

80.169-74.

BOWRA, C. M. 1961. Heroic poetry. London. BRÄILOIU, C. 1946. Sur une ballade roumaine (La Miori(a). Genève. BRAND, P . 1 9 6 2 . The ballad mongers: Rise of modern folk song. New York. BRÄTULESCU, M. 1966. Cîteva tipuri de metafora în folclor. Studii de poeticä si stilisticä, 81-93. Bucuresti. BRAUN, M . 1 9 6 3 . Zum Problem der serbokroatischen Volksballade. Opera Slavica 4 . 1 5 1 - 7 4 . Göttingen. BREDNICH, R. W. 1968. Der Plan eines europäischen Balladentypenindex. Zbornik XII Kongresa jugoslovanskih folkloristov, Celje, 1965, 363-9. Ljubljana. BRÉMOND, C . 1 9 6 4 . Le message narratif. Communications 4 . 4 - 3 2 . . 1966a. La logique des possibles narratifs. Communication 8.60-71. . 1966b. Combinaisons syntaxiques entre fonctions et sequences narratives. Paris.

726

V. N. TOPOROV

BRONSON, B. H. 1944. The interdependence of ballad tunes and texts. California Folklore Quarterly 3.185-207. . 1946. Folksong and the modes. The Musical Quarterly 32.37-49. . 1949. Mechanical help in the study of the folk song. JAF 62.81-6. . 1959. Toward the comparative analysis of British and American folk tunes. JAF 72.164-91. . 1959-62. The traditional tunes of the Child ballads. Vols. I-II. Princeton. BUDZYK, K. 1956. Rozwoj jçzyka i kszaltowanie siç stylôw pismiennictwa oraz niektôrych form artystycznych literatury piçknej. Pamiçtnik Literacki 47/1-2. BURKE, KENNETH. 1941. The philosophy of literary form. Baton Rouge. . 1945. A grammar of motives. New York. CARACOSTEA, D. 1948. Esquisse d'une typologie de la ballade populaire roumaine. Langue et littérature, Bulletin de la Section Littéraire, 4.5—11. CERNOV, I. A. 1965. O strukture russkogo ljubovnogo zagovora. Trudy po znakovym sistemam 2.159-72. Tartu. CHADWICK, N. K. 1969. The epic poetry of the Turkic peoples of Central Asia. Oral epics of Central Asia, ed. by N. K. Chadwick and V. Zirmunsky. Cambridge. [A reworked version of the section from Chadwick's The growth of literature, Cambridge 1940.] CHASTAING, M. 1958. Le symbolisme des voyelles: les 'i'. Journal de psychologie normale et pathologique 55.403-23. . 1962. La brillance des voyelles. ArchL 14.1-13. CHILD, F. J. 1882-98. The English and Scottish popular ballads. Vols. 1-5. Boston. CHITIMIA, J. C. 1957. Poezia popularâ narativâ, Balada. Studii çi cercetâri de istorie literarâ si folclor 6.595-651. CIZEVSKY, D. 1947. On alliteration in ancient Russian epic literature. Russian Epic Studies, Memoir 42. CJURLIONITE, JA. 1966. Povestvovatel'nye pesni-ballady. Litovskoe narodnoe pesennoe tvorcestvo, 158-65. Moskva-Leningrad. COFFIN, T. P. 1950. The British traditional ballad in North America. 2nd rev. ed. 1963. Philadelphia. . 1960 (1963). Critical analyses. Ancient ballads, traditionally sung in New England, ed. by Helen Hartness Flanders. Philadelphia. COFFIN, T. P., and E. M. LEACH, eds. 1961. The critics and the ballad. Carbondale, m. COLBY, B. N. 1966a. Cultural pattern in narrative. Science 151.793-8. . 1966b. The analysis of culture content and the patterning of narrative concern in texts. AmA 68.374-88. COLBY, B. N . , G . A . COLLIER, a n d S. K . POSTAL.

1963.

C o m p a r i s o n of t h e m e s in

folktales by the General Inquirer System. JAF 76.318-23. DAL, E. 1962. Scandinavian ballad research today. Scandinavica 1.5-16.

FOLK POETRY: GENERAL PROBLEMS

727

1961. Les symboles chrétiens primitifs. Paris. 1 9 6 3 . Schema di classificazione del materiale folclorico. Tradizione 1. Padova. DAVENPORT, W. H. 1953. Marshallese folklore types. JAF 66.219-37. DAVIE, D . 1955. Articulate energy. An inquiry into the syntax of English poetry. London. DOLEZEL, L . 1 9 6 6 . Perspektivy strukturâlni analyzy literârniho dfla. Struktura a smysl literârniho dila, 7 0 - 8 6 . Praha. DOLEZEL, L . , and K . HAUSENBLAS. 1 9 6 1 . O sootnosenii poetiki i stilistiki. Poetics — Poetyka — Poètika, 3 9 - 5 2 . Warsaw and The Hague. DONOHUE, J. J. 1943, 1949. The theory of literary kinds. I: Ancient classifications of literature. II: The ancient classes of poetry. Dubuque, Iowa. DORSON, R . M. 1963. Current folklore theories. CAnthr 1963.93-6. DUCHESNE-GUILLEMIN, J. 1961. Symbolik des Parsismus. Symbolik der Religionen, ed. by F. Hermann. Stuttgart. DUNDES, A. 1 9 6 2 . From etic to emic units in the structural study of folktales. DANIÉLOU, J .

D'ARONCO, G .

J A F 75.94-105.

. 1963. Structural typology in North American Indian folktales. SJA 19.126-9. . 1964. The morphology of North American Indian folktales. FFC 195.1134. . 1966. The study of folklore. New York. DUNDES, A . , and R . A . GEORGES. 1 9 6 3 . Toward a structural definition of the riddle. J A F 7 6 . 1 1 1 - 1 8 . EHRICH, E . L. 1959. Die Kultsymbolik im Alten Testament und im nachbiblischen Judentum. Stuttgart. ELIADE, M. 1970. De Zalmoxis à Gengis-Khan. Études comparatives sur les religions et le folklore de la Dacie et de l'Europe Orientale. Paris. EMENEAU, M. B. 1958. Oral poets of South India — the Todas. JAF 71.312-24. EMRICH, W. 1943. Die Symbolik von Faust II. Sinn und Vorformen. Berlin. ERDELY, S. 1962. Classification of Hungarian folksongs. The Folklore and Folk Music Archivist 5/3.1-2. ERETESCU, C. 1 9 6 6 . Compozitie în balada 'Äi trei fra^i eu nouä zmei'. Revista de etnografie si folclor 1 1 . 1 5 2 - 6 4 . EVGEN'EVA, A. P. 1963. Ocerki po jazyku russkoj ustnoj poezii. Moskva and Leningrad. FISCHER, J . L . 1960. Sequence and structure in folktales. Men and cultures, ed. by A. F. C. Wallace, 442-6. Philadelphia. . 1963. The sociopsychological analysis of folktales. CAnthr 4.235-95. . 1966. A Ponapean Oedipus tale. Anthropologist looks at myth, 109-24. Austin and London. FOCHI, A. 1964. Miorita-tipologie, circulatie, geneza, texte. Bucuresti.

728

V. N. T0P0R0V

I. 1963. Die Metaphern in der Phonetik. The Hague. H. 1 9 2 6 . Der kallimachische und der homerische Hexameter. Nachrichten von der Königlichen Gesellschaft der Wissenschaften zu Göttingen, Phil.-Hist. K L , 1 9 7 - 2 2 9 . FRIEDMAN, A . B . 1961a. The ballad revival: Studies in the influence of popular on sophisticated poetry. Chicago. . 1961b. The formulaic improvisation theory of ballad tradition — a counterstatement. JAF 74.113-15. GEORGE, A. L. 1959. Quantitative and qualitative approaches to content analysis. Trends in content analysis, 7-32. Urbana. GIMBUTAS, M . 1 9 5 8 . Ancient symbolism in Lithuanian folk art. GOLF, B. L. 1963. Symbols of prehistoric Mesopotamia. New Haven and London. GOLOPENTIA(-ERETESCU), S. 1966. Analiza semanticä a poeziei folclorice. Revista de etnografie si folclor 11.109-22. Bucureçti. GONDA, J . 1961. Stilistic repetition in the Veda. Amsterdam. GÖRNY, W. 1 9 6 1 . Text structure, the background of language structure. Poetics, 25-34. Warsaw. GREIMAS, A. J. 1963. La description de la signification et la mythologie comparée. l'Homme 3/3.51-66. . 1966a. Eléments pour une théorie de l'interprétation du récit mythique. Communications 8.28-59. . 1966b. Sémantique structurale. Recherche de methode. Paris. . 1970. Du sens. Paris. GRIAULE, M. 1952. Réflexions sur les symboles soudanais. Cahiers internationales de sociologie 13.8-30. . 1957. Symbolisme d'un temple totémique soudanais. Paris. GROOT, A . W . DE. 1957. Phonetics in its relation to aesthetics. Manual of phonetics, ed. by L. Kaiser, 385-99. Amsterdam.

FONAGY,

FRANKEL,

HAMMERSCHMIDT, E . , P . HAUPTMANN, P . KRXGER, L . OUSPENSKY, a n d H . J . SCHULZ.

1962. Symbolic des orthodoxen und orientalistischen Christentums. Stuttgart. HAMMOND, M. 1961. Poetic syntax. Poetics, 475-82. Warsaw. HAVLIKOVÄ, C. 1967. Pouziti schémat pro vyznaceni kontaminacniho a variacniho procesu (Pfispëvek ke katalogizaci textovych variant lidovych pisni). Cesky lid 54/5.285-96. HEMPEL, C. Y., and P . OPPENHEIM. 1936. Der Typusbegriff im Lichte der neuen Logik. Leiden. HERMANN, F . 1 9 5 8 . Die Dichtung der Naturvölker. Lehrbuch der Völkerkunde, ed. by L. Adam and H. Trimborn. 3rd ed. Stuttgart. , ed. 1961. Symbolik in den Religionen der Naturvölker. Stuttgart.

FOLK POETRY: GENERAL PROBLEMS

729

M. J., and F . S. HERSKOVITS. 1 9 5 8 . Dahomean narrative, a crosscultural analysis. Evanston. HERZOG, G. 1949. Song: Folk song and the music of folk song. Funk and Wagnalls standard dictionary . . . 2 . 1 0 3 2 - 5 0 . HONKO, L . 1 9 6 4 . Memorates and the study of folk beliefs. Journal of the Folklore Institute, Vol. I. . 1968. Genre analysis in folkloristics and comparative religion. Temenos 3.48-66. HORAK, J . 1 9 5 6 . Slovenské l'udové balady. Bratislava. . 1961. Les ballades populaires slaves. RESI 39.31-42. HORAK, J . , and K . PLICKA. 1 9 6 5 . Zbojnicke piesne slovenského l'udu. Bratislava. HORALEK, K . 1 9 6 4 . Pohàdkoslovné studie. Praha. . 1968. Pfispévek k sémantice epickych textu. Slavica Pragensia X, Philologica 1-3.337-52. HYMES, D . H . 1960. Phonological aspects of style: Some English sonnets. Style in language, ed. by T. A. Sebeok, 109-31. Boston and New York. , ed. 1965. The use of computers in anthropology. The Hague. INGARDEN, R. 1931. Das literarische Kunstwerk. Halle. . 1939. Das Form-Inhalt-Problem im literarischen Kunstwerk. Helicon I.5 Iff. . 1947. Szkice z filozofii literatury 1. Lòdi. . 1957. Studia z estetyki. Warsaw. . 1961. Poetik und Sprachwissenschaft. Pieties — Poetyka — Poètika, 3-9. Warsaw and The Hague. IONESCU, L . 1 9 6 6 . Paralelismul in lirica popularä. Revista de etnografie $i folclor II.48-68. [= Studii de poetica si stilistica, Bucure§ti.] IVANOV, V . V . , and V . N . TOPOROV. 1 9 6 5 . Slavjanskie jazykovye modelirujuscie semioticeskie sistemy. Moskva. . 1966. K rekonstrukcii praslavjanskogo teksta. Slavjanskoe jazykoznanie. V Mezdunarodnyj s"ezd slavistov, 88-158. Moskva. JACOBS, M. 1959a. Folklore. The anthropology of Franz Boas, ed. by W. Goldschmidt. . 1959b. The content and style of an oral literature: Clackamas Chinook myths and tales. VFPA 26. New York. . 1956. Thoughts on methodology for comprehension of an oral literature. Men and cultures, ed. by A. F. C. Wallace, 123-9. Philadelphia. . 1966. A look ahead in oral literature research. JAF 79.413-27. JAKOBSON, R. 1 9 4 5 . On Russian folktales. On Russian fairy tales. New York. . 1952. Studies in comparative Slavic metrics. OS1P 3.21-66. [Cf. the variant Slavic epic verse. Studies in comparative metrics. 1966.414-63.] • . 1960. Closing statement: Linguistics and poetics. Style in language, ed. by T. A. Sebeok. Boston and New York. HERSKOVITS,

730

V. N. TOPOROV

. 1961. Poezija grammatiki i grammatika poezii. Poetics — Poetyka — Poetika. Warsaw and The Hague. . 1962. O sootnosenii mezdu pesennoj i razgorovnoj narodnoj recju. VJa 3.87-90. . 1966. Selected writings IV. The Hague. JAKOBSON, R . , and J . LÖTZ. 1 9 4 1 . Axiomatik eines Verssystems am mordwinische Volkslied dargelegt. Stockholm. [English version 1 9 5 2 . ] JOBES, G. 1962. Dictionary of mythology, folklore and symbols. Vols. 1-3. New York. JOKIMAITIENE, P. 1967. Lietuvhj liaudies vaiky dainos ir jij klasifikacija. Lietuvos TSR Mokslij Akademijos Darbai. Serija A. 1/23.151-62. . 1968. Lietuviij liaudies balades. Dainuojomosios tautosakos klausimai. Literatüra ir kalba 9.297-350. Vilnuis. JOLLES, A. 1929. Einfache Formen. 2nd ed. 1956. Halle. JONES, J . H . 1 9 6 1 . Commonplace and memorization in the oral tradition of the English and Scottish ballads. J A F 7 4 . 9 7 - 1 1 2 . JONES, L . G. 1965. Tonality structure in Russian verse. I J S L P 9.125ff. JONSSON, B. R . 1 9 6 1 . Volkslied. Schwedische Volkskunde, 4 9 1 - 5 0 9 . Stockholm. [Bibliographical and historical research.] . 1967. Svensk Balladtradition I: Balladkällor och balladtyper. Stockholm. KARLGREN, B. 1963. Sound and symbol in Chinese. Oxford. KATONA, J . 1969. Kiserlet egyszakaszos nepdalszövegek vältozatkepzödesenek mennyisegi vizsgälatära. Studia Ethnographica 5.398ff. KATZ, J . J . , and P . M. POSTAL. 1964. An integral theory of linguistic description. Cambridge, Mass. KAYSER, W. 1948. Das sprachliche Kunstwerk. Bern. . 1958. Literarische Wertung und Interpretation. Bern. KAZLAUSKIENE, B . 1 9 5 9 . Dainuojamosios lietuviij tautosakos teksty klasifikavimo sistema. Lietuvos TSR Mokslij Akademijos Darbai. Serija A. 2 / 7 . 2 3 3 - 5 4 . . 1967. Vestuviniij lietuviij liaudies dainij klasifikavimas. Lietuvos TSR Mokslij Akademijos Darbai, Serija A. 1/23.135^9. KEMPPINEN, J . 1 9 5 4 . The ballad of Lady Isabel and the False Knight. . 1957. Lunastettava neito, Vertaileva balladitutkinus. Helsinki. KIEFER, E. S. 1947. Albert Wesselski and recent folktale theories. Bloomington. KIRDAN, B. P . 1962. Ukrainskie narodnye dumy (XV-nacalo XVII v.). Moscow. . 1965. Ukrainskij narodnyj epos. Moscow. KIRFEL, W. 1959. Symbolik des Buddhismus. Stuttgart. KIRK, G. S. 1962. The songs of Homer. Cambridge. . 1966a. Studies in some technical aspects of Homeric style. I: The structure of the Homeric hexameter, II: Verse-structure and sentence-structure in Homer. YCS 20.73-152. . 1966b. Formular language and oral quality. YCS 20.153-216.

FOLK POETRY: GENERAL PROBLEMS

731

C. 1960. The use of typology in anthropological theory. Men and cultures, ed. by A. F. C. Wallace. Philadelphia. KOLPAKOVA, N. P. 1962. Russkaja narodnaja bytovaja pesnja. Moskva-Leningrad. KÖNGAS, E . - K . , and P . MARANDA. 1 9 6 2 . Structural models in folklore. Midwest Folklore 1 2 . 1 3 3 - 9 2 . KÖNIG, O. 1929. Das Volk erzählt. Wien. KOTLJAR, E. S. 1970. O nekotoryx certax fol'klora Afriki juznee Saxary. Fol'klor i literatura narodov Afriki. Sbornik Statej, 167-83. Moscow. KURYLOWICZ, J. 1947. Jçzyk poetycki ze stanowiska lingwistycznego. Sprawozdania Wroclawskiego Towarzystwa Naukowego 2.4-11. LAFORTE, C. 1958. Le catalogue de la chanson folklorique française, Publications des archives de folklore de l'Université Laval. Québec. . 1960. La structure et la forme. Cahiers de l'Institut de science économique appliquée, Serie M. 7.1-36 (mars). LANGER, S. 1953. Feeling and form. New York. LAWS, G . M . , J R . 1 9 5 7 . American balladry from British broadsides. A guide for students and collectors of the traditional song. Philadelphia. . 1964. Native American balladry: A descriptive study and a bibliographical syllabus. Rev. ed. Philadelphia. LEVIN, J U . I . 1 9 6 5 . Struktura russkoj metafory. Trudy po znakovym sistemam 2 . 2 9 3 - 9 . Tartu. . 1969. Russkaja metafora: sintez, semantika, transformacii. Trudy po znakovym sistemam 4.290-305. Tartu. LEVIN, S. R. 1962. Linguistic structures in poetry. The Hague. . 1963. Statistische und determinierte Abweichung in poetischer Sprache. Lingua 12.276-90. LÉVI-STRAUSS,C. 1958. Anthropologie structurale. Paris. . 1960. L'analyse morphologique des contes russes. IJSLP 3.122-49. . 1964-68. Les mythologiques. Paris. LINTUR, P. V. 1963. Narodnye ballady Zakarpat'ja i ix zapadnoslavjanskie svjazi. Kiev. LIST, G. 1968. Toward the indexing of ballad texts. JAF 81.44-61. LITTLETON, C. SCOTT. 1965. A two-dimensional scheme for the classification of narratives. JAF 78.21-7. LORD, A. B. 1948. Homer and Huso III: Enjambement in Greek and Southslavic song. TAPA 79.113ff. . 1951. Composition by theme in Homer and Southslavic epos. TAPA 82.71ff. . 1956. The role of sound patterns in Serbocroatian epic. For Roman Jakobson, 301-5. The Hague. . 1960. The singer of tales. Cambridge, Mass. KLUCKHOHN,

732

V. N. TOPOROV

. 1962. A comparative analysis. [Introduction to] Umbundu. Folk tales from Angola. Boston. LÖTZ, J. 1942. Notes on structural analysis in metrics. Helicon 4.119-46. . 1954. Kamassian verse. JAF 67.369-78. . 1956. A notation for the Germanic verse line. Lingua 6.1-7. . 1960. Metric typology. Style in language, ed. by T. A. Sebeok 135-48. Boston and New York. LURKER, M. 1964-68. Bibliographie zur Symbolkunde. Vols. 1-3. Baden-Baden. LYNCH, J. J. 1953. The tonality of lyric poetry: An experiment in method. World 9.211-24.

MALINOWSKI, B. 1922. Argonauts of the Western Pacific. An account of native enterprise and adventure in the Archipelagoes of Melanesian New Guinea. New York. MARANDA, E. 1966. What does a myth tell about society. Radcliffe Institute Seminars. Cambridge, Mass. MARANDA, P. 1966. Computers in the bush: Tools for the automatic analysis of myth. Proceedings of the Annual Meeting of the American Ethnological Society. Philadelphia. MARCHAL, H. 1956. Le symbolisme dans les arts de l'Inde. France-Asie 5/VI.708-23.

MARCUS, S. 1965. Un précurseur de la poétique mathématique: Pius Servien. RLing 10.415-24.

MASSON, D. J. 1961. Sound-repetition terms. Poetics — Poetyka — Poètika, 189-99. Warsaw and The Hague. MAUNG THAN SEIN, and A. DUNDES. 1964. Twenty-three riddles from Central Burma. J A F 7 7 . 6 9 - 7 5 .

MAYENOWA, M. R. 1949. Poetyka opisowa. Warsaw. MEIER, J. 1941. Die Ballade'Des Grafen Töchterlein'und der Roman von Flore und Blancheflur. Jahrbuch für Volksliedforschung 7.5-10. MELETINSKIJ, E. M. 1963. Proisxozdenie geroiceskogo eposa. Moskva. . 1966. O strukturno-morfologiceskom analize skaski. Tezisy dokladov vo Vtoroj letnej skole po vtoricnym modelirujuscim sistemam, 37-40. Tartu. . 1968. 'Edda' i rannie formy eposa. Moskva. , 1969. Strukturnotipologiceskoe izucenie skazki. Morfologija skazki, by V. Ja. Propp. 2nd ed. Moskva. . 1970a. Die Ehe im Zaubermärchen. Ihre Funktion und ihr Platz in der Struktur des Märchens. AEH 19.281-92. . 1970b. Problème de la morphologie historique du conte populaire. Semiotica 2/2.128-34. . 1970C. Klod Levi-Stross i Strukturnaja tipologija mifa. Voprosy Filosofii 10.165ff. MELETINSKIJ, E . M.,

S. J u . NEKLJUDOV, E . S. NOVIK,

and

D . M . SEGAL.

1969.

FOLK POETRY: GENERAL PROBLEMS

733

Problemy strukturnogo opisanija volsebnoj skazki. Trudy po znakovym sistemam IV.86-135. Tartu. . 1971. Esce raz o probleme strukturnogo opisanija volsebnoj skazki. Trudy po znakovym sistemam V. Tartu. MELICHERCIK, A . 1 9 5 2 . Jânosikovskâ tradicia na Slovensku. Bratislava. . 1956. Juraj Jânosik hrdina protifeudâlneho odboja slovenského l'udu. Praha. MENÉNDEZ-PIDAL, R. 1936. Das Fortleben des Kudrungedichtes (Der Ursprung der Ballade). Jahrbuch für Volksliedforschung 5.85-122. MEYERSON, J. 1 9 5 6 . Le temps, la mémoire, l'histoire. J. de Psych 5 3 . MORRIS, C . 1 9 3 8 . Foundations of the theory of signs. International Encyclopedia of Unified Science 1 / 2 . 1 - 5 9 . . 1946. Signs, language and behavior. New York. MÜLLER, G. 1943. Morphologische Poetik. Helicon 5. . 1944. Die Gestaltfrage in der Literaturwissenschaft und Goethes Morphologie. Halle. MUSURILLO, H. 1961. Symbol and myth in ancient poetry. New York. NATHHORST, B. 1968. Genre, form and structure in oral tradition. Temenos 3.128-35. . 1969. Formal or structural studies of traditional tales. Stockholm Studies in Comparative Religion 9. Stockholm. NEKLJUDOV, S . J U . 1 9 6 6 . K voprosu o svjazi prostranstevenno-vremennyx otnosenij s sjuzetnoj strukturoj byliny. Tezisy dokladov vo vtoroj letnej skole po vtoricnym modelirujuscim sistemam Tartu. . 1969. Cudo v byline. Trudy po znakovym sistemam 4.146-58. Tartu. NEWMAN, N . , and W . POPPER. 1 9 1 8 - 2 3 . Studies in Biblical parallelism. University of California. NIKIFOROV, A. I. 1928. K voprosu o morfologiceskom izucenii narodnoj skazki. Sbornik statej v cest' A. I. Sobolevskogo, 172-8. Leningrad. NYGARD, H. O. 1958. The ballad of Heer Halewijn: Its forms and variations in Western Europe. Helsinki and Knoxville, Tenn. . 1965. Ballad Folkevise, chanson populaire. Folklore studies in honor of Arthur Palmer Hudson. Chapel Hill, N.C. OPPEL, H . 1 9 4 7 . Morphologische Literaturwissenschaft. Goethes Ansicht und Methode. Mainz. OZOLS, A . 1 9 6 1 . Latviesu tautasdziesmu valoda. Riga. PAREDES, A. 1 9 5 8 . 'With his pistol in his hand', a border ballad and its hero. Austin. PARRY, M. 1928. L'épithète traditionelle dans Homère. Paris. . 1929. The distinctive character of enjambement in Homeric verse. TAPA 60.200ff. . 1930-32. Studies in the epic techniques of oral verse-making. I: Homer

734

V. N. TOPOROV

and Homeric style. II: The Homeric language as the language of an oral poetry. HSPh 41.73ff. (1930); 43.1ff. (1932). . 1933. Whole formulaic verses in Greek and Southslavic heroic song. TAPA 64.179ff. PARRY, M., and A. B. LORD. 1953-54. Serbocroatian heroic songs. Vols. I-II. Cambridge and Belgrade. PEARSON, N. H. 1941. Literary forms and types. New York. PELC, J. 1961. Semantic functions as applied to the analysis of the concept of metaphor. Poetics — Poetyka — Poetika, 305-39. Warsaw and The Hague. PENTIKÄINEN, J. 1968. Grenzprobleme zwischen Memorat und Sage. Temenos 3.136-67.

PERMJAKOV, G. L. 1968. Izbrannye poslovicy i pogovorki narodov Vostoka. Moskva. PEUKERT, H. 1960. Bemerkungen zur künstlerischen Gestalt der russischen Volksepik. ZS1 5. . 1961. Die Funktion der Formel im Volkslied. Poetics — Poetyka — Poetika, 525-36. Warsaw and The Hague. PIKE, K. L. 1954-55. Language in relation to a unified theory of the structure of human behavior. 2 Parts. Glendale, HI. POLIVANOV, E. D. 1918. Formal'nye tipy japonskix zagadok. Sbornik Muzeja Antropologii i Etnografii 5/1.371-4. POLLOK, K. H. 1964. Studien zur Poetik und Komposition des balkanslawischen lyrischen Volksliedes, I: Das Libeslied. Göttingen. POP, M. 1967. Metode noi in cercetarea structurii basmelor. Folclor Literatur 5-12. Timijoara. . 1968. Der formelhafte Charakter der Volksdichtung. DJbVK 14.1-15. POPE, J. C. 1966. The rhythm of Beowulf. London. POPPE, N. 1958. Der Parallelismus in der epischen Dichtung der Mongolen. UAJb 30. PROPP, V. JA. 1946. Istoriceskie korni volsebnoj skazki. Leningrad. . 1961. O russkoj narodnoj liriceskoj pesne. Narodnye liriceskie pesni, 5 68. Leningrad. . 1964a. Principy klassifikacii fol'klornyx zanrov. SovEtn 4.147-54. . 1964b. Zanrovyj sostav russkogo fol'klora. Russkaja literatura 4.58-76. . 1966. La morfologia della fiaba. Torino. . 1968. The morphology of the folktale, 2nd ed. (1st Russian ed. 1928.) Austin. PUTILOV, B. N. 1960. Russkij istoriko-pesennyj fol'klor XIII-XVI vekov. Moscow-Leningrad. . 1964. Istoriceskie korni i genezis slavjanskix ballad ob inceste. Moskva. . 1965. Slavjanskaja istoriceskaja ballada. Moscow-Leningrad. RICHARDS, I. A. 1936. The philosophy of rhetoric. New York.

FOLK POETRY: GENERAL PROBLEMS

735

M. 1959. Criteria for style analysis. Word 15.167ff. G. 1965. Structuralism et folklore. IV International Congress for Folk-Narrative Research in Athens, 399-407. Athens. ROSENBERG, N. V. 1967. From sound to style: The emergence of bluegrass. JAF 80.143-50. Russo, J. A. 1966. The structural formula in Homeric verse. YCS 20.217ff. RUXÂNDOUI, P. 1966. Aspectul metaforic al proverbelor. Studii de poetica si stilistica, 93-113. Bucuresti. RYCHNER, J . 1 9 5 5 . La chanson de geste. Essai sur l'art épique des jongleurs. Genève and Lille. SADOWNIK, J. 1 9 5 6 . Z zagadnien klasyfikacji i systematyki polskiej piesni ludowej. Polska sztuka ludowa 6 . 3 4 3 - 5 4 . SAPIR, E . 1 9 2 9 . A study in phonetic symbolism. Journal of Experimental Psychology 1 2 . 2 2 5 - 3 9 . . 1959. Indian legends from Vancouver Island. JAF 72.106-14. SAPORTA, SOL. 1960. The application of linguistics to the study of poetic language. Style in language, ed. by T. A. Sebeok, 82-93. Boston and New York. SAPORTA, SOL, and THOMAS A. SEBEOK. 1959. Linguistic and content analysis. Trends in content analysis, 131-50. Urbana. SAUKA, D . 1 9 7 0 . Tautosakos savitumas ir vertè. Vilnius. SAUKA, L . 1 9 6 9 . Lietuviij vestuvinès dainos. Literatüra ir kalba 9 . 7 - 2 9 7 . Vilnius. SAUSSURE, F. DE. 1964. Les anagrammes. Mercure de France, 243ff. (fevrier). . 1967. J. Starobinski. Les mots sous les mots: Textes inédits des Cahiers d'anagrammes de F. de Saussure. To Honor Roman Jakobson, 1906-1917. The Hague-Paris. . 1968. CFS 24. . 1969. J. Starobinski. Le texte dans le texte. Extraits inédits des Cahiers d'anagrammes de F. de Saussure. Tel Quel 37.3-33. . 1970. Change. No. 6. . 1971a. R. Jakobson. La première lettre de Ferdinand de Saussure à Antoine Meillet sur les anagrammes. Homme XI.2.15-24. . 1971b. J. Starobinski. Les mots sous les mots. Paris. SAVORY, T. H . 1 9 6 7 . The language of science. London. SCHMAUS, A. 1959. Gattung und Stil in der Volksdichtung. Rad Kongresa folklorista Jugoslavie u Varazdinu 1957, 169. Zagreb. SEBEOK, THOMAS A . , ed. 1 9 5 2 . Studies in Cheremis folklore. V o l . 1 . Bloomington. . 1956. Sound and meaning in a Cheremis folksong text. For Roman Jakobson. The Hague. . 1957. Toward a statistical contingency method in folklore research. Studies in folklore, ed. by W. E. Richmond (Indiana University Publications, Folklore Series) 9.130^0. Bloomington. RIFFATERRE,

ROHAN-CSERMAK,

736

V. N. TOPOROV

. 1959a. Folksong viewed as code and message: A Cheremis sonnet. Anthropos 54.141-53. . 1959b. Approaches to the analysis of folksong texts. UAJb 31.392-9. . 1960. Decoding a text: Levels and aspects in a Cheremis sonnet. Style in language, ed. by T. A. Sebeok, 22Iff. Boston and New York. . 1961. Notes on digital calculator as a tool for analyzing literary information. Poetics — Poetyka — Poètika, 571-90. Warsaw and The Hague. . 1965. The computer as a tool in folklore research. The use of computers in anthropology, ed. by Dell Hymes, 255-72. The Hague. SEBEOK, THOMAS A . , and F . I . INGEMANN. 1 9 5 6 . Structure and content analysis in folklore research. Studies in Cheremis: The supernatural, 2 6 1 - 8 . New York. SEBEOK, THOMAS A . , and V . J. ZEPS. 1 9 5 8 . An analysis of structured content with application of electric computer research in psycholinguistics. L&S 1.181-93.

. 1959. On non-random distribution of initial phonemes in Cheremis verse. Lingua 8.370-84. . 1961. Concordance and thesaurus of Cheremis poetic language. The Hague. SEEMANN, E. 1955. Ballade und Epos. Schweizerisches Archiv für Volkskunde. 1955.147-83. SEGAL, D . M. 1 9 6 6 . O svjazi semantiki teksta s ego formal'noj strukturoj. Poetics — Poetyka — Poètika, 15—44. Warsaw and The Hague. SHEPARD, L. 1962. The broadside ballad: A study in origins and meaning. London. SIUTS, H . 1962. Volksballaden — Volkserzählungen. Fabula 5.72-89. SKENDI, S. 1954. Albanian and South Slavic oral epic poetry. Philadelphia. SKINNER, B. F. 1939. The alliteration in Shakespeare's sonnets: A study in literary behavior. The Psychology Record 3.186ff. . 1941. A quantitative estimate of certain types of sound-patterning in poetry. AJPsych 54.64ff. . 1957. Verbal behavior. New York. SOURXAU, E. 1 9 5 0 . Les deux cent mille situations dramatiques. Paris. SPET, G . G . 1922-23. Esteticeskie fragmenty. I - I I I . Petersburg. SPITZER, L. 1 9 4 8 . Linguistics and literary history. Essays in stylistics. Princeton. . 1949. A method of interpreting literature. Northampton, Mass. . 1959. L'archetype de la ballade Miorita et sa valeur poétique. Romanische Literaturstudien. Tübingen. STAIGER, E. 1 9 4 4 . Morphologische Literaturwissenschaft. Trivium 2 . . 1946. Grundbegriffe der Poetik. Zürich. STANKIEWICZ, E. 1960a. Linguistics and the study of poetic language. Style and language, ed. by T. A. Sebeok, 69-81. Boston and New York.

737

FOLK POETRY: GENERAL PROBLEMS

. 1960b. Expressive language. Style and language, ed. by T. A. Sebeok, 96-7. Boston and New York. . 1961. Poetic and non-poetic language in their interrelation. Poetics — Poetyka — Poétika. Warsaw and The Hague. STEINITZ, W. 1 9 3 4 . Der Parallelismus in der finnisch-karelischen Volksdichtung. Helsinki. STOKMAR, M. P . 1 9 5 2 . Issledovanija v oblasti russkogo narodnogo stixoslozenija. Moskva. STONE, P . J . ,

R . F . BALES, J . Z . NAMENWIRTH,

and

D . M . OGILVIE.

1962.

The

general inquirer. A computer system for content analysis and retrieval based on the sentence as a unit of information. B S 7 . 1 - 1 5 . SYDOW, C. W. VON. 1948. Selected papers on folklore. Copenhagen. TARANOVSKI, K. 1963. O vzaimootnosénii stixotvornogo ritma i tematiki. American Contributions to the Fifth International Congress of Slavists. The Hague. . 1965. The sound texture of Russian verse in the light of phonemic distinctive features. IJSLP 9.114-24. TAYLOR, A . 1 9 3 1 . Edward and Sven i Rosengárd. Chicago. Tendances principales de la recherche dans les sciences sociales et humaines. Premiere partie. Paris-Le Haye, 1970. THOMPSON, J. 1961. Linguistic structure and the poetic line. Poetics - Poetyka — Poétika, 167-75. Warsaw and The Hague. THOMPSON, S . 1 9 3 8 . Purpose and importance of an index of types and motifs. Folk-Liv 1 9 3 8 . 1 0 3 - 8 . . 1946(1957). The folktale. New York. . 1955. Narrative motif-analysis as a folklore method. Beiträge zur verglichenden Erzählforschung, ed. by K. Ranke, 2-9. Helsinki. . 1955-58. Motif-index of folk literature, a classification of narrative elements in folktales, medieval romances, exempla, fabliaux, jets-bodes, and local legends. Vols. 1-6. Bloomington. . 1960. Fifty years of folktale indexing. Humaniora. Essays in literature, folklore. Bibliography honoring Archer Taylor on his seventieth birthday, 4 9 57. New York. THOMPSON, S., and J . BALYS. 1 9 5 8 . The oral tales of India. Bloomington. THOMPSON, S. and W . E . ROBERTS. 1 9 6 0 . Types of Indian oral tales. F F C 1 8 0 . Helsinki. TIEGHEM, P. VAN. 1938. La question des genres littéraires. Helicon 1.95-101. TOPOROV, V . N . 1 9 6 3 . Iz oblasti balto-slavjanskix fol'klornyx tekstov. Lietuviq kalbos klausimai 6 . 1 4 9 - 7 6 . . 1966. K analizu neskol'kix poeticeskix tekstov (preimuscestvenno na nizsix urovnjax). Poetics — Poetyka — Poétika, 77-120. Warsaw and The Hague. [IV. Neskol'ko soobrazenij o strukture litovskoj narodnoj ballady.]

738

V. N. TOPOROV

P. 1961. Das Metrum der litauischen Volkslieder. Poetics — Poetyka — Poétika, 119-25. Warsaw and The Hague. TSCHANG TSCHENG-MING. 1937. Le parallélisme dans le vers du Chen King. Paris. TYNJANOV, JU. N. 1924. Problema stixotvornogo jazyka. Leningrad. Uxov, P. D. 1970. Atribucii russkix bylin. Moskva. VANSINA, J. 1958. La valeur historique des traditions orales. Folia Scientifica Africae Centralis, Vol. 4. . 1961. De la tradition orale. Essai de méthode historique. Tervuren. VARGYAS, L. 1955. Betiirim a magyar népkôlteszetben. Filológiai Közlöny 1.401ff. . 1960. Forschungen zur Geschichte der Volksballade im Mittelalter. Die Herkunft der ungarischen Ballade der eigemaurten Frau. Budapest. (Offprint.) . 1960-61,1960-62. Kutatások a népballada kôzépkori tôrténetében. AEH 1960.1-88, 1961.241-94; Ethnographia 1960.163-276, 1962.206-59. . 1964. Rapports internationaux de la ballade folklorique hongroise. Littérature hongroise — Littérature européenne 1964.69-103. . 1967. Researches into the medieval history of folk ballad. Budapest. VERNANT, J.-P. 1 9 6 9 . Mythe et pensée chez les grecs. 2nd ed. Paris. VIANU, T . 1 9 5 7 . Problemele metaforei çi alte studii de stilisticä. Bucuresti. . 1961. Quelques observations sur la métaphor poétique. Poetics — Poetyka — Poètika, 297-304. Warsaw and The Hague. VIET, J. 1965. Les méthodes structuralistes dans les sciences sociales. Paris. VOGT, E . Z . 1 9 6 0 . On the concept of structure and process in cultural anthropology. AmA 62. VOIGT, V . 1969a. Towards balancing of folklore structuralism. AEH 18.247-55. . 1969b. Modellálás a folklorisztikában. Studia Ethnographica 5.355ff. VRABIE, CH. 1965. Sur la technique de la narration dans le conte roumain. IV International Congress for Folk-Narative Research in Athens, 606-15. Athens. VRABIE, G . 1 9 5 7 . Cälätoria fratelui mort sau motivul Lenore în folclorul sud-est european. Limbä çi literatura III. . 1966. Balada populara i românâ. Bucureçti. . 1970. Folclorul. Bucureçti. WAUGH, B. 1966. Structural analysis in literature and folklore. W F 25.153-64. WEHRLI, M . 1 9 5 1 . Allgemeine Literaturwissenschaft. Bern. WEINREICH, U. 1963. On the semantic structure of language. Universals of language, ed. by J. H. Greenberg. 114—71. Cambridge, Mass. WEINRICH, H. 1 9 7 0 . Structures narratives du mythe. Poétique 1 . 2 5 - 3 4 . Paris. WEITZ, M. 1950. Philosophy of the arts. Cambridge, Mass. WELLEK, R., and A. WARREN. 1956. Theory of literature. New York. WERNER, H . , and B . KAPLAN. 1963. Symbol formation. New York. TROST,

FOLK POETRY: GENERAL PROBLEMS

739

A. 1925. Märchen des Mittelalters. Berlin. . 1931. Versuch einer Theorie des Märchens. Reichenberg. WILGUS, D . K . , and L. MONTELL. 1 9 6 8 . Clure and Joe Williams: Legend and blues ballad. JAF 8 1 . 2 9 5 - 3 1 5 . WIMSATT, W . K . , JR., and C . BROOKS. 1 9 5 7 . Literary criticism: A short history. New York. WINTER, M. 1959. Über eine Methode zum Nachweis struktureller Relevanz von Oppositionen distinktiver Merkmale. Symposion Trubetzkoy, 28-44. Basel and New York. XARKINS, V . 1 9 6 3 . O metriceskoj roü slovesnyx formul v serboxorvatskom i russkom narodnom epose. American contributions to the Fifth International Congress of Slavists, Sofia, 1963. The Hague. ZIRMUNSKIJ, V. M. 1964. Ritmiko-sintaksiceskij parallelizm kak osnova drevnetjurkskogo narodnogo epideskogo stixa. VJa 4. [ = Syntaktischer Parallelismus und rhythmische Bindung im altturkischen epischen Vers. Beiträge zur Sprachwissenschaft, Volkskunde und Literaturforschung, Wolfgang Steinitz zum 60. Geburtstag am 28. Februar 1965 dargebracht, 387-401. Berlin, 1965.] WESSELSKI,

F O L K POETRY: H I S T O R Y A N D T Y P O L O G Y

K. HORALEK

I. FOLK POETRY IN ITS HISTORICAL PERSPECTIVE

A. Some General Considerations One characteristic of folk poetry is that similar forms appear in different areas and among different peoples. Certain genres of folk poetry have a decidedly international character. This often comes about as a result of various converging processes, since folk poetry has an expansive force and frequently diffuses from one people to another. Also, the development of human culture very often reveals regularities of a universal character. The study of the emergence of similar phenomena in the folk poetry of different peoples would be unthinkable without international collaboration. There exist international folkloristic publications and folklore associations; a cataloguing system has been drawn up for the needs of comparative studies of folk poetry, along with catalogues of folklore types and folklore motifs. What is surprising however is that to date folklore terminology has not been unified, even in the major world languages. Sometimes the most essential folklore phenomena are understood and described in different ways. Most often the fact that folk poetry exists only in oral form is regarded as its most important feature. That is why, at times, folk poetry is termed oral literature {la littérature orale). But there are reasons to consider certain types of written and printed works as folk poetry. One cannot, after all, exclude certain forms from folk literature merely because they are written or printed. Some languages are beginning to distinguish these two spheres of folk poetry. The German Volksdichtung is being used more and more as a term for all kinds of folk poetry, whereas Volkstümliche Literatur designates only written and printed folk poetry. The latter is represented by folk letters with specific stylization, written amulets, popular songbooks (written and printed), popular reading editions (Volksbücher), leaflet songs (Flugblattlieder, Bankellieder), popular calendars, popular books that explain dreams and predict the future, and so forth. English too distinguishes between 'folk poetry' and 'folk (written) literature'. Russian uses the terms narodnaia slovesnost and narodnaya pismennost, respectively. Similarly Czech uses

742

K. HORÁLEK

the terms lidová slovesnost and lidová literatura. Russian employs, in addition, the term lubocnaya literatura for printed folk literature. In the Russian printed folk prints {lubki), the text is often subordinate to the illustrations, and many folk pictures are reproduced without captions. There are no sharp boundaries between oral and written (or printed) folk poetry. Various transitional forms link folk poetry with learned literature. So-called 'folklorized' songs, adventure tales and stories, certain types of literature for young people, and so on, have a decidedly transitional character. The term 'trivial' is used to describe this type of folk literature. Recently this sphere has come to be examined scientifically, but generally outside the frame of official folkloristics. It should be noted that the term 'folk poetry' should not be confused with the English term 'folkore' if by the latter is designated not just oral expressions of a poetic character but also manners, customs, superstitions, weather lore, pranostics, and so forth. The English term 'folklore' is indeed sometimes used with reference to the oral tradition, but in this sense it is not very suitable, as it may refer to all forms of oral transmission. 'In a culture without writing (termed nonliterate cultures by the anthropologists), almost everything is transmitted orally; and although language, hunting techniques, and marriage rules are passed orally from one generation to another, few folklorists would say that these types of cultural materials are folklore' (Dundes 1965). Folklore ('folk poetry', 'folk literature') is in the European tradition associated with rural culture, but in modern times (especially in the Socialist countries) one also speaks of a working class or urban folkore.1 1

The Russian term fol'klor coincides otherwise with the limitations that W. Bascom suggested for the English use of the term. 'Folklore, to the anthropologist, is a part of culture but not the whole of culture. It includes myths, legends, tales, proverbs, riddles, the texts of ballads and other songs, and other forms of lesser importance, but not folk art, folk dance, folk music, folk costume, folk medicine, folk customs, or folk belief. All of these are unquestionably worthy of study, whether in literate or nonliterate societies . . . All folklore is orally transmitted, but not all that is orally transmitted is folklore' (JAF 66.283f., 1953; cf. Dundes 1965). Soviet textbooks on folklore (for instance Sokolov 1950) only deal with literary forms of folk culture. Later textbooks therefore have specific titles, as for instance the book edited by P. G. Bogatyrev, Russkoe narodnoe poéticeskoe tvorcestvo (1943,1956), and the one by A. M. Novikova and A. V. Kokorev, Russkoe narodnoe poétiíeskoe tvoríestvo (1969). Several essays on folk music have also been included in the annual publication Russkii fol'klor which has appeared since 1956. Sometimes the term folklor is used in Polish to indicate folk poetry. The encyclopaedia Slownik folkloru polskiego (Krzyzanowski 1965) deals mainly with folk literature, in the broad sense of the word, including in this concept printed matter as well, insofar as it is intended for the general reader. This printed matter is not just the writing of a poetic character; the encyclopaedia includes entries such as Senniki (books about dreams), Kalendarz ksiqzka-informator o podstawowych datach i sprawach, przewidywanych na najblizszy rok, PiesA robotnicza rewolucyjna polska (Polish revolutionary workers' songs), and similar other entries. Although the encyclopaedia does not neglect such phenomena as folk music, folk theater, customs, and superstitions, it does not treat them systematically. There is an extensive entry on Polish folk music ( M u z y k a ludowa polska) but no special entry, for instance, about popular musical instruments, about folk music bands, about the scales characteristic of folk music (for example the oft-discussed pentatonic), and so forth. What is especially surprising is that there is no entry, Formula. The encyclopaedia covers quite thoroughly such phenomena as superstitions and customs; for instance, we find entries like Migsopust (carnival), Pianetnik, Pogrzeb, Skarbnik, Skarby zaklgte, yet, on the other

FOLK POETRY

743

In comparative studies of folk poetry a great deal depends on the character of the material examined, on its authenticity, and also on the reliability of the relevant evidence. Early descriptions of folk poetry, whether they were literary texts based on the folk poetry or reports about such poetry, were often made in order to justify one of several conflicting theories. The same is true even of some modern records of folk literature. Therefore, all the sources and documents which constitute the object of the comparative science of folk poetry must be critically examined, and their documentary value and reliability tested. This is often a complex and arduous task. If the text of an older literary work has its counterpart in modern folk poetry, the question of their genetic connection is immediately posed. In such a case we are faced with two basic alternative explanations: either the older literary version represents the beginning of an entire tradition, the later versions or variations being derivatives of this original literary version, or else the older literary version is a derivative from an even older, established oral tradition. There can be more than one literary derivative from such an oral (folk) tradition which, in turn, can influence the oral tradition. The modern oral variant of a certain type need not be the direct continuation of an unbroken series; one can expect to find examples of literary versions that were transferred into the oral tradition. How complicated

hand, no entries relating to old Polish mythology. Somewhat surprising is the inclusion of several heraldic entries, probably with reference to the heraldic tales. In Czech the term folklore is most commonly used to describe folk poetry (under the influence of Russian usage), but apart from this, in a broader context, it also covers folk music and dance. Thus, speeches that are linked to folk rituals (for instance the wedding toasts) and texts of a superstitious character, like magic formulae, have a rather uncertain position. In the official work on Czech and Slovak folk culture "Lidová kultura", Ceskoslovenská vlastiveda III, which appeared in 1968 as a result of collaboration between the Institutes for Ethnography and Folkloristics of the Czechoslovak Academy of Sciences and the Slovak Academy of Sciences (editorin-chief, A. MelicherSik, Professor at Comenius University, Bratislava, who died in 1967), the chapters on folklore, in both the Czech and Slovak sections, deal with folk prose, proverbs, pranostics, and riddles (the Slovak section also contains other sayings), folk theater and children's folklore, folk and folklorized songs, folk music and folk dances. The anthology Slovensky folklor (1959) edited by Melicheriik contains only folk poetry but omits the tunes of folk songs. The anthology of Slavic folklore by Prof.Cvetana Romanska (1963) was compiled in a similar manner. In Serbo-Croatian the meaning of the term folklore is not fixed. The federated organization Savez folklorista Jugoslavije groups together researchers on both material and spiritual folk culture; yet sometimes the term folklore is used in the narrow sense, either as in Russian (folk poetry) or as in Czech (folk poetry + folk music). The Serbian narodno stvaralastvo in Croatian is narodna umjetnost — as a rule it does not cover pictorial art. Wavering in terminology quite often signals a lack of clarity of a theoretical nature. If by folklore we mean only poetic, literate works, then such things as magic formulae are omitted. In the Russian textbook (Bogatyrev 1943,1956), mentioned above, the chapter on magic formulae, "Russkoie narodno-poeticeskoe tvoróestvo doVelikoi oktjabfskoT socialistiSeskoT revoljucii", was included; it was of course also contained in Sokolov's textbook but it is no longer to be found in the 1961 edition. Certain discrepancies in the above-mentioned textbooks cannot be explained from an ideological viewpoint; for instance, they lack chapters on ballad songs, to which Soviet folklorists have paid considerable attention (N. P. Andreev, B. N. Putilov, and others).

744

K. HORALEK

the relations are between the oral and written traditions can be seen from the following examples. The ancient Greek chronicler Herodotus took considerable material for his history from oral tradition. 'In his travels he questioned and listened to all manner of men, Greeks and barbarians alike, though he knew no language but Greek. Like all ancient historians, except perhaps Thucydides, Herodotus is uncritical in his attitude towards his sources. Yet he does not vouch for the truth of everything he repeats; he can often stand aside and leave the reader to judge' (Cassell's Encyclopedia of Literature 1953 :1.1019). Herodotus uses as historical material even narratives of a clearly fictional character and his chronicle is therefore an instructive source for research workers in folk poetry, and especially in folktales. One of Herodotus's most interesting stories is the tale about the thief of the Egyptian King Rhampsenit's treasury. This story was known in Greek literature before Herodotus, but his version no doubt contributed considerably to its dissemination. Later literary versions arose in the Middle Ages in the collection Dolopathos, and these, in turn, influenced the oral tradition. In the Aarne-Thompson international catalogue, the following characterization of this type, at 950, is given (a few details have been changed): 1) The King's treasury is robbed by a former architect who has left a stone loose. The thief is caught in a trap. At his request, his brother (son) cuts off his head so that his identity may be concealed. 2) In order to identify the thief, the body is carried through the streets to find out if anyone weeps over it. The son has forewarned the family, but the mother insists upon the rescue of her son's body. His brother succeeds in stealing the body. 3) To catch the youth the king allows free access to his daughter. She makes each of the visitors declare his most dangerous exploit. When she learns of the theft, she is to leave a black mark on the man who sleeps with her; the youth marks all the knights and the king himself. A child will hand a knife to the guilty man; he exchanges a toy with the child and thus deceives the king (cf. also Thompson 1951:171-2). Similar tales are known from ancient Indian and Chinese literature (in the latter case they are translations of the Indian models). A special Indian version was retained in the collection of Somadeva. It conforms to Herodotus's in that: 1) The number of the thieves is two; 2) One of them is caught; 3) Guards watch the body to see if anyone laments; 4) They are overcome by trickery; 5) The king's efforts are futile; 6) Pardon (reward) is offered. There is no mention in the Sanskrit tale of a treasury, and consequently the trap and beheading of the brother do not occur. No mother appears, and there is neither the shaving of the guards nor the prostitution of the king's daughter (cf. Penzer 1952:31, Liungman 1961:244f.). The first question that arises in folkloristic scholarship concerns the origin of this tale. Some scholars believe that it could not have emerged directly from the Egyptian milieu because some of its motifs do not conform to the morals of ancient

FOLK POETRY

745

Egypt. But we cannot dismiss the fact that it first appeared in Egypt among Greek colonists. A satisfactory answer to the question whether Herodotus's version is the starting point for a whole tradition can only be given when this type has been given monograph treatment. It is highly probable that even prior to Herodotus a tale about the theft of the king's treasury was widespread in the eastern Mediterranean and that the popularity of Herodotus's chronicles of old traditions merely strengthened and revived it. This type of tale was later influenced by other literary versions, including the above-mentioned medieval version in Dolopathos, which must have stemmed from an oral tradition, and certainly not from Herodotus's writings. Animal tales and fables were popular genres even in ancient times. There cannot be the slightest doubt that animal tales were part of the folk tradition of a pre-literate period, as is evidenced by the widespread character of this genre among primitive peoples. One can therefore assume that certain literary genres (e.g. of ancient Egypt and Greece) were based on an ancient oral tradition. Comparative studies have shown that the oldest literary forms are not necessarily the most archaic ones. Certain fables are recorded in ancient times only in an abbreviated form. A proverb by Archilochos says: 'The fox knows much, the hedgehog but one thing' — a clear reflection of the fable about the fox and the hedgehog which is known mainly from the folk literature of the Balkan peoples (AT 105, cf. Bolte and Polivka II.119f. = Grimm No. 75). The fox boasts of having a bag of tricks, whereas the hedgehog admits it knows only one — how to roll into a ball and pretend to be dead. In time of danger this saves him, whereas the fox is caught. The oldest literary version (Romulus) replaces the hedgehog by a cat and the latter saves itself by climbing up a tree. The conformity between the ancient Greek proverb and the more recent Balkan versions can be explained only by the continuity of the oral tradition since the time of antiquity. It is hardly likely that the entire oral tradition is a derivative (like the older proverb) from a preserved literary version (cf. Krohn 1892:177-81). O. Hackman likewise showed (1904) that the old tale about escaping from a one-eyed giant has been preserved in a more original form in the folk tradition than in Homer's Odyssey. Recent analyses of ancient Egyptian fairy tales about two brothers and their modern parallels have led to similar conclusions. Among the more recent records we know of only one text based on the folk tradition which coincides with the ancient Egyptian version in certain features that are not otherwise recorded in folk tradition. This proves quite clearly that an isolated oral version is derived from a recent translation, especially since it occurs in an area where this type (AT 318) is not otherwise attested (cf. W. Anderson in Hessische Blätter für Volkskunde 44:123, 1953). What we have in all such instances is the more or less direct reproduction of written versions. But, is it possible to decide on the basis of formal features whether such a text can be considered folkloric? In collections of folk poetry one

746

K. HORALEK

encounters reproductions of texts which were inserted by the collector himself. Such incidents are quite frequent in nineteenth-century collections whose character was primarily determined by the appeal they may have had for their readers. Most tales are not faithful records of the oral tradition but are stories retold by authors with literary ambitions. When suitable folk texts were not available these authors drew upon older collections, and sometimes even upon foreign texts. (Such instances have been found, for example, among the Czech tales by the writer B. Nemcova, and among collections from China in European texts; cf. Asian and African Studies 2.23-37, 1966). Literary imitations in the nineteenth century were more often presented as written folklore, and included not only prose tales but also versified compositions. There is no agreement among contemporary scholars on the question of the origins and roots of folk poetry; however, the classic theory of ritual origin continues to be in wide vogue. According to this theory, folk poetry is a survival, a vestige of ancient rituals. 'For those folklorists who regard all folklore as derived from ritual, the search for origins consists in identifying or reconstructing the ritual underlying a given example of folklore. For example, one suggested ritual origin of the counting-out rhyme beginning with "eenie, meenie, miny, moe" is supposedly an ancient magic rhyme that was used in Druid times for choosing victims for human sacrifice. Similarly, the children's game of London Bridge is supposed by some to be a survivial of the actual custom in which a human victim was sacrificed to the spirit of the stream over which a new bridge had been built.2

B. Major Historical Theories and Results The scientific study of folk poetry, which was first seriously undertaken in the early nineteenth century, has always been historically oriented. This approach, however, was not evenly applied to all genres. The historical method was first utilized in the study of fairy tales in view of the similarity between ancient and modern, exotic and European fairy tales. Old literary poetry invited comparison with oral poetry and raised the question of its historical priority. The established relationship between fantastic fairy tales and myths gave rise, in the middle of the nineteenth century, to the so-called mythological theory of the origin of fairy tales (J. Grimm). According to this theory, fantastic fairy tales stemmed from Aryan myths. Comparative literary studies led to the conviction that various extant European anthologies of fairy tales and stories were derived from Oriental prototypes, and that some of them were translations of Indian or other Oriental originals. 2

Note the following selected works by adherents of the ritual approach to the study of folklore: James 1958, Propp 1946, Spence 1947, Lord Raglan 1955, Hyman 1955, and a critical study, Bascom 1957. There are many obvious connections between ritual and folk drama. For these problems, see the entry on 'ritual drama' in the Standard dictionary ... II.246-9 (Leach, ed., 1949-50).

FOLK POETRY

747

The interest of European readers in Oriental fairy tales was strengthened by the various translations from Arabic (The Arabian Nights), from Persian (The Thousand and One Days), and Turkish (Tuti nameh). A similar flow from east to west can be assumed in the case of oral tradition, and there is ample proof of the folklorization of fairy tales from translated collections. A one-sided generalization from these facts gave rise to the Indian theory regarding the origin of fantastic fairy tales (Th. Benfey). Some scholars accepted the more moderate view of the Oriental origin of fantastic fairy tales. One cause of this modified view was the discovery of fantastic fairy tales in ancient Egyptian literary writings which are much older than the anthologies of Indian fairy tales. Benfey did not claim that all the fantastic fairy tales came from India; in fact he assumed that in view of ancient European anthologies, the fairy tales about animals were of Western origin. In the second half of the nineteenth century, the comparative study of fairy tales acquired a more solid basis mainly as the results of the publication of a large amount of European folklore material enriched by material from other continents. More and more examples of migratory fairy tales were satisfactorily authenticated, though the number of unclear and controversial cases likewise increased. This led some scholars to assume that fairy tales emerged, and continued to emerge, in different areas at different times. The so-called anthropological theory, which has affinities with the mythological theory (E. B. Tylor, A. Lang) and which has affected other fields of research, views fairy tales as a product of cultural creation at a given stage of social development. It cannot be denied that cultural diffusion and migration contributed in a fundamental manner to the spread of folklore subjects and motifs, which originated at a given time and place. Most adherents of the anthropological theory admit the possibility of polygenesis for the simple types of subjects, whereas monogenesis is assumed for more complex forms. The forms that arose at a given time in a given environment spread from place to place, both through literary and oral means. It would be difficult to prove conclusively whether a particular case is the result of convergence (in the polygenetic sense) or of diffusion and migration. Even to date polygenesis cannot be said to have been proved and one may assume only various degrees of its probability. Polygenesis can be tested only where similar folklore forms appear in places that are distant from one another, and which do not exhibit any cultural contacts. Newly discovered facts may, of course, modify the investigator's original assumption. The fairy tale about the brave tailor (AT 1640) may serve as an example. J. Polivka devoted a chapter of his fairy tale studies (1904) to this type. He discovered that this fairy tale had appeared in Europe by the sixteenth century and that it is a parody of a chivalrous story and heroic song. Oral variants come from several areas (Central Europe, the Balkans, Russia, the Caucasus, Mongolia), and Polivka believed that some versions arose independently of the elements which were most widespread geographically. Later it was found that there was an even

748

K. HORALEK

greater number of oral variants and also several Indian versions which prompted Johannes Hertel to assume an Indian origin of this fairy tale. Polivka lacked a connecting link between the Caucasian, Russian, and Balkan versions, a link that was later provided by a Turkish variant. New findings fill in the blank spaces in Persia and the Middle East as well. Polivka adhered to the migratory theory but was willing to admit polygenesis where contacts were not clearly confirmed by the material. The discovery of more recent variants showed in this case the unlikelihood of polygenesis. The fairy tale about the brave tailor arose most likely in Persia or in India and came to Europe via different routes (through the Caucasus, the Balkans, and Italy). The fairy tale tradition of the entire ancient world (Europe, Asia, and North Africa) forms a unity in which the existence of independent parallels cannot be conclusively proved. Various links with the fairy tale tradition have also been discovered in America and Polynesia, and even Australia is not completely outside its area of impact. The American tradition was strongly influenced by European colonization and the importation of slaves from Africa. Ties to Asia were maintained across the Bering Straits and the Asian cultural flow can be followed along the Pacific coast down to Chile. What has been ascertained by the study of American myths also applies with great probability to certain types of fairy tales. A special problem is represented by the Central American tradition, including the Mexican and Peruvian areas. In the post-Columbian period, European and African influence was much more intensive, particularly in North America. But it would be absurd to regard the entire American fairy tale tradition as being merely a derivation from that of the Old World, and even less likely would be similar opinions in the field of American mythology. The American tradition gives rise to certain special problems. When a given legend or fairy tale is found in two places which are particularly far apart, it is sometimes difficult to decide whether we are dealing with independent traditions or with two fragments of what was once a single version. As an example we might choose the myth about getting to heaven with the help of a chain of arrows, shot one after another. This myth is widespread chiefly north of California and in Brazil (according to sixteenth century reports). P. M. A. Ehrenreich (1905a:76) claims that this is not a case of an old American tradition but an instance of more recent import, from the North to the South. The situation, however, is complicated by the fact that a similar myth is found in northern Asia and Melanesia. The Japanese have a related myth which goes back to the seventh century. Leo Frobenius, in his day, observed differences in myths that refer to the sun and the moon as a married couple. The basic difference concerns the gender of both members: in some places the sun is believed to be the male and the moon the woman, and in other places the opposite is the case. Even among the Europeans there is no agreement on this point. The Greeks, for instance, regard the sun as a man and the moon as woman, whereas the Germans reverse their gender. The

FOLK POETRY

749

Slavic peoples follow the latter, except that the sun is neuter. The Greek myth about the marriage of the sun to the moon penetrated to the Balkan Slavs who treated the moon as feminine (Serbian mesecina). This 'Greek' type spread along the whole Mediterranean, and extended via Iran to India, and through Indonesia to Melanesia and Polynesia, reaching as far as Central America and the adjacent areas of North America (the Mississippi watershed) and (northeastern) South America. Such relative compactness would argue for a genetic connection, whereas the separation of these areas by large bodies of water argues against this possibility.

C. Reconstruction as an Historical Method The historical-comparative approach to folk poetry originated in the nineteenth century together with the comparative linguistic method, when many texts from antiquity and from the Middle Ages came to the fore. These were not, however, straightforward records of folklore texts, but their stylized and adapted versions in which the folklore basis is often blurred. Historical-comparative study often has to content itself with sources of this kind. Nevertheless, it endeavors to eliminate all extrinsic, non-folk components, and to reconstruct the original sources at least in an approximate form. The comparative study of folklore sets for itself even higher goals, as when it attempts to reconstruct a picture of folk poetry even for times for which we have no written records. It then proceeds according to the principles of analogous deduction (i.e. when the situation of less developed societies suggests conclusions concerning the past of more advanced societies) or along the lines of reconstruction practiced in linguistics. The accuracy of each construction is more likely as the amount of information and the number of records is increased. Early literary evidence concerning the existence of folklore texts is for the most part rather vague; sometimes it does not even indicate whether a form in question was a myth or a tale with fantastic elements, or whether the original version was in verse or in prose. Let us take as an example fairy-tale type AT 400 (The Man on the Quest for His Lost Wife), the basis of which is usually considered to be the introductory part that contains the Swan Maiden motive. The hero wins as his wife a swan maiden by taking possession of her clothing of feathers which she has left on the bank of a lake. He keeps the clothing of feathers hidden, but after some time his wife finds them and flies away. This introductory part, however, is missing in many variants of Type AT 400, and it is questionable whether all variants may be subsumed under a single type. Doubts have been raised by V. Tille, and J. Krzyzanowski (in his catalog of Polish fairy tales), who divide the variants of Type AT 400 into more or less independent versions or subtypes (A,B,C,D). The verse dialogue between Purutavas and UrvasI in the Samhita Rigveda is regarded as the earliest example of Type AT 400. The dialogue implies that UrvasI

750

K. HORALEK

was a supernatural being who falls in love with King Puruvaras; she lives with him until she conceives and then she suddenly leaves him. The forsaken husband searches for her until he finds her with her maiden attendants at a lake. There he tries to persuade her to return to him, but she refuses. She reminds him that she has warned him against breaking the agreement that existed between them. She promises to send him the son, who will be immortal; the father himself, however, will not attain happiness until in the Other World. Mahabharata's brief version concludes with the punishment of the king because of the harm he has done to the Brahmans. Later Indian versions take the form of fairy tales; some of them contain the motif of transformation of celestial maidens into swans. In Kalidasa's drama about UrvasI (fifth century), the original plot is transformed into a story of marital jealousy. (The fairy UrvasI is the king's beloved, the king wavers between her and his legitimate wife; the two women are jealous of each other. UrvasI flees to a forbidden place and is transformed by the gods into a liana. Eventually, the king redeems her by services rendered to the gods and is allowed to live with her again.) In Somadeva's verse version (Chapter 17) the original plot is transformed into a story of the lover's faithfulness. UrvasI is separated from Purutavas by the gods because he has offended a dancer of divine origin. In the end, Visnu allows the lovers to meet again and live happily together. Somadeva's collection of fairy tales contain yet another version (Chapter 108). There is nothing to suggest that the original plot about UrvasI had the form of a fantastic fairy tale. The earliest literary versions clearly have the form of a myth; here the forsaken husband endeavors in vain to win back his wife of divine origin. Happiness is promised to him only in his posthumous existence. In Kalidasa's drama the old myth is enlarged by story motifs, which make it a fairy tale. A similar conclusion is chosen by Somadeva. In modern fairy-tale versions the focus of interest is shifted to the search for the lost wife. Here the tradition displays great variety, so that if the search is considered the basis of the narrative, it must be admitted that these versions are not of a single type. The old mythical basis gave rise to a fairy tale thanks to various additions which stem from the fairy-tale tradition. It is not necessary to assume a genetic connection between the particular transformations which lead from the old myth to the fairy tale. From Kalidasa's drama we can hardly posit the existence of a folk fairy tale similar to fairy tales of the modern tradition. Myths with the motif of the transformation of swans into maidens, moreover, need not have a single basis (prototype). Popular versions of fairy tales undoubtedly spread from one area into another. Thus we find among the modern records clearly retold versions from The Arabian Nights (cf. Tille 1925:312-28). Oriental origin can be assumed also in the case of some other versions (cf. Bolte and Polivka III.415; Liungman 1961:84; Andersson 1967:3ff.).

FOLK POETRY

751

Myth, legend, and folktale are not necessarily the only major categories of prose narrative under which all other kinds of prose narrative must be classified as subtypes. Reminiscences or anecdotes, humorous or otherwise, and jokes or jests may constitute the fourth and fifth such categories. Reminiscences or anecdotes concern human characters who are known to the narrator or his audience, but apparently they may be retold frequently enough to acquire the style of verbal art and some may be retold after the characters are no longer known at first hand. They are accepted as truth, and can be considered a sub-type of the legend, or a proto-legend . . . In contrast, jokes or jests do not call for belief on the part of the narrator or his audience, and in this resemble folktales. It may be possible to distinguish jokes from folktales and other prose narratives on formal grounds, but I am not aware that this has been d o n e . . . . In view of the importance of jokes in American folklore, they may perhaps deserve a separate category along with myths, legends, and folktales, but this may be an ethnocentric view because little has been written about them outside of literate societies. Both jokes and anecdotes obviously require more attention by folklorists than they have received, but until more is known about them, particularly in nonliterate societies, I prefer to consider them tentatively as sub-types of the folktale and the legend. In some societies the conventional opening formula which introduces a folktale gives warning to the listener that the narrative which follows is fiction, and that it does not call for belief; and this notice may be repeated in the closing formula. These signals serve as a frame to enclose folktales, and set them apart from myths and legends, from normal conversation, and from other forms of serious discourse . . . As an hypothesis for further investigation one may postulate that if a prose narrative begins with a conventional opening formula (even if its meaning is unknown), and if it is told after dark, it is a folktale rather than a myth or a legend (Bascom 1965:5-6).

D. Folk Narrative Of all the folklore genres, the folktale enjoys by far the widest extension. Similarities between European and Oriental folktales are especially striking. Close analogies between some folktales have also been found in classical literature. These facts attracted the attention of scholars even before folktales and fairy tales became the subject of scientific study. The Grimm brothers laid the foundation for the comparative study of fairy tales in their commentary to their German collection (Vol. 3, 1822). In the preface to the second volume of the collection (1815) W. Grimm wrote that the ancient German myths had been preserved in the German fairy tales. This assertion gave rise to the mythological theory of the origin of fairy tales which was developed further by later researchers (e.g. A. van Gennep, A. Wesselski, J. de Vries), whereas by contrast others (e.g. W. Wundt, F. Panzer) have ascribed a greater age to the fairy tales. The anthropological theory likewise posits the mythical origin of fairy tales, except that it allows for the formation of myths and fairy tales among various nations (in the polygenetic sense) while the Grimms saw only an Indo-European heritage in the fairy tale. Theodor Benfey presented a different view of the folktale's Indo-European origin in his commentary to the translation of the Indian collection Panchatantra

752

K. HORALEK

(Vol. 2, 1859). According to Benfey, folktales were created only by certain IndoEuropeans, moreover at a later date, long after the breakdown of Indo-European linguistic unity. Animal folktales originated, according to Benfey, mostly in the Greek sphere, whereas the fantastic fairy tales sprang up mainly in ancient India during the spread of the Buddhist religion. Fairy tales had then spread to Europe through various literary channels and by word of mouth. The transmission of folktales to Europe proceeded mostly through literary translations from Persian, Arabic, and Hebrew. Benfey also attributed great importance to migration by word of mouth along the so-called northern route, giving the main credit for intermediary to the Mongolians. Later research proved the Indian origin of some fairy tales which Benfey had not considered, and his theory, with various limitations and modifications, gained followers among outstanding connoisseurs of the fairy tale tradition (R. Köhler, E. Cosquin). J. Bedier was one of the most prominent critics of Benfey. In his book Fabliaux (1893) he claimed that the themes of Old French tales show Oriental origin only in isolated cases. Bedier assumed that folktales came into existence at various times and in various places, but doubted whether any conclusions could be drawn about their specific origin or direction of spread (the agnostic theory). Bedier's position met with little response. It had a deeper influence on the study of folklore in Russia where the migration theory had been discredited especially by V. Stasov who tried (1868) to explain the plots of the Russian byliny (folk epic) as derived from Oriental sources. The opposition to 'migrationism' was also strengthened by nationalistic tendencies. On the initiative of the Finnish scholars A. Aarne and K. Krohn research work on fairy tales and folktales began to be organized on an international scale. A. Aarne worked out a system of classification of the tales which soon gained wide international recognition. The Finnish Academy of Sciences provided publishing facilities for issuing the series Folklore-Fellows Communications (FFC). On the whole most works published in this series adhered to the so-called geographicalhistorical method, whose principles were elaborated and tested in several monographs by Krohn and Aarne. The Finnish school had a distinguished advocate in W. Anderson (1885-1962), who was on the editorial board of the FFC series where he published the monograph Kaiser und Abt (Volume 42, 1923). Anderson's Russian pupil P. Andrejev also collaborated with the Finnish group. The principles of the geographical-historical method were set forth concisely in the encyclopaedia Handwörterbuch des deutschen Märchens (11.508-22, 1934-40). Some researchers abided by these principles even before the Finnish School had come to the fore. Anderson himself cites as his predecessor the Russian scholar L. Kolmayevskii, whose comparative study of animal epos was published in 1882. J. Polivka also came close to the Finnish School in his working method. The specific aim of the 'orthodox' followers of the Finnish School was to reconstruct the primordial version of the folktale under scrutiny.

FOLK POETRY

753

Among the most outstanding opponents of the Finnish School was Albert Wesselski, author of the book Versuch einer Theorie des Märchens (1931), which provides a commentary to various collections of fairy tales (e.g. to the anthology Märchen des Mittelalter, 1945, which he himself compiled) and of various significant articles (e.g. "Der Knabenkönig und das kluge Mädchen", 1929). Wesselski sought the main source of folktales in written versions and saw the main task of the comparative researcher in establishing the oldest written version. H e examined the migration of tales by describing literary routes, considering here as well the role of traders, missionaries, and soldiers, and he adhered to the theory of oral transmission in the sense of a wave-theory. This view was shared also by C. W . von Sydow (a collaborator and a member of the editorial board of FFC) who held, however, a somewhat different opinion on the question of the origin and diffusion of folktales. Von Sydow returned to the theory of the Indo-European origin of the fantastic fairy tale (tracing it back to the period of Indo-European unity). In the Indo-Europeans he saw not only the original creators of the fantastic fairy tale (chimera), but also its main propagators, and he believed that its various regional versions (or oicotypes) were due to an original differentiation. V o n Sydow found the main support for his theory in the ancient Egyptian fairy tale of the two brothers and its modern versions. Although incomplete, and not convincing in its conclusions, we shall pause to consider this study because of its special significance for the historical study of the fairy tale. The tale of the two brothers, Anup and Bata, was discovered in 1852 on hieratic papyrus and soon aroused the attention of researchers. The text dates back to about 1200 BC (there are other, older records of fairy tales from Egypt, but none contain complete texts). E. Cosquin (1888) wrote the first comparative study of the tale of the two brothers, adducing modern parallels, mostly from eastern Europe (a semi-literary Russian version dates from the eighteenth century and exhibits some features that are not found elsewhere). Some of the Egyptian themes have parallels in the Oriental and Balkan versions, though the introductory section also shows similarity with North African versions. The Anup and Bata type was included in the international catalogue (twice, A T 318 and 590 A ) only after the publication of the 1961 edition. Comparison with modern parallels indicates that the ancient Egyptian version was founded on three differing tales. The introductory part (the story of Potiphar's wife) was based on one tale, the middle and final sections on a tale that is still alive in the European tradition, though the middle section was supplemented with the tale which figures in the international catalogue under A T Nos. 302 B and 516 B. (According to von Sydow this was a version of the oicotype, which was later listed in the international catalogue as A T Nos. 318 and 590 A ) . According to von Sydow both parts arose immediately following the breakdown of Indo-European unity in the satem group of Indo-European languages: one oicotype originated in the BaltoSlavic group, and the second in the Indo-Iranian group. This conception is contra-

754

K. HORALEK

dieted, however, by the fact that type AT 302 B (516 B) is associated mainly with a non-Indo-European environment, i.e. North Africa. Types AT 318 (390 A) and 302 B (516 B) cannot be considered as two versions (or oicotypes) of one type, for they are tales of quite a different character. In the ancient Egyptian tale of the two brothers and in its modern European variants the mainstay is formed by an account of the wife, who, incited by a foreign ruler, betrays her husband and is then punished; in the type AT 302 B the wife in a similar situation remains faithful to her husband and is carried off by force (kidnapped); the conclusion is the reunion of husband and wife. The modern European parallel to the tale of Anup and Bata might, but need not, be of Indo-European origin, as admitted even by W. Liungman (who otherwise does not accept the Indo-European theory of his compatriot). It is highly unlikely that the Slavs preserved this tale from ancient times and that they received it from Asia Minor. It is more likely that they received it via a literary route, perhaps through Byzantium, although oral transmission cannot be ruled out. The tale could have been brought to eastern Europe by pilgrims from east Mediterranean regions. But these are only conjectures. In the light of our experience with the ancient Egyptian tale of the two brothers and its modern parallels we can say that von Sydow's theory in no way contributes to the explanation of the history of the tale. This tale (as well as other tales that are preserved in ancient Egyptian versions) nevertheless remains as an important piece of evidence for old fantastic fairy tales. Other important evidence in favor of the great age of fairy tales are those which have been preserved in ancient Babylonian records (e.g. the mythical tale of Etaa in which the hero helps an eagle wounded by a snake and is then flown by the bird across the ocean). Remarkably close parallels to the Etan account are found in the European fairy tale tradition (mostly as a part of longer tales), and more distant parallels are found in other parts of the world. The striking similarity between these old literary and oral versions precludes any serious doubt as to their genetic connection. On the basis of the ancient Egyptian tale of the two brothers we can also ascertain that some old literary versions of tales were based on an oral tradition. We must not, however, therefore conclude that this tradition was uninterrupted. In some cases a fairy tale could have faded away in the oral tradition only to reemerge after a while in a literary source. A considerable number of tales penetrated the European folktale tradition from literary sources. The folktales of Oriental origin, and to a lesser extent those of classical antiquity, penetrated Europe in literary form. Oriental tales came to Europe via Spain and through the Mediterranean to the Balkans and Eastern Europe. Although many such cases have been safely ascertained, the conclusion cannot be drawn that Europe did not have its own fund of fairy tales. Some literary works preserved from the period when they could not have influenced the oral tradition point clearly to the existence of the folktale in Europe, though for the present we cannot specify the origin and extent of this tradition.

FOLK POETRY

755

Although some narratives are preserved without any fundamental changes, it would be false to conclude that for thousands of years the sense and function of the tales remained constant. The old versions of the tales usually contain more mythical material than the modern ones. In ancient times the mythical (or ritual) material might have in turn contained some elements which escape our present understanding. Of course it does not follow from this that, for example, the ancient Egyptian tale of the two brothers was a myth and not a fairy tale (as is held by e.g. J. de Vries and E. Brunner-Traut). We must be content with the assumption that formerly the differences between myths, legends, and tales were not so sharp. In interpreting old as well as modern texts the basic guide must be the mutual connection of the elements and their arrangement as a whole. The functional criterion should be given priority over the genetic one. The magic environment imparts to the fairy tales its narrative function; but a fairy tale need not inform us about ancient superstitions or customs. If the tasks of heroes find their parallel in ancient wedding ceremonies or customs, nothing serious can be deduced for the interpretation of the tale in which the performance of these tasks requires the interference of magic powers. The development of the European fairy tale is connected with an increasing simplification of its style (e.g. decrease of decorative formulae and of some formal features which are considered its constituent characteristics by M. Liithi). An increasing process of 'rationalization' brings the fantastic narrative at the same time closer to the realistic story. The formulaic style, however, is a product of specific historical conditions and can not be considered to be primitive (Rohrich 1956: 123f„ 137f.). It is difficult to decide which formal signs qualify the narrative as a poetic composition. The same folklore material which would be treated as a fairy tale in a peripheral cultural area (e.g. in the polar region), may be taken in the European milieu as no more than a memorate (for the term 'memorate' cf. von Sydow 1948: 73). The subject of a narrative may often be a simple tale about appeasing hunger, about conflicts motivated by vital needs, and so on. If the poetic character were to be decided by such concepts as departure-return, a task and the fulfillment of the task, danger and the removal of danger, a struggle and victory, one might easily find a structural pattern even in such a simple narrative as going to the forest for wood. The same holds for the criterion 'a move from disequilibrium to equilibrium' (Dundes 1965:208). The stylistic-formal criteria of the narrative (the frequency of epithets, metaphors, formulas, and so forth) give rise to various problems. Basically one always expects to find such stylistic-formal criteria, but in a historical-comparative investigation if we consider the spatial and temporal aspects as well as the exotic nature of the material, the formal aspect necessarily recedes into the background, especially since we often deal with translations which blur the stylistic quality of the original. But the poetics and stylistics of folklore present problems which are peculiar to

756

K. HORALEK

folklore itself, such as the quality of the records, the role of mimicry and gesticulation, of voice modulation, and so on. A study of these phenomena is now facilitated by film recordings, but there are too few of them to allow a comparative study. Not even the most perfect recordings can change the fact that they do not present living folklore, but a simplified reproduction thereof (notes, tape recordings, and so on). That is why, finally, traditional folkloristics remains closer to literary research than to ethnography. Since World War II increasing attention has been paid to legends and sagas. Among other reasons this is due to the fact that the irrational components of human culture are now expounded and evaluated differently. In the hypertechnical phase of history, man appears to have more understanding for his own foibles than in the period of his belief in the absolute power of reason. This important component of recent folktale research is rightly pointed out by G. Heilfurth in the introduction to his broadly conceived work Bergbau und Bergmann in der deutschsprachigen Sageniiberlieferung Mitteleuropas (1967:19). The new approach to the irrational elements of human culture enables us to see how myths survive in a modified form. Mythical elements in present-day French life was made the subject of a series of articles, collected by R. Barthes in book form as Mythologies (1957). His approach to the problems of modern myth is more descriptive than interpretative. More important than the 'everyday' myths of modern culture are those created by modern art and philosophy. By overestimating the irrational elements of modern culture some philosophical trends have created the myth of the instinctual basis of human existence, upon which all moral conventions have a destructive effect. Some of the most important contributions to the structural theory of the myths have been made by Claude Lévi-Strauss. According to his theory, there are two sources of human knowledge, conceptual thought and imaginative intuition. Mythical thought, which is based on sensible intuition, is, as Lévi-Strauss supposes, no less scientific or valid than knowledge gained by rational thought. Like Cassirer, Malinowski, and others, Lévi-Strauss 'does not favor an evolutionary approach to the study of human mentality which stresses the discontinuity and differences between primitive and modern thought. Archaic mentality is but one mode of thought, an attitude and tendency of the human spirit which manifests itself in myth, art, and religion. It is to be found in prehistoric as well as historic societies; all that changes are the proportions of the archaic and intellectual modes of thought' (D. Bidney in JAF 77.182f., 1964). See the principal works of LéviStrauss (1958:227f.; 1962; 1964-68); and compare Simonis 1968.

E. Folk Song Both folk prose and folk songs of all peoples are historically and territorially dif-

FOLK POETRY

757

ferentiated and include, in addition to some relatively stable constituents, some variable and unstable ones. Some song genres are widespread and known all over the world, others are limited to a few areas, appearing under particular historical conditions and surviving only so long as these conditions prevail. However, not much reliable information is available concerning the song tradition of many regions and peoples. But it may be assumed that the song tradition is as universal as the prosaic one. The differentiation of folk songs applies to both form and content. Formal differentiation is also strengthened by dependence on instrumental accompaniment. There is a great variety of musical instruments accompanying song even in regions which are remarkably homogeneous in culture. The accompaniment is sometimes optional, sometimes obligatory. The instrumental accompaniment is generally tied to traditional genres (ritual songs, heroic songs, wandering singers's songs, etc.). Among religious songs, the situation is complicated by the prohibition of all but oral music as is the case in the Orthodox Church (Greek and Slavic), a circumstance that influenced the development of folk music and song in general. Instrumental accompaniment has consequences for the existence of the so-called recitative, which is a transitional form between scanned recitation and proper song stylization. Less known are some verse formations recited with or without scanning. In Central Europe these compositions which form part of wedding ceremonies are recited by heart usually in accordance with some written or printed models. In magic formulas, a song or a recitative form often appears next to simple scanning. Nursery rhymes, known in most European folklore, often have special rhythmical features. They are characterized by a distinct modulation of accent, which leads to the weakening of the syllabic principle in regions where this principle is dominant (e.g. in the West Slavic tradition). Nursery rhymes are still a living component of all European tradition, but they have not been studied systematically on a comparative basis. Nor has epic recitative, which is still alive in Slavic tradition, been thoroughly investigated. The origin of song (and of any poetic language expression) is usually connected with the rhythm of work; this is however only one source of the poetization of human speech. Other sources may be found in magic and ritual expressions. Mnemonics, i.e. fixing ideas in one's memory, also influences the selection of the verse and song stylization. The common nature of primitive poetry is to some extent determined by forms that satisfy the elementary needs of life, and which consequently, reveal an interest in nature, in atmospheric phenemona, hunting, and in the basic problems of human existence (cf. C. M. Bowra 1962). In Cassell's Encyclopaedia of Literature (1953:I.225f.) we find the following explanation of the origin of folk song: The connection between folk songs and the dance and the precise origin of working songs are still matters for discussion, but there seems little reason to doubt that many folk songs

758

K. HORALEK

may be traced back to such sources — weaving songs, plough songs, the medieval cantilenae molares, even the better known cradle songs, all belong to this group. Any activity, in fact, in which rhythmic movement was called for, might readily and easily give rise to melody and song, as it still does amongst the labouring classes in many parts of the world. Songs would also arise on ritual occasions such as weddings and the scattering of the seed in the fields, the idea of fertility being prominent on these occasions, while drinking songs and satirical songs of a political nature need no further explanation. The literary treatment of such songs has already been referred to; it should also be noted that their forms and rhythms, as distinct from their content, have been utilized to good purpose on more than one occasion when the politer literature of a country was becoming sterile; the example of Sir Thomas Wyatt in England comes most readily to mind, and an even stronger and more continuous influence of this nature is apparent in Spanish literature.

Many similar cases are found in Slavic and other literatures. Traditional poetry of culturally less advanced peoples is usually permeated by a mythical conception of life. The main function of myth here is explicative; in most cases it is an explanation of natural phenomena in relation to human life. Na'ive rationality in myths goes hand in hand with unbounded imagination and emotion. All this determines — to some extent — the nature of primitive poetry, be it a song or a prose story. The distinction between prose and poetry is often variable: some peoples prefer song in the mythical and epic traditions, others prefer prose narrative. Many pre-Christian mythological survivals have been supplanted in the European tradition by Christian motifs. The poetry of more advanced societies is characterized by functional and generic differentiation. Some poetic forms have gradually become emancipated from family, calendar, and church rites, though the development of all forms is conditioned by the internal relations of their particular components. But all folk poetry reveals connections with the customary organization of social life. Different kinds of song and prose narratives are reserved for particular situations and occasions and many songs enter as components of ritual and ceremony (weddings, harvests, funerals, etc.), or function as complements thereof. The link between ceremony and some forms of work (mainly field work) is typical of conservative societies. So-called work songs often have a more prominent ritual rather than technological character. Song and narration also function as entertaining accompaniment to collective song-regulated work, as for instance, the work of Russian people trailing ships up the river by means of ropes (burlaki). The origin and development of political poetry is conditioned by the existence of class structures and political organizations. The oldest form of political poetry is the war song, which is found in primitive societies and can sometimes be transformed into heroic poetry of a professional type. The political song has two extreme forms: one expresses the dislike and resistance of the suppressed classes and groups, the other celebrates and honors political leaders. Military songs often express patriotic feelings, and sometimes people's resistance to wars or military

FOLK POETRY

759

service as well. Political motifs may appear as a subsidiary component in family, amatory, and religious songs. Folk poetry renders political history only in a fragmentary or distorted shape. Thus the Czechs and Poles have no folk songs related to the loss of the political independence of their countries, while anti-German and anti-Austrian political songs emerged during the Czech National Revival, although most of them enter the oral tradition through popular song-books and leaflets. The character of this political poetry was largely circumscribed by press censorship. War is one of the central historical topics of folk poetry. Songs or traces of songs on the war with the Turks are still extant in Slavic folk poetry. Songs about the combat against the Turks form various cycles and chronological strata among the Balkan Slavs. Several lyrical-epic songs belong to the epic cycle about the battle on Kosovo pole (Field of Merle) in Serbian folk poetry (e.g. Smrt majke Jugovica [Death of mother J.]). The wars with the Turks and Tatars are a dominant subject of historical Ukrainian songs. Turkish themes appear among the Western Slavs in ballads, and also in lyrics. Many European countries have songs about a lover or husband going to or returning from war; or the events at home during his absence (the wife might have married someone else, or she might have had children by a new husband, or the beloved wife is no longer alive, etc.). In Central Europe there are popular songs about the girl who wants to join the boy in his military service. A specialty of the East Slavic tradition are the laments of mothers (or sisters) over a son who must leave home for a long military service. Laments for the dead have a much larger area of diffusion in Europe and form a component of the folk tradition among the Eastern and Southern Slavs. Among the Eastern Slavs, laments for the bride are also sung, thus forming a constituent part of the wedding ceremony. Slavic laments also have their formal peculiarities. The East Slavic region (mainly in the North) offers records of syllabic metres, though most songs in these parts are asyllabic. Laments with a variable number of syllables are known from South Slavic (mainly Croatian), territory, while nearly all other songs observe the syllabic principle. The love poetry of West European and Central European nations is characterized by its connection with medieval lyrical literary poetry (of minstrels, spielmanns, and vagrants). This connection also contributed to the differences within Slavic folk poetry. In the West Slavic and Adriatic areas, differences were promoted by contacts with the neighboring Germans and Italians. A characteristic component of the Roman and Central European tradition are songs about a lover's night visit to his beloved and of his early morning leave-taking (analogous to the medieval albas). An interesting variant of songs with this motif are jocose compositions containing a dialogue between daughter and mother, the latter having heard voices in the daughter's room and wanting to know to whom the girl was talking. To the same cycle belong songs in which the girl advises her lover how to reach her at night (he must give something to the dog so that it does not bark,

760

K. HORALEK

open the gate carefully, not rattle his spurs, etc.). Another group is formed by songs containing a dialogue between the girl and the boy, who either consoles or mocks her. In the West Slavic tradition one finds ballads in which the lover is caught in the girl's room and killed by her brothers. More widely spread are songs in dialogue form between a girl and her lover, when the girl refuses to admit the lover, or takes flight and is transformed into a bird, a fish, etc. But the persistent lover responds with similar transformations in order to catch up with the girl. A South Slavic specialty are songs in which a girl who has several suitors chooses the winner of a certain competition (e.g. swimming across the river). This type is known in the Danube region among the Southern Slavs, Slovaks, and Moravian Czechs. It was recorded in the second half of the sixteenth century by the Czech cleric and man of letters, Jan Blahoslav. The motif of the race and competition occurs in many other types of songs among the Southern Slavs. It is partly a revocation of the ancient wedding rite, as exemplified by Homer's Odyssey, and well known also among the Turks. The structure of lyrical songs is generally simple; a two-part construction is very frequent. It is not, however, possible to claim that the simple form is more archaic than the complex one. Even songs of one stanza contain two parts. The second part is often loosely attached to the first as a parallel statement without any apparent connection, though occasionally with a contrasting or graded character. Even dialogue songs may contain only two parts; i.e. a question and answer, or a statement and its negation. The contrasting relation can be carried out according to the scheme A 1 + A 2 ( . . . + A n) + B, with B constituting the central message or point. (This principle of composition is applied, for example, in the song type 'choice of lover over relatives' for a young man in difficult family situation.) In areas where the stanzaic composition is uncommon the 'variation' presents an intermediary stage to the stanza form. So-called cumulative and chain songs are diffused all over Europe. A cumulative song is a kind of variation song in which each successive part enlarges the preceding part by one item (the pattern is A + Ai + A 2 + . . . , where Ai = A + a, A2 = A + a + b, and so on). Slavic songs of a servant's employment can serve as an example. For his service the servant is rewarded by various domestic animals and the description of their activity is presented in a cumulative form. Jewish Passover songs, which were imitated and modified in many ways within the European tradition, are another example of this type. Of a simpler form are choir songs in which the particular parts are attached in a certain order, with numerals functioning as the basis of their arrangement. A cognate form is the enumerative song, in which a group of things or people is successively presented from a certain point of view (e.g. a song in which a girl successively rejects the suitors recommended to her by her mother.) Cumulative and chain songs, which resemble the Jewish Passover compositions (Chad gadju, Echad mi iodea), were very popular in Europe. They were mostly recited or scanned and could also be incorporated into prose stories.

FOLK POETRY

761

Lyrical songs are often composed freely on the basis of uncertain and vague thematic associations, this being the origin of so-called song conglomerations. In epic poetry the loose attachment of episodes can be of a retarding character. This device is commonly used in heroic epic poetry and in prosaic stories, but not in folk ballads. The development of the stanzaic song in the European tradition was also promoted by dances with complicated rhythmic figures. The popular songs of leaflet prints and Church hymns likewise contributed to the stanzaic song composition. The stanzaic forms differ remarkably among various European nations. Even among the Western Slavs we discern considerable variation; for example, the stanza forms of the Czech songs are far more complex and variable than those of the Poles and Slovaks. Formal differences in the folk poetry of European peoples have arisen in connection with the divergence of their cultural development and most of all, with their linguistic differentiation, which was accelerated in the historical period. In this respect the situation in Slavic song folklore is instructive. West Slavic poetry primarily uses syllabic metre, rhyme, and stanzaic division (only in Lusatian folk songs is rhyme not compulsory); South Slavic songs employ syllabic verse without rhyme and stanzaic division (though the Slovenes are closer to the Western Slavs); East Slavic folk songs are non-syllabic, without rhyme and stanzaic division (the Byelorussian and Ukrainian songs, mainly in the west, come closer to the Western Slavs). For the Common Slavic period (which lasted approximately until the tenth century), we may assume (with N. S. Trubetzkoy, R. Jakobson) the existence of verse forms which were similar to those that now prevail in the South Slavic folk song, the non-syllabic (unmeasured) verses being another functional variant. ProtoSlavic syllabic verse probably arose on the basis of a similar type of Indo-European verse, which A. Meillet endeavoured to reconstruct. As a result of Central European cultural convergence, rhymed and stanzaic forms penetrated West Slavic folk poetry. Rhymed folk songs might have arisen as the direct imitation of foreign models (mostly German and Latin). Modern times have witnessed the penetration of these forms also among the Southern Slavs and Russians. Particularly widespread among the latter are the rhymed four-verse songs (the so-called chastushki) whose analogue can be found in the German Schnaderhiipfel and among several non-European peoples (e.g. the Iranians and the Turks). The adoption of rhyme and stanzaic division had certain consequences for the genre structure of folk poetry. The new type made improvisation more difficult and complicated the transference of longer narrative songs common among the Southern and Eastern Slavs; heroic epic poems are not found among the Western Slavs, even at an older period when one would expect to find this type of verse structure. Narrative song poetry is represented among the Western Slavs only by ballads of the West European type. The ballads are predominantly verses of eight

762

K. HORALEK

and six syllables. West European ballads were adopted also in East Slavic areas where they formed a younger layer upon the older, narrative songs (the so-called byliny). A South Slavic analogue to the West European ballad is the Balkan ballad whose roots must be sought in Greek folk poetry. The basic verse form of South Slavic heroic poetry is the ten-syllable verse with an asymmetrical division 4 + 6. In Bulgarian folk poetry this form gave way to eight-syllable verses with a symmetrical division (4 + 4 or 3 + 2 +3). South Slavic verse is characterized by a regular number of accents in the line (usually dactylic or trochaic). Historical songs and Ukrainian dumy (historical songs with elements of the heroic) have a looser construction. The South and East Slavic epics allow us to distinguish several phases of development which are defined more by contextual than by formal characteristics. The oldest groups are distinguished mainly by the presence of mythical elements. The beginnings of South and East Slavic heroic poetry are probably linked with the institution of the professional singer (shpilman, 'bard', the Russian skomorokh). Even in a folk milieu we must assume the existence of schooled and socially privileged people who carried the tradition of heroic poetry. We know of outstanding singers of epic songs in the East Slavic and South Slavic environments who can be compared to the old shpilman. The bearers of the epic tradition, whether they were professional or occasional singers were highly skilled: they had to master the technique of improvisation, had to know how to use traditional formulae and how to change and develop them. That is why there is good reason to compare Homer with the living epic Slavic tradition. The epic song tradition requires far greater creativity on the part of its carriers than the lyric tradition. But even here it is important to take into account the differences in time, an aspect which is underestimated by C. M. Bowra who writes (1962:3): This (heroic) poetry may be divided into two classes, ancient and modern. To the first belong those poems which have by some whim of chance survived from the past. Such are the Greek Iliad and Odyssey, the Asiatic Gilgamesh, preserved fragmentarily in Old Babylonian, Hittite, Assyrian, and New Babylonian, the remains of the Cananite (Ugaritic) Aqhat and Keret, the Old German Hildebrand, the Anglo-Saxon Beowulf, Maldon, Brunanburh and fragments of Finnsburh and Waldhere, the Norse poems of the Elder Edda and other pieces, some French epics of which the most remarkable is the Song of Roland, and the Spanish Poema del Cid, and other poems. The last hundred and fifty years have added a large second class of modern heroic poems, taken down from living bards. In Europe the art is still flourishing or was until recently; in Russia, especially in remote regions like Lake Onega and the White Sea; in Yugoslavia, both among Christians and Mohammedans; in Bulgaria; in the Ukraine; in Greece; in Esthonia; in Albania. In Asia, it is to be found in the Caucasus among the Armenians and the Ossetes; in the Caspian basin among the Kalmycks; among some Turkic peoples, notably the Uzbeks of what was once Bactria, and the Kara-Kirghiz of the Tien-Shan mountains; among the Yukuts of the river Lena in northern Siberia; the Achines of western Sumatra; the Ainus of the northern Japanese island of Hokkaido, and some tribes of the Arabian peninsula. In Africa it seems to be much less common, but there are traces of it in the Sudan. This

FOLK POETRY

763

list is by no means complete and could easily be increased. There are, n o doubt, also regions in which the art exists but has not been recorded by European scholars. There are as well other regions where it once existed but passed out of currency before the impact of new ideas and ways of life. Nevertheless, the present evidence shows that it is widely spread and that, wherever it occurs, it follows certain easily observed rules. It is therefore a fit subject for study, though any such study must take as much notice of variations as of underlying principles.

Bowra's fundamental dichotomy between the 'ancients' and 'moderns' could be complemented by various additions and corrections but the main fault in his developmental scheme is that it is based on heterogeneous material. The older phase in the development of epics is represented by literary verse; some of it was probably based on an epic tradition which can be assumed to have been similar to modern narrative oral poetry. The ancient literary epic (including the German Nibelung and Gudrun epics and the ancient Indian epos) has its analogy in Lonnrot's Finnish Kalevala, whose folkloric base — according to Bowra's scheme — would belong to the 'modern' folklore epic. But the source of old epics were, of course, not only short epic songs, but also prose legends. The source of the Anglo-Saxon Beowulf was mainly (as Fr. Panzer has shown) the fairy tale type AT 301 (The Stolen Princesses). Even the layman who is familiar with several larger collections of fairy tales could identify the folklore material in Homer's Odyssey. Not every lengthier epic must have a shorter epic form — either song or prose — as its basis. And furthermore one must allow for the basing of short epics upon 'large' epics. A. H. Krappe (1930:474) believes that such a development was actually widespread in European folklore.3 Among the folk ballads that did not result from the disintegration of large epics, Krappe includes the South Slavic narrative songs. However, he is inclined to derive most Eastern Slavic bylinas from the Greek epos about Digenis and from the extinct epos about King Solomon, though so far no closer plot analogy has been discovered between the Russian bylinas and Digenis Akrites. And as for the Solomon epic (which is known in West European literature for its motif of the unfaithful wife), there is not the slightest evidence that such an epic ever existed. And although the bylina about Vassili Okulievich does contain the motif of Solomon's unfaithful wife, this motif is more nearly analogous to the various transformations of fairy tales into bylinas which we encounter. The modern research on heroic epic poetry had devoted much attention to the problem of the relation between myths, fairy tales, and ballads. The origin and development of heroic epic poetry have been discussed most systematically by E. M. Meletinskii, who distinguishes two fundamental forms of heroic poetry: archaic and classical (1963). The development of lyrical folk poetry in Europe was 3

A considerable portion of existing ballads are unthinkable without the great epic as predecessor. Thus many of the Danish ballads are derived from the disintegrating cycle of the Nibelungs which migrated from Germany to Denmark. The Spanish Cid ballads of the Infantes de Lara presuppose the poem de Mio Cid and the several cantares de los Infantes de Lara, and the modern Greek ballads about Digenis Akrites owe their origin to the Great Byzantine epic.

764

K. HORALEK

promoted by the medieval lyric of vagrants and troubadours and by vocal church music which continued to flourish during the Reformation and the Baroque period when folk and church poetry mutually influenced each other. The different forms of Latin and Greek church music contributed to the differentiation of the European folk lyric. The influence of Western music among the Slavs accounts for the spread of rhyme and stanzaic form in lyrical songs. Work songs are considered to be the oldest type of folk song. According to K. Biichner, the primordial forms of singing originated directly in the work rhythm. Later research sought to connect the origin of songs with cult and magic functions, as well as with the themes of hunting (cf. Bowra 1962). So far there is no adequate compendium of European folk songs, nor any complete survey of their main problems and history. Good surveys of national folksong traditions are those by Suppan (1966), Danckert (1966), Wiora (1953), Sydow (1962), Wilgus (1959), and Dean-Smith (1954).« F. Proverbs and Riddles Proverbs and riddles are not confined to folk poetry alone. Many proverbs belong to learned literature as well as to folk poetry, or are part of both. Proverbs belonging to the folk tradition often penetrate into literature, which then just as often return to the oral tradition of the same or of a different ethnic milieu. A large number of proverbs of Greek and Roman origin have since antiquity formed a part of the European oral tradition, which then in the Christian era gave rise to certain biblical proverbs and sayings thereby enriching the European tradition. The simplicity of form links the proverb to the riddle. Most proverbs consist of one sentence, usually containing an imperative (for instance festina lente; 'measure twice, cut once', an indicative 'birds of a feather flock together', 'hawks will not pick hawks' eyes out'; 'no learning ever fell from heaven', or an infinitive (cf. the Czech proverb mluviti stfibro, mlceti zlato, 'To speak is silver but to be silent is gold'). The simple imperative proverb consists often of two parts (ora et labora, "live and let live') and this two-fold construction is generally widespread (e.g. 'all's well that ends well'; 'the nearer the fire, the hotter the flame'; ubi bene, ibi patria; tant va la cruche à l'eau qu'à la fin elle se casse). Descriptive proverbs that express an event have allegorical connotations (e.g. 'pride goes before a fall', 'a good beginning makes a good ending'). The expansion of the 'event' imparts to the proverb the character of an anecdote or tale. Such proverbs are common in the Oriental tradition. The modern Greek and Balkan proverbs stand, according to K. Krumbacher, on the threshold between the East and the West: 4

Slavic folk-songs and folk music are treated in Russkoe narodnoe poètiieskoe tvorcestvo I—II (Bogatyrev 1943, 1956); "Lidovâ a zlidovëlâ pisen" (Ceskoslovenskâ vlastivëda III, Praha 1968), and "Muzyka ludowa polska" (in the encyclopaedia Slownik folkloru polskiego (Krzyzanowski 1965:240-64)).

FOLK POETRY

765

Gemeinsam sind den Sprichwörtern der Mittel- und Neugriechen und der mit ihnen zu einer Kultureinheit zusammengeschlossenen Nachbarvölker einige allgemeine Eigenschaften, der Reichtum an originellen, dem Westeuropäer fremdartigen Bilder und namentlich die Vorliebe fur die anekdotenhafte, epilogische, konkret erzählede, fragende oder befehlende Form der Einkleidung, die Vorliebe für ein Schema, in welchem der persönliche und patikulare Fall noch nicht zur abstrakten, allgemeinen Regel verdichtet ist. Orientalisch ist, um ein Beispiel zu gebrauchen, die Form: Einem schenkte man einen Esel und er schaute ihm auf die Zähne', occidentalisch die Form: 'Einem Geschenkten Gaul schaut man nicht ins Maul'. Durch diese Eigentümlichkeit scheidet sich das byzantinisch-neugriechisch-südslavisch-orientalische Sprichwort prinzipiell von dem abendländischen, und man kann dadurch in der ganzen Sprichwörterweisheit eine griechischorientalische und eine abendländische Gruppe unterscheiden (1897:906-7). Krumbacher ends this interesting characterization with an appeal to investigate the proverbs of other European peoples, particularly the Slavic, on the basis of this cultural dichotomy; so far this appeal has received little response. (Only the Czech scholar V. Flajshans used Krumbacher's dichotomy to characterize Czech proverbs as falling within the Western orbit. Other West Slavic proverbs would naturally belong to the same sphere.) Proverbs of the narrative type can be found also in various Western European countries and need not be of Oriental provenience. Narrative proverbs of the Byzantine-Oriental type are found more frequently among the Eastern Slavs. Here are some examples of the East Slavic (Russian) narrative type: Privel losad' kovat', kogda kuznia sgorela. ('They brought the horse to be shod when the smithy had burned down'.) Zloj s lukavym vodilis', da oba v jamu svalilis'. ('The rascal and the rogue walked together and both fell into a pit'.) Oí vola usel, na medvedia napal. ('He fled from the ox and met a bear'.) Szalilsia volk nad jagnenkom, pokinul kosti da kozu. ('The wolf took pity on the lamb and left the skin and bones'.) Kto vstal ran'se, usel podal' e. ('The early bird catches the worm'.) A special type of narrative is the so-called Wellerism (from Sam Weiler, a character in Charles Dickens' novel The Pickwick Papers) known in German as Sagworte. These are proverbial citations with a narrative, explicative addition (for instance, "All's well that ends well', said the monkey when the lawnmower ran over his tail"). In such a context the proverb carries an ironic, mocking character. 'Wellerisms' are found in the literature of antiquity but in recent times they spread throughout Europe. Wellerisms are easily transformed into anecdotes; and they might have originated from anecdotes or other narrative forms (cf. Kuusi 1960, and Neumann 1968). It is necessary to distinguish the proverb and the so-called saying from clichés such as 'through and through', 'slowly but surely', 'as black as coal', 'for love or money', which do not have a clearly defined form. Learned literature and the speech of educated people make frequent use of foreign expressions or maxims

766

K. HORALEK

(mainly Latin) which constitute quotations from literary sources (e.g. gaudeamus igitur, voletis volena), legal and religious formulas (e.g. furor teutonicus, fiat lux, homo novus, requiscae in pace, ferro eligni). But the borderline between the proverb and cliché is fluid and allows for many transitional forms. Proverbs often appear as synonymous expressions which render the same meaning in a completely different form (e.g. the German Verkaufe die Aale nicht, eh du sie gefangen hast — Ehe man den Bären hat, darf man das Fell nicht vertrinken), or with minor variations (the Latin, Manum manus lavat, and the German, Eine Hand dient der anderen). An instructive explanation of this phenomenon by A. Taylor may be found in the Standard dictionary . . . 1.903 (Leach 1949-50). Like all folklore materials, a proverb has many traditional variations, all of which are of equal authority. The parallels to the English 'A bird in the hand is worth two in the bush' differ in the number of the birds mentioned as we see in the Gaelic 'A bird in the hand is worth a dozen on the wing' or the Spanish 'A bird in the hand is better than a hundred (or thousand) flying'. The speaker may elaborate the proverb by naming particular birds: in the Persian 'A sparrow in the hand is better than a hawk in the air', or the German 'A sparrow in the hand is better than a pigeon on the roof'. When a proverb employs completely different means of comparison to express this idea, we cannot easily discover whether we have variations that have arisen in the course of oral transmission or whether we have two proverbs of entirely different origins. Examples are the Turkish 'The egg of today is better than the goose of tomorrow', the Latin 'One hour today is worth two tomorrow', and the Serbian 'Oat bread today is better than cake tomorrow'. Proverbs of different national traditions are decidedly different from one another in form, including the use of rhyme. In the Czech tradition, for instance, rhymed proverbs are decidedly more frequent than in the East and South Slavic traditions and F. L. Celakovsky (in his collection of Slavic proverbs, 1852) often translates the unrhymed proverbs of other Slavs into a rhymed form. Even the proverbs of antiquity are often rendered in rhymed form in the modern European languages (for the Latin festina lente the German has eile mit Weile). Russian proverbs are noted for their pithiness and frequent use of elipsis. A specific feature of the Russian proverb is its frequent reference to historical events. Folk proverbs enter into literature as citations, or as 'themes' which are enlarged upon or 'illustrated' by a narrative, poem, or drama. Folk proverbs permeated medieval literary works, the ballads of François Villon, the Gargantua of François Rabelais, and the works of Shakespeare; A. de Musset (and others) developed from this base a special dramatic form known as proverbe. Proverbs are employed in a different way in folktales and in folk songs. Some stories are nothing but developments of illustrations of a single proverb (e.g. the modern Greek tales illustrating the proverb 'What is written in your book of fate cannot be unwritten'). Slavic folk poetry makes much use of proverbial formulas

FOLK POETRY

767

expressing the impossible: A girl promises her suitor to become his wife 'when a stone floats on water' or 'when a dry willow blossoms'. The basic theoretical work on proverbs is Taylor (1931). Among the Slavic proverbs the most carefully studied are the Polish ones: the four-volume comparative survey by J. Krzyzanowski, et al. on Polish proverbs are still in progress (the first volume appeared in 1970). The comparative study of Czech proverbs was initiated by V. Flajshans. A Czech anthology of proverbs dates back to the fourteenth century, and the first Czech anthology was first printed in the sixteenth century. Several collections of Polish proverbs appeared in the seventeenth century, the most important among them compiled by G. Knapski in 1632. The oldest European collections of proverbs are mostly translations. Many local and period features are contained in weather proverbs which among Christians are often related to saints' days (e.g. the Czech Saint Martin comes riding on a white horse; Saint Lucy drinks up the night). Many weather proverbs use personifications, e.g. the English 'March comes in like a lion and goes out like a lamb'; or the Czech 'White February makes the fields stronger'. Although proverbs unquestionably belong among the oldest folklore genres, they are not equally common among various peoples. The oldest anthologies come from Sumer and Egypt, but these were for the most part literary forms that might have drawn on an oral heritage. The German paroimiologist C. E . Moll, believed that even recent European proverbs were of bookish origin (Proverbium 6.114, 1966) and that Asia Minor (perhaps Sumer) was the cradle of the entire European (if not world) tradition of proverbs. But Moll was unable to say anything definite about how these Sumerian proverbs made their way to Europe. He was correct in pointing out, however, that the Bible did not function as an intermediary, since there are only rare similarities between the Sumerian and European proverbs. The parallels between the Bible and classical literature (Gordon 1959) indicate, however, that the former played a greater role in the European tradition than Moll would have us believe. That the proverb of antiquity exerted a deep influence on the modern European proverb, beginning with the Middle Ages, is demonstrated by H. Walther's monumental edition of medieval proverbs (1963-1966); cf. Proverbium 8.185-91, 1967. The later national literatures drew especially upon Latin literary sources. But the extent to which the parallel oral tradition flourished, no doubt remaining in contact with the literary tradition, is a matter of conjecture. Riddles are close to proverbs both in form (lapidary pithiness) and construction. However riddles more often have a narrative form and can be derived from stories or anecdotes. Among the 'simple forms' (einfache Formen) riddles, according to A. Jolles, are closest to myths: 'A myth provides an answer to a question it contains, whereas a riddle asks the question to which it suggests the answer'. But the connection between riddle and myth is debatable, for certain riddles are closer to etiological stories and others to animal tales. The play of poetic elements on

768

K. HORALEK

meaning brings the riddle closest to metaphorical expression. Even children's riddles are usually poetic. Riddles may figure in myths in different ways (e.g. the riddle about the three periods of human life in the old Greek myth about Oedipus). Most riddles are witty and poetic rather than explicative; old riddles even contain elements of mythical ideology. In folk epics, riddles are most often used to create a plot. In fairy tales, as in the myth about Oedipus, the hero is threatened by death if he does not solve the riddle; and if he succeeds, he wins the princess as his bride. A theme common to old sources is a contest in solving riddles. 'Enigma contests at Greek and Roman feasts were later brought to high stages of development with a master of riddles presiding and awarding the coveted laurel wreath to the winners and condemning the losers to drink their wine mixed with salt water. In the second century A.D. the winners were getting rare books and sums of money, and in the third or fourth century (the date is very uncertain), the Father of Enigmatology, Symphosius, put into three-line stanzas of pretty good Latin poetry the famous Hundred Riddles, evidently already current for some time as plain folk riddles' (Ch. Fr. Potter in Standard dictionary . . . 1.942; Leach 1949-50). The first Slavic documents about proverbs and riddles appear among the Slavs at the beginning of their literary history (in the Russian Primary Chronicle and the Igor tale and so forth). In the Czech literature of the fourteenth century, the first collection of riddles was part of a Latin work with Czech glosses (cf. Clareti Enigmata 1957, which lists additional bibliography on p. 1: mainly the work of V. Flajshans mentioned above). For Polish proverbs and riddles cf. Slownik folkloru polskiego (Krzyzanowski 1965:334^1, 452-7). For Russian, see Novikova and Kokorev 1969:119-54. Among the smaller folklore genres, one can also list magic formulae and incantations, apart from proverbs and riddles. But this may not always be the case, for certain formulae may be longer than short stories or songs, and some magic formulae are part of rituals (e.g. of the marriage ritual). The structure of magic formulas is based on syntactic parallelism, euphonic figures, and contrasting semantics. The magic formula is one of the last remaining archaic genres. The largest collection of Slavic magic formulae comes from Russia (cf. Mansikka 1909).

G. Folk Drama Specialized and regional interest in folk drama exists in contemporary research on folk poetry, but more general comparative works are needed in this field. Folk drama is mentioned, yet inadequately discussed in the existing histories of dramatic literature and of the theater. And in general works on folk literature, the sections concerning drama are disproportionately brief. The most systematic work on European folk drama is Das deutsche Volksschau-

FOLK POETRY

769

spiel by the Viennese scholar Leopold Schmidt (1962). Schmidt also prepared an anthology of Le Théâtre populaire européen (1965), in which the (Western) European representation was limited to a discussion of the Romance and Germanic peoples; two Osman Turkish folk plays are included, and one Greek (an imitation of the Turkish shadow-theatre). N o systematic treatment of folk drama and theatre of the Slavs has appeared, but the material has been anthologized in scholarly editions. The chapter on Slavic folk drama in Kultura ludowa Slowian by Kazimierz Moszynski (1929-39:11/2.973-1008) treats only folk plays that are components of calendar rites. There is considerable disagreement about the derivation of folk theater from traditional plays and ritual ceremony. See Cassell's Encyclopaedia of literature (1.156) on the origin of drama: Primitive man, in his struggle to live and increase his kind, recognized some of the forces in nature as his assistants and others as his enemies. All of them, whether good or evil, were the easier to contend with if regarded as human beings possessed of supernatural powers. Sometimes they were conceived of as gods and goddesses; sometimes as giants. At other times they were thought of as spirits, daemons, saints, and devils. By adopting such symbols primitive man not only rationalized his fears, but thereby provided himself with means to combat them. Nature provided a reminder in the annual triumph of Spring over Winter that the most malignant spirits could be banished or at least kept at a distance with the help of those more benevolently inclined. His ritual therefore symbolized this conflict between the forces of life and death with the express object of persuading the former to assist him and his fellows in their own struggle with the latter. In the British Isles the sword dance and mummers' plays of the folk, as degenerate as they have now become and as undeveloped as were even in their prime, bear distinct traces of this ritual origin. Lack of records restricts detailed reconstruction of those rites from which European drama derives. Nevertheless such indications of its general nature as do survive are corroborated by similar characteristics found in the early history of drama in both the Far and Middle Eastern countries. Imitative children's games of trial or punishment often reenact ancient drama (cf. Frazer 1919:442). And there are children's games in which ancient legal procedures are preserved (cf. Kiinssberg 1920, Sebeok and Brewster 1958:68). A fixed line cannot be drawn between folk theater, preserved in oral tradition, and theater cultivated by common people often based on anonymous written or printed texts. Most European folk plays from the seventeenth and eighteenth centuries, both secular and religious, such as plays of wandering puppet-players, etc., derive from written versions. Schmidt (1962) distinguishes between folk drama and semi-literary dramatic forms, and offers a negative characterization of folk drama: Es handelt sich beim Volksschauspiel nicht um das Schauspiel des Mittelalters, obgleich zu diesem eine nahe Verwandschaft besteht und vor allem das Bürgerspiel des Spätmittelalters in jede Darstellung des gesamten Volksschauspiels einbezogen werden muss. Es handelt sich auch nicht um das städtische Fastnachtspiel oder um das Aufführungswesen der Meistersinger usw.; es handelt sich nicht um das Schul- und Ordensdrama der barocken

770

K. HORÄLEK

Gegenreformation. Es handelt sich nog weniger um das Hanswursttheater der Wandertruppen und ihrer Nachfolger, auch nicht um das populäre Vorstadttheater des Spätbarock und Biedermeier, sowie keinerlei um Schauspieltheater und noch so volkstümliches Literatendrama ('Bauerntheater') des 19. Jahrhunderts, und schließlich schon gar nicht um Bühnenfestspiele, chorische Festaufführungen mit oder ohne Freilichtbühne usw. Es bestehen jedoch zu vielen von diesen Gebieten mehr oder weniger starke Verbindungsfäden, die sich im Einzelfall genauer aufweisen lassen. Beim Volksschauspiel handelt es sich um jenes Schauspielgut, das im Rahmen der 'überlieferten Ordnungen' der Volkskultur von den Trägern dieser Überlieferungen gespielt wurde und wird. Unter den 'Ordnungen' sind hier im wesentlichen die brauchmäßigen Festlegungen des gesamten Lebenslaufes zu verstehen, also die Markierungen des Jahresablaufes vor allem. Als Volksschauspiel im engsten Sinn wäre also nur das Brauchspiel anzusprechen, wobei freilich sehr viel an festlichem Spiel zu brauchmäßigen Anlässen hierher gezählt werden darf. Von diesem Mittelpunkt des Brauchspiels her sind dann alle Ausweitungen zu verstehen. Durch diese bewußte Einengung wird auch die nahe Verwandschaft mit dem Schauspiel des Mittelalters erklärlich, das seinerseits so gut wie zur Ganze 'Brauchspiel' war, schauspielhafte Gestaltung der Feste des Kirchenjahres, soweit es sich um geistliches Schauspiel handelte, um Steigerung der Terminfeste und Übergangszeiten, soweit es sich um weltliches Schauspiel im Sinn der Fastnachtsspiele handelte (1962:11-12). Schmidt's anthology of European drama (1965) conforms to his theory only in the first section, entitled "Pièces de théâtre se rattachant à des coutumes populaires" (pp. 12-187). And the claim that all these plays are genuine folk creations is disputed. Some may have a literary source. A short Turkish play in this section, in which an old man fights with a Negro, is read allegorically by the editors as an analogue to the struggle between summer and winter in a play from Germanspeaking parts of Switzerland: Il est évident que par le motif du combat entre le 'Blanc' et le 'Noir', ce jeu anatolien se rattache au groupe des jeux de l'hiver et de l'été, et par celui du meurtre et de la résurrection du vieillard, aux divers jeux d'action. Ceux-ci atteignent leur développement maximum en Allemagne dans les jeux de combat à l'épée et en Angleterre dans les 'Mummers plays'. Mais il existe également dans d'autres pays de nombreux spectacles qui présentent le même motif. Tout à fait caractéristique apparait par exemple le petit jeu-combat entre un Turc et un hussard que l'on jouait encore au dix-neuvième siècle dans le Burgenland autrichien ou dans la région de Oedenburg en Hongrie. Le hussard provoque en duel les Turcs qui, détail significatif, sont décrits comme 'étant noirs' — comme le diable. Le hussard vainc les Turcs mais chante ensuite avec les Turcs allongés sur le sol un chant martial à la suite duquel les Turcs se relèvent, et tout le monde part ensemble. Il faut donc tenir largement compte de nombreuses formes apparentées et le jeu turc n'est pas très éloigné des pièces du théâtre populaire européen de ce groupe (p. 32). However, the characters of the Negro and the black Turk frequently occur in epic folk poetry (mainly in the Balkans), where they sometimes retain daemonic features and thus are not necessarily symbolic. Plays with ritual and allegorical significance may be based on actual experience as well as earlier narrative, e.g. the majority of medieval religious plays based on the Bible (Christmas Nativity plays, the Twelfth Day plays of the Three Magi, Easter Crucifixion plays, plays of the Flood, of Lazarus, etc.); Baroque drama

FOLK POETRY

771

(both Latin school dramas and popular dramas, in some cases actually created by peasants); and plays derived from saints legends and Christian martyrs (of St. Dorothea, of Maxmilian, and others). The plots of a number of folk plays are based on chivalrous novels and popular epic prose (e.g. German Volksbücher), the folk plays of Griselda (and other innocent persecuted women), of Doctor Faustus, and others. Plays especially adapted for puppet-theatres are derived from similar sources: 'The folk or Mummer's play in England, in spite of variations from county to county, even from village to village, is made up of certain well marked sections; the presentation, in which the characters taking part are introduced; the actual drama, in which the combatants argue and fight, in which one of them is killed, lamented and brought back to life; and finally the quête, or collection of gifts or money from the audience by the players. Like all folk literature, this play has undergone change in detail, while retaining its essential outline; lines have been misunderstood or misheard and altered without any further clarity being attained thereby, a frequent enough occurrence in oral transmission; new characters have been introduced, often enough merely to bring the play up to date by including topical references. Sir E. K. Chambers suggests that the traditional text, so far as the figure and adventures of St. George are concerned, is based ultimately upon sixteenth century romance, Richard Johnson's Famous histoire of the seaven champions of Christendom, which was extremely popular and many times reprinted. Other items in the traditional play were perhaps borrowed from the regular travelling theatres of the fifteenth and seventeenth centuries. It has also been maintained that much of the Mummer's play is old enough to have affected the religious drama, the Miracle and Morality plays of the Middle Ages. For the central themes of a death and revival Chambers himself has suggested the influence of primitive European ludus, probably with some original significance other than that of mere amusement, even though it may only dimly survive in a vague notion that the whole thing is done for "luck". This is behind not only the English folk play, but also the seasonal customs still observed in many parts of the Continent in which a death and a revival appear, in the masquerades of Les Rouges and Les Noirs in parts of southern France, the widespread continental sword dances and garland dances, the St. George plays, and the dramatic ceremonies of Haghios Gheorghios in the southern Balkans. Interesting too, though a topic which has aroused and still arouses much controversy, is the possible relation of such a ludus as that found at Haghios Gheorghios to the origin of the drama in Greece' (Cassell's Encyclopedia of Literature 1953:226). In the epic section of Schmidt's anthology (1965) great priority is given to an Italian puppet-play about Roland and Angelica, whose plot is derived from Aristo's Orlando Furioso. Similar literary drama, recalling old-time struggles between Christians and Moslems, is performed in Sicily today, and on these are based some genuine folk plays. One Greek folk play, the plot of which strongly resembles Erophile by the Cretian poet G. Chortadzes (sixteenth-seventeenth centuries), is included in Schmidt (1965).

772

K. HORALEK

Chortadzes studied with his contemporary Italian dramatists (especially G. B. Giraldi), but his work draws on elements of his own folk tradition. A folk play derived from Chortadzes's drama was recorded in northern Greece and is one of the few pieces of evidence demonstrating that dramatic folk poetry flourished even when the church opposed its development. Dramatic folk poetry (as folk prose) in Greece was influenced by the Turkish tradition (cf. Fabula 7.1, 1935). The Turkish shadow play (with special flat puppets) as imitated in Greece is probably of Chinese origin, having spread into Central Asia and Iran during the Mongolian expansion. Analogous to the main character in the Turkish shadow theatre, Karagheuz ('Black Eye'), is the Italian harlequin. Theater historians see models for the Turkish harlequin in the Mediterranean dramatic tradition, primarily the spring carnival processions. The Greeks retained the harlequin's Turkish name (Karagiuzos), but specifically Greek elements permeate the play: social and political satire replaces the light humor of the Turkish shadow theatre. In the nineteenth century some renowned writers wrote plays for the Greek shadow theatre. No reliable information is available on the existence of the shadow theatre among the Balkan Slavs. The Slavic folk puppet-theater, apparently uninfluenced by the European traditions, was most popular with the Czechs and Eastern Slavs. Czech puppet-theater is closely related to the German: Its clown, Kasparek, is directly analogous to the German Kasperl (which in turn resembles the English Punch). Some Czech puppet-theater plays were even adapted from German models (e.g. plays about Doctor Faustus). Neither Czech nor German puppet theater draws significantly on religious subjects; in the Slavic tradition, however, Christmas plays (vertep) enjoyed great popularity in the prerevolutionary period. Use of hand puppets in the East Slavic tradition contrasts with the West Slavic preference for marionnettes. One special form of the East Slavic folk theater is the rayek, 'binocular theater'. A series of pictures (Russian lubki) is reeled from one cylinder onto another, magnified and viewed with binoculars inside a small chest. An oral interpretation accompanies the performance. The name of this popular theater is derived from a favorite serial story of Adam and Eve and their expulsion from Paradise (raj in Slavic). Analogous to rayek are serial song illustrations often used by Western European peddlers to sell song leaflets. Prominent in the repertory of the folk theater of Western Slavs are plays on religious subjects (e.g. on Christmas and Easter, or on the legend of St. Dorothea, a very popular theme among Czechs and Slovaks). The martyrdom of Prince Maximillian is an East Slavic analogy to — and possibly a derivation of — the Czech play about St. Dorothea, which spread to Poland and the Ukraine. The drama Lodka ('ship') is a favorite play of the Russian folk theater. In theme it is classed with 'highwayman's lore', the most customary form of which is the ballad (cf. the English Robin Hood). A distinctive feature of the Lodka play is a song about sailing down the Volga river (Vniz po matuske po Volge), and

FOLK POETRY

773

possibly the play was first an elaborated enactment of this song, drawing from popular illustrated printed versions. The main character of both song and play is the leader {ataman) of a group of highwaymen. The play may have arisen in the second half of the eighteenth century, when — with Pugacev's uprising — the brigand's tradition was topical. A relationship to the song tradition of Stenko Razin is also suggested. In his treatise on Slavic folk plays (1929-39), K. Moszynski mentions that the Russian enthusiasm for the theater is shared by their Finno-Ugrian neighbors (Voguls, Votyaks, etc.); however, no Slavic analogy to the forms of the Finnish theater has been found, e.g. the pantomime. Remarkable differences between the forms of folk theater within the Slavic tradition can be explained by its relatively recent development. Literary Slavic folk theater did not derive directly from early indigeneous plays and ceremonial ritual but in contact with European culture. Today research on the Slavic folk theater has not produced such systematic and complex works as exist for example on the German tradition. The highest level of all has been reached by scholars of the Czech and Russian theater. Basic to the study of Czech folk theater is a collection of folk plays from Moravia by a German scholar, Feifalik (1864), whose work was continued by F. Mensik (1894-95, 1895). Czech puppet plays have been investigated by J. Vesely and J. Bartos, and C. Zibrt's work on Czech ritual plays (1891, 1950, etc.) provides a valuable basis for further research. The investigation of Czech baroque drama was begun by St. Soucek (1929). The functional and structural viewpoint was applied to folk theater by P. G. Bogatyrev (1940), and an informative survey of the Czech folk theater is that of O. Sirovâtka (1962; cf. 1968:290). Basic literature on the Russian folk theater is to be found in the textbook by Novikova and Kokorev (1969:348). Concerning Ukrainian folk theater see Narodna poetichna tvorchist (1965:296-304). Finally, a brief survey of Slavic folk drama has been made by T. S. Romanska (1963:245ff.). A collective work by French scholars (Queneau 1955-56) describes the folk beginnings of the theater in general but gives no detailed information on the folk theater of the East (Indian, Iranian, etc.). Interesting information on dramatic readings of longer stories can be found in the chapter on Malagasy literature (p. 1453). 'Un peu en marge des contes se situe une autre catégorie d'œuvres d'imagination, auxquelles on a, faute de termes propres, donné le nom de fabliaux. Ce sont plus exactement de longs romans à tendance réaliste et satirique, composés sur des thèmes sociaux, autour d'une intrigue amoureuse; ils sont colportés de ville en ville, à la mode médiévale, par des chanteurs ambulants mpilalao. Pierre Camo assista naguère, à Tananarive, à quelques séances de ces troubadours malgaches: Une discours faisait le prologue, les chants exposaient le sujet et la dance mimait l'aventure. — Genre de transition, à la fois romanesque, dramatique et lyrique, le fabliau ouvrait la voie au théâtre, qui ne devait se développer sous une forme affirmée qu'à l'époque moderne.' Early forms of the Arabic theater are noted (p.

774

K. HORALEK

838), and mention is made of the shadow theater (and the figure Karagueuze), but there is no indication that the Arabs adapted this from the Turks. In the chapter on Turkish literature there is no mention of folk drama.

II. TYPOLOGY AND SYSTEMIZATION

The essential component of folklore typology is the theory of types and variants. The concept of type is most systematically applied in the study of folk narrative, whether in prose or song. Types, however, are encountered in all the genres of folk poetry, in lyrics, proverbs, puzzles, and the like, as well as melodic forms. Type may be defined as a set of variants, and the origin of folklore types can be explained in two ways: either one representative text is sustained in variant forms (as is inevitable in the oral tradition), or a type is gradually constituted by the merging of separately derived forms. Parallels of independent origin most frequently occur in elaborate fairy tales and heroic epic poetry, still determined by general principles of constructing narrative forms. Most investigators (e.g. Veselovskii, W. Anderson) dealing with the problem erf independent parallels have stressed the general nature of the motifs and of the simple narrative patterns. 'The very nature of human thought has undoubtedly produced many natural parallels, and there has also been a great deal of diffusion of motifs from narratives of one people over those of another. From the simplest type of preliterate culture to the most sophisticated of our occidental writers, the material that goes into narratives forms a continuum (Thompson 1955:4). Roman Jakobson points out the connection between the origin of independent parallels and the influence of structural laws, discovered by V. Propp and A. Nikiforov in their analysis of Russian fairy tales. 'In folklore as well as in language, only a part of the similarities can be explained on the basis of common patrimony or of diffusion (migratory plots). And, since the fortuity of the other coincidences is impossible, there arises imperatively the question of structural laws that will explain all these striking coincidences, in particular, the repetitive tale plots of independent origins. The remarkable studies of the Soviet folklorists, V. Propp and A. Nikiforov, on the morphology of the Russian folktale have approached the solution. Both of these scholars base their classification and analysis of the tale plots on the function of the dramatis personae. By the concept of function, they mean the deed defined from the viewpoint of its signification for the plot' (Jakobson 1942 [1968: 86]; cf. 1969:94). Some theoreticians deny the existence of folklore types, asserting that the acknowledged types are often hard to distinguish (see Stith Thompson's comments, above) and that transitional forms and recombined forms are common phenomena. The boundaries between the single types do become less definite the greater the number of folklore records at our disposal and the larger the territory in which the

FOLK POETRY

775

investigated material is collected. Thus from the discontinuous process of tradition we derive a continuum, and single types become the result of more or less arbitrary segmentation. Negative or skeptical views of the possibility of classifying folklore material on a typological basis find support in the fact that the classification systems successfully applied to European folklore material have not proved valid for material from other continents. Therefore Propp elaborated his classification of fantastic fairy tales by distinguishing the constant and the variable components of narrative plots. He analysed (1928a) the fairy tales on the basis of thirty-one functions of the dramatis personae and found that all Russian fairy tales are uniform in their structure. Lately Propp's theory has found adherents in the U.S.A., among other countries, e.g. Dundes 1964. Propp's attitude to typological classification is not altogether negative (cf. his edition of Afanasjev's 1956 collection of fairy tales in the commentary to which we find an Aarne-Thompson type classification — with N. P. Andrejev's supplement — with references to other Russian fairy tale anthologies). The Aarne and Thompson classification system (1961) has met with nearly general acceptance, not only in Europe, but also in America and in some Eastern countries. The use of an international catalogue for the classification of exotic fairy tales was substantially improved in the new edition by the introduction of a number of new types on the basis of various material recently gathered (Chinese, Turkish, Indian fairy tales, etc.), as well as by its extensive references to additional material. The use of the international catalogue is spreading due to its authorization by the journal Fabula (Journal of Folktale Studies, 1957-), the chief editor of which is Professor K. Ranke (Göttingen), and thanks to the influence of the Society for Folk-Narrative Research (President, K. Ranke). Several special fairy tale catalogues were issued in the Finnish FFC series based on the international system (Estonian, Magyar, Rumanian, Spanish, Irish, Indian, etc.). A catalogue of Polish fairy tales prepared by J. Krzyzanowski was published in a two-volume enlarged edition in 1962-63 (the first edition appeared in 1947). In the series Folklore Studies of the University of California (Berkeley and Los Angeles) a catalogue of Spanish fairy tales from Central and South America was published by T . L . Hansen in 1957 (as the eighth volume of the series). The catalogue of French fairy tales, edited by P. Delarue (1:1957, 11:1964, completed by M . L. Tenese) has reached type at 736A. A s for the German fairy tales, they are catalogued in K. Ranke's edition of Schleswig-Holsteinische Volksmärchen (1:1955, 11:1957, III-1962, to type at 960). (See also a recent work by Baughmann 1966.) A n edition of Swedish fairy tales has appeared (Sveriges samtliga folksagor in ord och bild, 1:1949, 11:1950), and W . Liungman has added an interesting commentary to this publication as its third volume, on the origin and development of the individual types. This commentary was also printed in German (unfortunately without an index) as Die schwedischen Volksmärchen, Herkunft und Geschichte

776

K. HORÄLEK

(Berlin, Akademie-Verlag, 1961). Liungman also published a collection of Swedish sagas according to type (1957-61). In the FFC series an international catalogue of sagas was published by R. Th. Christiansen (1958). Several recent monographs have appeared which contribute towards a more profound knowledge of individual types, e.g. concerning type AT 939A Killing the returned soldier by M. Kosko (FFC 198), and AT 782 Midas and the ass's ears by M. Boskovic Stulli (Zagreb 1967). Rumanian folklore studies can take pride in an extensive catalogue of humorous folktales (Stroescu 1969). Commentaries which conform to the international catalogue supplement selections of fairy tales are issued in the following series: Fabula Supplement Serie (e.g. collections from Estonia, Israel, China, etc.), Märchen der Welt (primarily its more recent volumes, containing for instance Russian, Magyar, Polish, Czech, and other tales — this series is published by Diederichs Verlag, Köln-Düsseldorf), Folktales of the World (The University of Chicago Press, general editor Richard M. Dorson, including selections from Japan, Israel, Hungary, Norway, Germany, Ireland, Denmark, etc.), Das Gesicht der Völker (Erich Roth Verlag, Kassel — primarily editions of African fairy tales recorded by Professor L. Kohl-Larsen), Volksmärchen — eine internationale Reihe (published by Akademie-Verlag Berlin — this series contains Russian, Magyar, Czech, Turkish, and Byelorussian tales). Commentaries with reference to the international catalogue were appended to some publications in the French series Contes des cinq Continents (Editions Erasme, selections of Catalan, Malgatian, Turkish, and other tales). The Leningrad folklorist Isidor Levin is preparing commentaries with reference to the international catalogue for the Soviet Academy of Sciences (e.g. as a supplement to a selection of Turkish fairy tales arranged by N. K. Dmitriov, 2nd ed. 1967). International catalogue references are regularly added to book reviews published in the periodical Fabula. The use of some lists and catalogues that were not made up according to the Aarne-Thompson system has been made easier by Walter Anderson's reviews, cf. ZSlPh 9.509 ff., 14.227 ff., 18.245 ff., on the three-volume collection of Czech fairy tales by V. Tille (1929-37), and Hessische Blatter für Volkskunde 44:112 ff., on the catalogue of Turkish fairy tales by W. Eberhard and P. N. Boratav (1953). Polivka's important catalogue of Slovak fairy tales (Süpis slovenskych rozprävok I—II, Turc. Sv. Martin, 1923-31), is arranged to correspond to Tille's list of Czech fairy tales (FFC 44, 1927), but unfortunately has not been included in the international catalogue. Every user of the international catalogue knows that it is not a perfect work, and although the new edition of 1961 has been considerably expanded and contains new types whose classification does not always correspond to the original scheme, the sometimes difficult task of locating the individual types has been eased by the introduction of an alphabetic index of types and motifs. However, some international fairy tales types are missing (Ali Baba, Anup and Bata, King Lear, Dido's List, Drosselbart, Great Sinner, Madey, Mother of Saint Peter, Forgotten Bride,

FOLK POETRY

111

Veritas-Falsitas (AT 613), Twelve Month-Men, Hassen of Baara), and in some cases the references for a particular type are sadly incomplete. Another drawback of the international catalogue of fairy tale types is its inattention to narrative songs, which are often derived directly from either fairy tales or from short stories. There is no mention, e.g., of the classical struggle between father and son, both in its literary versions (Ancient Greek Telegonia, Firdausi's Sahname, the Old Irish saga of Cuchulain, the Old Germanic song of Hildebrand, etc.), and also folklore sources (Greek songs, the Russian saga of Ilja Muromec, prosaic Estonian, Caucasian, and Kirghizian fairy tales, and others). Cf. the Russian fairy tale about Jeruslan Lazarevich, which is of Oriental origin. A. van der Lee's monograph on this theme (1957) does not sufficiently examine the folklore tradition. Compare the Greek folk songs of this type recorded by D. A. Petropoulos (1961). The only source mentioned for some of the recently included types is the Indian tradition (the types are adopted from Thompson and Roberts 1960) even when there is available documentation from other countries as well: AT 881A: The abandoned bride disguised as a man [a prince and his bride are separated in a forest; the bride, disguised as a man, is chosen king; she displays her picture in a public place and in this way she finds her husband], for which similar Turkish texts exist, see e.g., Fr. Giese 1925:145, 230; as well as Eberhard and Boratav 1953:262 (No. 215 IV and No. 195). There is no reference in the catalogue to the related type AT 425D, and the basic version AT 881 does not mention the literary treatment of the motif in the medieval novel Pierre et Maguelone, which was also published in booklet form (in Russian, Peter with Gold Keys). Some variants originated within the oral tradition on the basis of the above-mentioned booklets (cf. V. Tille, catalogue I/I.51 sq.). The characteristic features of each individual type are summarized in the international catalogue. Exotic fairy tales usually have three-part plots, cf. e.g. AT 327 The children and the ogre (1. Arrival at the ogre's house, 2. The ogre deceived, 3. Escape); AT 328 The boy steals the giant's treasure (1. Expeditions to the giant, 2. Giant robbed, 3. Giant captured); AT 329 Hiding from the devil (1. The task, 2. Youngest brother undertakes task, 3. Accomplishment); AT 432 The prince as bird (1. The bird lover, 2. The lover wounded, 3. The lover healed); and AT 670 The animal languages. Often we encounter four-part fairy tales (e.g. AT 410 Sleeping Beauty, AT 449 Sidi Numan, AT 879 The Basil maiden, AT 707 The three golden sons). Plots of five or more parts have usually been extended from an original three-part structure, cf. type AT 301 The three stolen princesses: the hero rescues the princesses from an underground hiding place, but later, having been betrayed by his brothers (or helpers), he must remain in the underworld himself where he has new adventures (reminiscent of the Old Babylonian myth of Etana, cf. type AT 313B, 537, whereupon the hero succeeds in returning to the upper world and regains his rights (the conclusion conforms to type AT 974 The

778

K. HORALEK

home-coming husband). In type AT 303 The twins or blood-brothers, one of the brothers marries a princess, whereupon he launches out on an adventure, is transformed by a witch into a stone, from which his brother, whom the princess believes to be her husband, saves him. The complication is signaled in the beginning of the narrative by the motif of the two brothers' close resemblance as well as by the motif of the life-token. In type AT 313, the hero transgresses a law after his escape from the monster's power — he forgets his bride — and she must undergo a complicated test to regain her husband. Propp's structural analysis divided the elaborate fairy tales according to general motifs. The different forms of violence (for example 'rape of a bride or sister', 'stealing an object', 'damaging one's dwelling', etc.) are included in the formulaic expression 'the villain causes harm or injury to one member of the family'. The category "The hero reacts to the actions of the future donor' comprises such cases as 'the hero sustains (or does not sustain) an ordeal', 'the hero answers (or does not answer) a greeting', 'he performs (or does not perform) a favor for a dead person', 'he frees a captive', 'he saves himself from an attempt on his life', etc. In such an abstract formulation of plot construction the binding of individual elements becomes indistinct. Of course, it is not possible to apply a general formula to every fairy tale type (motif, in Dundes' terminology) by wording an arbitrary concrete variant. This may easily be demonstrated with the types AT 560 The magic ring, AT 561 Aladdin's lamp, and AT 562 The spirit in the blue light. These motifs have essentially the same structure (according to Propp's chain of functions), and yet they represent clearly different types. In the fairy tale about the magic ring it is necessary for the hero to win the gratitude not only of the donor of the talisman (this is generally a serpent), but also of a dog and a cat (not of other animals) because these two animals are alloted a special task at the end of the fairy tale, i.e. to help the hero to recover the stolen talisman. This type acquires a special character when the wife helps the performer of the violence to steal the talisman (the version including her betrayal approaches type AT 318). In the fairy tale about Aladdin the situation is substantially different: the task of the dog and the cat would be superfluous here because besides the magic lamp the hero is in possession of another talisman with the help of which he recovers the lamp and also conjures up palace and princess. In type A 562, which is derived from the Aladdin fairy tale, the motif of the second talisman is missing, while the task of the helper is performed here by the hero's friend. Still, Propp's formalized analysis is immensely useful in detecting connections between the individual fairy tale types. This holds true with respect to experiments by Propp's Soviet followers (E. M. Meletinskii, S. I. Nekliudov, G. Novik, et al.) which seek to formulate still more abstract patterns for the fairy tale (cf. Meletinskii's epilogue to the new edition of Propp 1928a); cf. similar endeavors by structuralists in the U.S.A. (T. A. Sebeok, M. Jacobs, J. L. Fischer, et al.) and in France (A. J. Greimas, CI. Brémond, etc.), the Soviet folklorists, in their analyses of tales of witchcraft, have dealt with the

FOLK POETRY

779

ordeal, the loss (or want) of something, charms, an object (its gain and loss), satisfaction of the want, and the final aim of the tale. More complicated forms are discussed in the article "The problem of the description of fairy tales" (Semeiotike, Trudy po znakovym sistemam 4.86-135, 1969). Witchcraft fairy tales must be differentiated: there are heroic fairy tales, fairy tales about pursued individuals, about magic objects, etc. Within the class of heroic fairy tales there are cases when the hero fights to win something for himself and those when the interest of another person is at stake. A formulation in the texts of the recent Soviet research workers may be interpreted as rehabilitation of the classification in the international catalogue. One may, however, express the doubt whether even the new descriptive classification — broad as it is — may still be inadequate to differentiate between all witchcraft fairy tales. Can we indeed expect structural analysis to describe the substance of the fairy tale plot? The theory that fairy tale plots arise from the confrontation of opposing elements, such as departure/return, loss/victory, want/satisfaction of the want, etc., can be criticized. But basic plot construction does not consist of two but of three parts: there are not two elements — 'the task' and 'the fulfilment of the task' — but three 'the task', 'obstacles encountered by those who fulfil it', and 'the overcoming of the obstacles'. The removal of the obstacles that hinder the fulfilment of the task may equal the act of fulfilment, but need not. The plot may be expanded by inserting a new complication before the conclusion, cf. A T 301, A T 313, and A T 519 The strong woman as bride, in which the hero marries with the help of his friend, but the wife detects the fraud and takes revenge on both men, while they try to retaliate. Double construction is a special case among fairy tale plots. One of the characters imitates another, but with the opposite and unwanted result (cf., e.g., Type A T 480 The kind and the unkind girls and A T 676 Open sesame, i.e., Ali Baba). Another frequent pattern in fantastic fairy tales is the repeated attempt which only succeeds the third time. The simple pattern 'intention/realization of the intention' (task/fulfilment of the task) becomes here 'intention/unsuccessful attempt at realization (repeated)/carrying out of the intention (attainment of the aim)'. There are cases when three attempts end in failure (and the hero loses his life and must be revived by means of magic — in fairy tales of Type A T 317 or 302C* the device is usually a magic horse). If structural analysis of narrative plots is truly to aid in historic comparative study, attention must be paid to the tendencies and laws of variant forms. Propp was aware of this himself and wrote a supplement (1928b) to his Morphology (1928a) in which he attempted to formulate the fundamental principles for the distinction of later forms (variants and versions) from the basic forms: 1) The fantastic explanation of the elements of magic fairy tales predates the rationalist explanation. 2) Heroic precedes humorous (parody).

780

K. HORALEK

3) A logically employed form is prior to its confused and nonsensical explanation. 4) International is prior to national. The most important of these for historical comparative study is the third, which implies that fairy tales do not as a rule derive from simple forms by gradual expansion and 'crystallization', but that the historical prototype is most often an individual creative act, in accordance with a limited number of motifs and the accepted principles of their combination. Note the similarity to language, whose basic components are the lexicon (means of denomination) and grammar (rules for the combination of words into larger units). The precondition of any intelligible communication (both artistic and non-artistic) is the code, i.e. set of communication rules. Even in art it is necessary to distinguish the system of communication rules from the actual realizations (texts, works of art). The fact that folklore (folk poetry) is tied to oral tradition and not the written text does have certain consequences; more favorable conditions for variability of composition prevail, but firm limits are maintained by centripetal tendencies from which arises tradition. Structural analysis of a form does not deny descriptive typology, but complements it. Evidence for this is found in Propp's edition of Afanasjev's collection of Russian fairy tales, mentioned above. The usefulness of the international fairy tale types does not lie in their systemization, but in the typing of variants. Therefore, an indispensable part of the catalogue is the index, the shortcomings of which have been mentioned above — however Thompson's Motif-index of folk-literature (1955-58) compensates for the inadequacies of the index to a considerable extent. It could actually be regarded as the supplement to the catalogue if it were based to a greater degree on the motif analysis of fairy tale types. The Motif-index does not of course include only material from fairy tales, but also from some other genres of folk literature, e.g. the narrative (including ballads). Earlier short tales from Western European literatures are included but Slavic literatures have been entirely omitted; Oriental literatures are surveyed in Western translation. The sixth volume of the Motif-index (892 pages) contains an alphabetical index of the most important words with references to the entries. The material in the Motif-index is grouped under several themes, each of which is in turn further subdivided. The Motif-index includes material from other special motif-indexes, most of which were initiated by Professor Thompson, some with his participation (Thompson and Balys 1958, Cross 1952, Rotunda 1942, Keller 1949, Noymann 1954, Balys 1936). A combined catalogue of types and motifs of Cheremis (Mari) fairy tales was prepared by A . N. Nyrges (1952), cf. Baughman (1966). Different principles underlie the work of N. Elisseeff (1949). Attempts at typological systemization of myths and legends are based on experience with the classification of fairy tales. Difficulties are encountered since legends do not represent narrative forms with a fixed plot pattern. In addition, the material is greatly differentiated and much more closely tied to other forms of

FOLK POETRY

781

culture than in the case of fairy tales. Legends differ from fairy tales mainly in that they relate fictitious events as fact. Legends are usually localized (both in space and in time), and often connected with well-known persons. Legends differ from myths mainly in their characteristically descriptive statement, whereas ait important role is played in myths by explicative elements. In fairy tales, the mythical elements usually lose their explicative character. Still, myths and legends are not easily separated. The international catalogue includes some tales that might be regarded as myths, e.g. AT 471 The bridge to the other world. In the Section on "Religious Tales" (Nos. 750-849), tales of an intermediate character between anecdotal fairy tales and religious legends predominate. Within the Section on "Romantic Tales" (Nos. 850—999), mythic features appear in "Tales of Fate" (Nos. 930-949), e.g. the retold classical tales of Oedipus (AT 931) and the related Christian legend about Pope Gregory (AT 933 Gregory on the stone), which in the South Slavic tradition are sung. Type AT 930A The predestined wife assumed the form of a song in the Russian tradition (bylina about Svjatogor). Type AT 934A Predestined death more often occurs as a localized legend. This type is also represented by the Old Egyptian fairy tale about a prince to whom the goddesses of fate (Hathors) foretold that his death would be caused by a crocodile, a snake, or a dog. The conclusion of this tale has not been preserved. According to the modern parallels (tales about triple death), one might speculate that the Old Egyptian tale also ended tragically and consequently that it was a legend rather than a fairy tale. In the modern European version tales in which predestined death is averted are usually Christian legends. The problems of typological classification of legends are being dealt with by a special commission of the Society for Folk Narrative Research. In 1963, the commission held a large international conference (Ortutay 1963). Information on the typological systemization of legends was provided by Wayland D. Hand (1965). An important contribution to the cataloguing of international legends was made by Reidar Th. Christiansen (1958), and a catalogue of Finnish mythic legends was worked out by Lauri Simonsuuri (1961). Work on an international catalogue of folk balladry is only beginning. The pioneer work done in this field by Francis James Child in his five-volume annotated collection The English and Scottish popular ballads (1882-94) has never been equaled. A classification according to song type is always involved in folk song collections which assign variants (or at least information about them) to representative texts. This was done e.g. in the large collection of Danish songs whose publication was begun by E. S. Grundtvig (1953, Danmarks Gamle Folkeviser, reprint 1967). A similar work is the collection of German songs edited by L. Erk (1893-94). A comparative survey of French narrative songs on the basis of reconstructed texts was made by George Doncieux (1904). More recently, various selections of Slavic songs have been published, e.g. Polish songs by S. St. Bystron, Slovak songs by J. Horak, Croatian by M. Boskovic

782

K. HORÁLEK

Stulli, etc., in which the songs are classified according to traditional content. Abstracts were provided in the catalogue of Bulgarian folk songs by A. P. Stoilov (1916-18); in the 'twenties Stoilov's work was continued by St. Romanski. Lithuanian ballads were described in a similar manner by Jonas Balys (1954). A method similar to that worked out by Child, but in a more elaborate form on a broader comparative basis, was employed in the monumental collection of German ballads, edited by J. Meier (1935-67). Ninon A. M. Leader (1967) combined Child's method with more recent practices. The foundation of the typology of Czech folk songs was laid by Fr. Susil in his edition of Moravian folk songs (1860). After a preparatory edition which appeared in two volumes (1835 and 1840), Susil proceeded with an edition in which he classified the song types in such a way that each basic text was accompanied by variants (when available). There are some references to analogous Slavic material. His method satisfies modern requirements much better than that of K. J. Erben (1864), who used variants to reconstruct 'complete' versions. In systemization, Susil went further than Kollár, who collected a large amount of Slovak material in his edition Národnie spievanky (1834-35; see 1953). The typological point of view was applied by the Polish collector O. Kolberg (1857). For some types he gives as many as fifteen variants. In the edition of Russian folk ballads and lyric songs, similar principles were adhered to by A. I. Sobolevskii (1895-1902). As for the Russian byliny, the basic typological characteristics with bibliographical data are given by A. M. Astakhova (1938-51). Similar commentaries are added to the edition of Bulgarian representative texts, Bálgarsko narodno tvorcestvo (1961-63 — epic songs are contained in volumes I-IV). Detailed comparative commentary with data on archival variants is provided in most volumes of the series Hrvatske narodne pjesme (Zagreb 1896-1948). The classical collection Srpske narodne pjesme by Vuk St. Karadzic does not yet include information on the known variants. Collections of Slavic songs from Macedonia (Miladin brothers, Verkovic) have so far been edited without commentary. Typological comparison of European folk songs has been undertaken since the end of the nineteenth century, cf. some early works by A. Veselovskii and other Russian investigators. More recently, these studies have been criticized because they place too much emphasis on borrowing and do not allow for the possibility of independent variation. Zirmunskii (1958) raises objections often valid for traditional comparative studies. His main argument is that if elements of different European epics have Asian parallels that cannot be assumed to have genetic connection with the European tradition, then the genetic explanation does not always apply within the European area. We would modify this argument: because similar phenomena exist in the Oriental tradition, it is necessary to allow for the existence of independent parallels in the European tradition, even though there has undoubtedly been borrowing. In some cases Zirmunskii himself accounts for certain elements in the European tradition by the assumption of contact. He finds a com-

FOLK POETRY

783

mon basis for some aspects of the Russian and German medieval epic in the heroic fairy tale, and similarly explains some analogies between Homer's Odyssey and the Middle Asian version of Alpomys. Opinions differ but traditional comparative study is the foundation of the typology of European song epic. In the works of Veselovskii, I. P. Sozomovic, M. G. Khalanskii, M. P. Drahomaniv, J. Máchal, T. Maretic, V . Hnatiuk, S. Banasevic, Horák, J. S. Bystroñ, J. Meier, E. Seemann, and other investigators, conclusive evidence has been collected on the relationship between Slavic ballads and heroic songs and German and Romance songs. Future investigation may show that all the cases studied are the result of independent development. Whatever the final explanation of origin, a large amount of comparative material has been collected, and new material is always forthcoming, thus subjecting all prior conclusions to reexamination. Short lyrical songs vary too greatly to be grouped according to type. They are usually arranged in alphabetical order according to first line in the indexes to folk song editions. A general index of Czech folk songs on this basis is Zibrt (1895). Dialect forms are sometimes standardized, but Zibrt did not choose to do so, e.g. he lists incipits with the word 'Kdyz' separately from the dialect variant 'Dyz'. Short (usually four-line) songs are generally under-represented in such indexes, the German Schnaderhiipfel, Russian castuski, Ukrainian kolomyjky, Polish krakowiaki, etc. However, Piesni Podhala, a Polish anthology edited by J. Sadownik and his team (1957) includes primarily four-line songs, with some five-line, and a few one- or two-line songs. In addition to a first line index, there is a double thematic index: one groups the songs according to subject (erotic, social, pastoral, military), the other according to some key word, in alphabetical order (e.g. Adamy, Ameryka, Baba, Bieda, etc.), with reference to the relevant thematic group. There are 1,250 songs. A similar system was used in the Russian edition of Castuski v zapisiach sovetskogo vremeni (Vlasova and Gorelov 1965). There are 8,230 songs in this collection, arranged by geographical area of origin. There are several indexes, although none by first lines: thematic types, geographical names, personal names, and finally an alphabetical index of subjects or key words (pp. 443-93). This collection is characteristic in that the editors do not attempt to refer to cognate songs in other works. The monumental Polish edition of proverbs and sayings, Nowa ksiqga przyslow i wyrazen przyslowiowych polskich (1969), lists cognate proverbs under one main entry, cf. the Russian collection Poslovitsy, pogovorki, zagadki v rukopisnych sbornikach XVIII-XX vekov (1961). In the edition Zagadki, edited by V . V . Mitrofanova (1968), riddles were arranged thematically with an index of variants attached. In proverb collections three methods of classifications are the most popular. Warren Roberts, in his essay on Cheremis proverbs, explains:

784

K. HORALEK

They have been arranged alphabetically according to the first word, as in Marta Warren Beckwith's Jamaica proverbs, alphabetically according to the most important word, as in Edwin M. Fogel's Proverbs of the Pennsylvania Germans, and by a subject matter classification, as in le Roux de Lincy's Le Livre des proverbes français. If the structure and wording of proverbs did not vary in oral transmission, an arrangement alphabetically according to the first word would be satisfactory so long as the proverbs of one language only were treated. This system must, however, be discarded if the variant forms of a proverb are not to be scattered throughout the collection. An alphabetical classification according to the most important word has advantages. It is easy to use, and variant forms of a proverb in which the structure is changed but the wording remains the same can be brought together. With a group of proverbs in which two or more languages are represented, however, an alphabetical arrangement becomes less satisfactory. Moreover, closely related proverbs whose wording differs often must be widely separated when such a system is used , and only a complex series of cross references could compensate for this defect. For instance, anyone using this collection who was interested in the proverb 'Man's hand bends toward himself would probably be interested in knowing that the form 'the dog's tail curves toward himself' also appears. In any alphabetical classification these two examples would be separated widely, and only a dilligent search through the entire collection would make one feel sure that he had located all the variant and related forms of a given proverb. A subject matter classification, such as is used in this collection [of Cheremis proverbs], minimizes the problems of handling material which is in more than one language. In most cases, too, related proverbs are brought together irrespective of their wording. In a few instances cross-referencing has been necessary. Unfortunately, a subject matter arrangement is difficult to use, for the person desiring to locate a proverb may well have a concept of its meaning or basic thought different from that of the person who classified it. As an aid in locating a particular proverb, however, an index of key words has been provided. The system of classification which has been used in our collection is a condensation of one devised by Jonas Balys for use in the Lithuanian Archives. Balys' system was used for a model because it was based upon an examination of an adequate number of examples. In the first group, proverbs which deal with different qualities and abilities of individuals are included under various subheadings. The second large group deals with plans and preparations and with various ways of conducting oneself in his daily life. The third group deals with property and allied matters, such as ways of attaining it and ways of handling it. Proverbs dealing with the various traits of women, the relation between man and wife, marriage, and related matters are included in the fourth section. The fifth group includes a large number of subgroups which treat with various phases of the relationships between the individual and those about him. Those proverbs dealing with the relations between different classes compose section six. The proverbs in group seven concern the related aspects of youth and old age and the benefits of experience. In the eighth section are proverbs concerning more or less abstract concepts, such as death, fate, or God's will, and cause and effect. The final section includes some proverbs which do not fall readily into the other classes, some rather obscure proverbs, and a few examples, such as 'The cripple holds it, the lame reaches it', which seem to be riddles without answers, but which may be proverbs (Roberts 1952:118). The classification of folklore texts according to type is possible in all genres, but the degree of applicability depends on their variability or stability. Within the tradition all genres exist only in the form of variants, and variants exist as such

FOLK POETRY

785

only in relation to types. With the comparatively stable genre such as proverbs, a simple classification based on dominant content components is practical, disregarding possible type division. But the notion of type gains importance even in these cases, if the classification is undertaken on a larger, international scale. Using the notion of type and variant is difficult because in some genres there are many transitional compositions and 'conglomerates' so that any division into individual types is more or less subjective, often arbitrary. Even in the international catalogue of folktale types (Aarne and Thompson 1961), which is used throughout the world, there are many debatable cases which may be modified when new variants are found. In describing and classifying narrative genres (tales and stories, ballads, etc.) it was necessary to introduce a notion of version or sub-type. One and the same type may have several versions historically or geographically conditioned. The version is a group of variants sharing some features but not enough to isolate such a group as a type. The geographical version delimited by a linguistic community has been termed 'oicotype' by C. W. von Sydow. Sydow's general theory is based on his criticism of the migration theory, yet the term oicotype is used by scholars who regard migration and diffusion as useful notions but who are otherwise uninfluenced by Sydow's theoretical system. Von Sydow was convinced that fairy tales (his 'chimerate') are of Indo-European origin, and in working out his theory he has turned the notion of an oicotype into a predominantly historical category. He associates the constitution of oicotypes with the earliest cultural and linguistic differentiation of the Indo-European peoples. Von Sydow's view can be instructively illustrated by his explanation of type A T 302: The folk-tale about the Giant without a heart is spread among all Indo-European peoples. A giant who has carried away a princess cannot be vanquished because his heart is hidden and strongly guarded. But the hero learns the secret from the princess, and succeeds in gettting hold of the heart. When it is destroyed the giant dies. This folktale may be divided into two great oicotype groups: the Asiatic, which has it that the giant's life or heart is in, or identical with, a bird or insect in a distant place, guarded by demons or a dangerous beast; the European which has it that the heart is in an egg in a bird which is usually enclosed in a whole series of beasts that have to be killed first. Here the migration theory cannot be used, for why should just the boundary between Europe and Asia compel so thorough a remodelling of the tale? The most natural explanation is that the folktale was originally a much simpler narrative, based directly on people's belief; but it incited the imagination to extend the account, to make it more thrilling by increasing the difficulties that the hero has to overcome. In so doing, different ways have been followed in Europe and Asia. This puts the origin of the tale back to a couple of thousand years B.C., and this is not the least unreasonable, as the way of thinking is extremely primitive in it, and the tale seems to have influenced the Egyptian tale of the Two Brothers, which was written down about 1300 B. C., and with which I shall deal more closely later on in this paper. This tale of the 'giant without a heart* can also give another proof of its age.

786

K. HORÄLEK

Among the Slavs and the Celts there developed oicotypes of a kind peculiar to these groups of peoples, but both these oicotypes exhibit such mutual correspondence compared with the Teutonic oicotype that they must have a special connection with one another. A migration over Teutonic territory is not to be thought of: the explanation must lie in the fact that Celts and Slavs dwelt next to one another about 600 B.C. with the result that they then had a whole number of traditions, and among them several folktales, in common (von Sydow 1948:56f). Some folklorists (V. Tille and J. Polivka, among others) have cast doubt on whether it is possible adequately to describe and classify the whole narrative tradition by means of a type catalogue. Propp's structural analysis of fairy tales is usually cited, but discussion continues (cf. Jason 1968). B. Holbek, in his paper delivered to the IV International Congress for Folk-Narrative Research in Athens in 1964, criticizes the approach: I do not intend here to combat the monographic method, but to draw attention to the fact that folktales are not amenable to division into clearly distinct types, as would be supposed from our type index. Nor should we forget that each narrator of tales has been acquainted with several folktales and has had the possibility of combining and re-casting, as inspiration would suggest at any given moment. Nor will this be novel to my audience, but it is a reflection which did not occur to the scholars who laid down our typology at the beginning of this century. They should not be blamed for that, but we shall be blamed if we do not draw the consequence of our substantially increased knowledge of the informants and of the dynamics of their narrative arts. In my opinion we shall sooner or later be obliged to free ourselves of the philologically inspired typology which formed the basis of Antti Aarne's Verzeichnis, and set about to re-arrange the whole material of folktales, presumably into groups according to main themes, where Aarne rather robustly classified according to the kinds of magic and witchcraft encountered by the hero of the tale. Presumably, it will not prove possible to carry out such a re-arrangement simultaneously for the whole field comprised by the international type index, and since it will necessarily be more searching in its analysis, it will certainly become too comprehensive to be practically feasible, but we must be set to work region by region. In this connection I would point to the Schleswig-Holsteinsche Volksmärchen by Kurt Ranke and to Le conte populaire français by Paul Delarue, which would seem useful as models for the work to be done. Thus type 300 is in Ranke divided into five sub-types, and that within such a limited area as Schleswig-Holstein! (1965:160).

III. FOLK POETRY AND LITERATURE

Literature is basically derived from oral folk poetry. Even translated literature, e.g. Old Church Slavonic which is made up of translations (mainly from Greek) as well as particular stylistic traditions derived from these translations, can in some way be linked to the oral tradition (through verse forms, metaphors, etc.). But literature can also be in complimentary opposition to oral poetry, for example in the Old Slavic tradition. The Ancient Christian literature developed apart from both the oral tradition and also the literatures of the Ancient Near East (especially Hebrew), of Greece, and of Rome.

FOLK POETRY

787

We know the old Christian legends, reflecting as they do the classical tradition, primarily from various medieval versions. The legend of Judas is linked to the myth of King Oedipus by the unavoidable fateful prophesy. Another Christian legend, that of Pope Gregory, is related to the Oedipus myth by the theme of incest. Legends have been retained in Old Church Slavonic literature from which Slavic folk versions are derived, e.g. the Serbian folk song about the foundling Simeon. Although no Greek (Byzantine) version of this legend has ever been found, one is assumed on the basis of the Coptic and Aramaic versions. Another case in point is the legend about the great sinner (AT 756B) which was examined in a monograph by the Russian scholar N. Andreev. The boy, who is promised to the devil, grows up and sets out for hell in order to break the contract. Along the way he meets a robber and a murderer and is asked to find out, when he gets to hell, what sort of punishment awaits them. The sinners are to repent by watering a burning branch until it turns green. There is a similar legend about two sinners, one of whom wins forgiveness by killing the other, more guilty one (AT 756C). We also know of a Moslem version of this legend which probably arose somewhere in the Near East. We find unlikely L. Vargyas's suggestion that type AT 756B emerged in the Shaman milieu of Siberia and was brought to Europe by the Magyars. Old Christian legends contain elements known in Old Egyptian literature, e.g. the legend of the burning and rebirth of a saint reflects in several of its motifs the Old Egyptian tale about the two brothers. Many Buddhist legends have had a development similar to the adaptation by Christians of early fables. See the jatak collection (tales about Buddha's former lives) in the Pali language; other similar collections of Buddhist tales are found only in Chinese translations. The medieval Christian legend about Barlaam and Josaphat (the Greek form is Joasaph) comes from an old tale about Buddha. A Latin version followed the Greek, and various European versions were derived from the Latin. The Old Church Slavonic version is based directly on the Greek. The connecting link between the Indian and the Greek versions was the Persian (Pehlevi) of which there is unfortunately no record. There were several Arabic translations of the Persian, and the Syrian version was probably the direct model for the Greek. By means of translation from Buddhist collections, Indian tales also made their way into the oral tradition of Central Asia and Asia Minor, particularly in Tibet, Mongolia, Persian, Arabia, Turkey, and the region of the Caucasus. Based on the Greek version of the frame-tale cycle Vetalapancavimsati, an original Mongolian collection Siddhi Kür emerged which contains several interesting variations of fairy tales otherwise known mainly from the European models. Th. Benfey explained the proximity of certain stories in the Siddhi Kür collection to the European tradition by direct Mongol influence in Europe, but this is most uncertain. In this regard the Gypsy variation published by H. von Wlislocki is unreliable. We encounter traces of the oral tradition in the old Indian Veda; here we have

788

K. HORALEK

authenticated material about the girl-bird that we know from various European fairy tales (AT 313, 400, etc.). The very oldest literary records that exist from the literature of Ancient Babylonia, Sumeria, Egypt, Greece, Persia, and also India developed from the oral tradition. For instance, the Babylonian myth about Etana (cf. AT 537) is based on the oral tradition, as is the epic about Gilgamesh. The link of several Ancient Egyptian tales to the oral tradition is indisputable, particularly the tale about the two brothers (AT 318), and about truth and falsehood (AT 613). The story of the two brothers contains material from the Biblical tale of Joseph and Potiphar's wife. Folklorists believe for the most part that the structure of this subject is too simple to allow indisputable classification but a more detailed analysis of the Old Egyptian tale leads to a very clear conclusion. The old Egyptian version has striking similarities only within the North African tradition; a genetic connection is certain and can be explained by the oral tradition alive in North Africa today. The dependence of Ancient Greek literature on the oral tradition is absolutely clear (Homer's Odyssey, classical dramas, the histories of Herodotus, etc.). Dante's Divine Comedy carries on the ancient tradition of story-telling about travels to the next world (the epos about Gilgamesh, Homer's Odyssey, the Egyptian story about Si-Osir, etc.), and unquestionably draws on medieval legends, some of which in turn influenced the oral tradition. Dante's friends testified to his knowledge of Arabic literature. There are unexplained Asian parallels to the European folktales about journeys to the other world (see type AT 460, 471). Vargyas suggests that type AT 471 The bridge to the other world is of Shaman origin and made its way to Europe through the Magyars. Shakespeare indirectly drew on the oral tradition, utilizing folklore subjects already in literary form. His Hamlet, very likely, has a foundation in folklore: it was influenced by Kid's drama, now lost, which most probably reflected the tales in the Saxus Grammatikus chronicle (Gesta Danorum) which Shakespeare might have known in Bellefort's French version. Saxus's tale about Amleth unquestionably derives both from the oral tradition and from literature. Hamlet's feigning madness has a parallel in Roman literature (the tale of Brutus). The folkloric material related to Hamlet resembles both ballad songs (in the Scandinavian tradition) and prose tales. Literary works about Don Juan have very many connections to the folklore tradition. Molière's drama (derived from a drama by the Spanish poet Tirso da Molina) and Mozart's opera (with libretto by L. da Ponte) are the most famous. The main element of this tradition is the invitation by a blasphemer to a dead man (or his skull). The oldest known version is a medieval exemplum describing life after death dating back to the middle of the fourteenth century, but the roots of this tradition are much older and have an archetypal character. The motif of an avenging statue lives in the folklore tradition and in literature independently: P. Mérimée used this motif in his novel La Vénus d'Ille, cf. de Maupassant's novella

FOLK POETRY

789

about the mummified hand. L. Petzoldt (1968) relates the tradition of the dead guest (AT 470, 470A). The connection between Goethe's Faust and the popular booklet about Doctor Faust (1589) is well known, its plot based on an oral tale which most likely grew out of some literary legend. Related to this are old Christian legends about Simon the Magician and about Cyprian; in addition there is the medieval legend about Theofil who sold his soul to the devil in order to regain the church position from which he has been removed. Goethe may also have used the English drama by Marlowe (1594). He would have been familiar with the folk text for the puppet theater and songs sung at fairs, but no evidence has been produced to prove that Goethe's Faust was greatly influenced by popular versions. A connection between the Polish tradition of Pan Twardowski and the German tradition still awaits a detailed explanation. The Faustian motifs in the Old Russian tale about Sava Grudtsin (seventeenth century) are derived directly from older traditional legends. Analogies to Faust can also be found in Ancient Egyptian stories about Setna. Some famous novels, beginning with Cervantes' Don Quixote, should be included in a history of relations between literature and folk poetry. Don Quixote is patterned on a semi-romantic folk epic in which the hero experiences various adventures on his journeys in the picaresque tradition. A classic example is the popular book about Eulenspiegel. The novels of Dostoevsky, in the tradition of Gothic criminal literature, are linked to folk literature; cf. in particular his novel The brothers Karamazov. The links between national literature and oral folk poetry can only be characterized on the basis of a special examination of individual cases. Even in the literatures of such closely related peoples as the Slavic there are great differences. Literature supercedes folk poetry only when it creates favorable conditions for the development of reading skills. The introduction of printing and popular education is a great historical landmark. Professional storytellers and singers (bards) who borrow from literature maintain contact, as do religious institutions and customs. In the Middle Ages, Christian preachers enriched the folktale tradition by adding new subjects, and church music and song influenced folk song tradition. Various bans by religious authorities often repressed the oral folk tradition as did the introduction of international legends, songs, and rituals (in e.g. Buddhism, later in both Christian and in Islamic tradition). The histories of literature are inadequate for the formulation of a general concept of the relations between folk poetry and literature. Even the Soviet literary encyclopaedia, which was especially designed to further such investigation, fails to cover its own territory.

790

K. HORALEK BIBLIOGRAPHY I. FOLK POETRY I N ITS HISTORICAL PERSPECTIVE (pp. 741-74)

A. General (pp. 741-46) 1.

Folk

literature

and V . K. SOKOLOVA, eds. 1 9 6 5 . Slavianskii fol'klor i istoriceskaia deistvitel'nost'. Moskva. BALYS, J. 1948. Lietviij tautosakos skaitymai [Handbook of Lithuanian folklore]. Vols. I—II. Tübingen. BASCOM, W. 1957. The myth-ritual theory. JAF 70.103-14. BAUSINGER, H. 1968. Formen der Volkspoesie. Berlin. BHAGWAT, D . 1958. An outline of Indian folklore. Bombay. BOBERG, I. M. 1953. Folkemindeforskningens historie i Meilern- og Nordeuropa. Copenhagen. BOGATYREV, P. G., ed. 1943, 1956. Russkoe narodnoe poeticeskoe tvorcestvo. Vols I—II. Moskva. BÜLTE, J . , and G . POLIVKA. 1 9 3 2 . Anmerkung zu den Kinder- und Hausmärchen der Brüder Grimm, Vols. I—II. Leipzig. CARVALHO, NETO, P . DE. 1 9 5 6 . Concepto de folklore. Montevideo. CIÖEROV, V. T . 1 9 5 9 . Voprosy teorii i istori narodnogo tvorcestva. Moskva. [See also Tschitscherow, W. I.] COCCHIARA, G. 1954. Storia del folklore in Europa. 2nd ed. Torino. CORSO, R . 1953. Folklore, storia, obietto, metodo, bibliografia. 4th ed. Napoli. DINEKOV, P . N. 1959. Bälgarski foMor I . Sofija. DORSON, R . M. 1959. American folklore. Chicago. . 1963. Current folklore theories. CAnthr 4.93-112. DUNDES, A . , ed. 1 9 6 5 . The study of folklore. [Readings.] Englewood Cliffs, N . J . DURIC, V . J . 1 9 5 7 . Postanak i razvoj narodne knjizevnosti. 2nd ed. Beograd. GREENWAY, J . 1 9 6 4 . Literature among the primitives. Hatboro. GUSEV, V . E. 1 9 6 3 . Problemy fol'klora v istorii estetiki. Moskva. . 1967. Estetika fol'klora. Leningrad. HACKMAN, O. 1904. Die Polyphemsage in der Volksüberlieferung. Helsinki. HYMAN, S. E. 1955. The ritual view of myth and the mythic. JAF 68.462-72. JACOBS, M. 1959. The content and style of an oral literature; Clakamas Chinook myths and tales. VFPA, New York. JAMES, E. O. 1958. Myth and ritual in the Ancient Near East. London. KRAPPE, A. H . 1930. The science of folk-lore. 2nd ed. 1962. London. KROHN, K . 1892. Eine griechische Tierfabel; am Urquell. KRZYZANOWSKI, J . , ed. 1965. Slownik folkloru polskiego. Warszawa. LEACH, E. M., ed. 1949-50. Standard dictionary of folklore, mythology and legend. Vols. I-II. New York. ASTAKHOVA, A . M . , B . N . PUTILOV,

FOLK POETRY LIUNGMAN,

791

W. 1961. Die schwedischen Volksmärchen, Herkunft und Geschichte.

Berlin. 1 9 5 9 . Slovensky folklor; chrestomatia. Bratislava. 1966. Grundriss der sorbischen Volksdichtung. Bautzen. NOVIKOVA, A . M . , and A . V . KOKOREV. 1 9 6 9 . Russkoe narodnoe poéticeskoe tvorcestvo. Moskva. PENZER, N. M. 1952. Poison-damsels, and other essays in folklore and anthropology. London. PROPP, V . IA. 1 9 5 6 . Istoriceskie korni volsebnoi skazki. Leningrad. RAGLAN, LORD. 1 9 5 5 . Myth and ritual. J A F 6 8 . 4 5 4 - 6 1 . RAMOS, A. 1958. Estudos de folklore: Definigao e limites, teorías de interpretacao. 2nd. ed. Rio de Janeiro. ROMANSKA, T. S. 1963. Slavianski folklor; Ocerci i obrazci. Sofija. RYL'SKYJ, M., ed. 1965. Ukrainska narodna poetycna tvorcist'. Kyjiv. SCHWARZBAUM, H. 1968. Studies in Jewish and world folklore. Berlin. SIROVÁTKA, O . 1 9 6 7 . Ceská lidová slovesnost a její mezinárodni vztahy. Praha. SOKOLOV, Iu. M. 1950. Russian folklore. New York. [Russian original 1938: Russkij Fol'klor. Moskva.] SPENCE, L . 1 9 4 7 . Myth and ritual in dance, game and rhyme. London. THOMPSON, S . 1 9 4 6 , 1 9 5 1 . The folktale. 2nd ed. 1 9 5 1 . New York. , ed. 1953. Four symposia on folklore. Bloomington. TOSCHI, P . 1 9 6 2 . Guida alio studio delle tradizioni popolari. 3rd ed. Torino. TSCHITSCHEROW, W. I. (See also Cicerov.) 1968. Russische Volksdichtung. Berlin. [Russian original 1959: Russkoe narodnoe tvorcestvo. Moskva.] VANSINA, J. 1961. De la tradition orale; essai de méthode historique. Tervuren. [English translation 1965, London.] VEGA, C . 1 9 6 0 . La Ciencia del folklore; con aportaciones a so definición y objecto y notas para so historia en la Argentina. Buenos Aires. VRABIE, G . 1 9 7 0 . Folclorul: obiect, principii, metoda, categorii. Bucuresti. WEBER-KELLERMANN, I . 1 9 6 9 . Deutsche Volkskunde zwischen Germanistik und Sozialwissenschaften. Stuttgart. WINNER, T. G . 1958. The oral art and literature of the Kazakhs of Russian Central Asia. Durham. MELICHERCIK, A . NEDO, P .

2. Folkloristics Folklore research around the world. JAF 74.287-460, 1961. [Rich in information concerning Slavic countries except Russia.] ORTUTAY, G. 1955. The science of folklore in Hungary between the two world wars and during the period subsequent to the liberation. AEH 4.5-89. PEUCKERT, W . - E . , and O . LAUFFER. 1 9 5 1 . Volkskunde: Quellen und Forschungen seit 1 9 3 0 . Bern. [Cf. pp. 1 2 3 - 2 6 2 for a survey of the investigation of folk poetry.]

792

K. HORALEK

The Yugoslav-American Folklore Seminar. Journal of the Folklore Institute 3.217-418. [Cf. pp. 398—418 for a survey of the study of folklore by F. J. Oinas.] 3. Major periodicals USA: Journal of American Folklore (1888-), Western Folklore (1947-), Journal of the Folklore Institute (1964-, formerly Midwest Folklore), Publications of Texas Folklore Society (1911—). England: Folklore (1890-). Austria: Oesterreichische Zeitschrift für Volkskunde (1947; formerly Zeitschrift für österreichische Volkskunde, 1895- and Wiener Zeitschrift für Volkskunde, 1919—). France: Arts et Traditions Populaires (1953-). Czechoslovakia: Cesk# lid (1892-), Slovensky narodopis (1952-). German (FGR and GDR): Hessische Blätter für Volkskunde (1902-, formerly Blätter für hessische Volkskunde, 1899-), Zeitschrift für Volkskunde (1929formerly Zeitschrift des Vereins für Volkskunde, 1891—), Deutsches Jahrbuch für Volkskunde (1954-), Rheinisches Jahrbuch für Volkskunde (1950-), Jahrbuch für Volkslied-forschung (1931—), Fabula: Zeitschrift für Erzählforschung (195 8-), Demos (1950-). Greece: Laographia (1909-). Ireland: Journal of the Folklore of Ireland Society (1927-). Italy: Lares (1930-), Tradizioni (1961-). Poland: Lud (1895-), Literatura ludowa (1957-). Hungary: Studia Ethnographica (1959-). Rumania: Rivista di etnografie si folclor (1964-, formerly Rivista de folclor, 1956-). Soviet Union: Russkii fol'klor (1956-). Switzerland: Schweizerisches Archiv für Volkskunde (1897-). Yugoslavia: Slovenski etnograf (1948-), Narodno stvaralastvo-folklor (1962-), Narodna umjetnost (1962-), Rad kongresa folkorista Jugoslavije (1965-), Makedonski folklor (1968-). 4. Recent Festschriften Studies in folklore; in honor of Distinguished Service Professor Stith Thompson, ed. by W. E. Richmond (Bloomington 1957). Humaniora; essays in literature, folklore and bibliography, honoring Archer Taylor on his seventieth birthday, ed. by W. D. Hand (New York 1960). Folklore and society; essays in honor of Benjamin A. Botkin, ed. by Bruce Jackson (Hatboro 1966). Folklore international; essays in traditional literature, belief, and customs, in honor of Wayland Debs Hand, ed. by D. K. Wilgus, with the assistance of Carol Sommer (Hatboro 1967).

FOLK POETRY

793

Volksüberlieferung; Festschrift für Kurt Ranke zur Vollendung des 60. Lebensjahres, ed. by F. Harkort (Göttingen 1968). Märchen - Mythos - Dichtung; Festschrift zum 90. Geburtstag Friedrich von der Leyens (1963). Literature, Komparatystyka, Folklor; Ksiçga poSwiçcona Julianonowi Krzyianowskiemu (Warszawa 1968). Brüder Grimm Gedenken 1963; Gedenkschrift zur hundertsten Wiederkehr des Todestages von Jacob Grimm (Marburg 1963). [ = Hessische Blätter für Volkskunde Vol. 54.] Beiträge zur Sprachwissenschaft, Volkskunde und Literaturforschung, Wolfgang Steinitz zum 60. Geburtstag am 26. Februar 1965 dargebracht (Berlin 1965). Beiträge zur vergleichenden Erzählforschung; Festschrift für W. Anderson zu seinem 70. Geburtstage, ed. by Kurt Ranke (FFC 151-163) (Helsinki 1955). Kontakte und Grenzen; Probleme der Volks-, Kultur- und Sozialforschung. Festschrift für Gerhardt Heilfurth zum 60. Geburtstag (Göttingen 1969). Estudios dedicados a Menéndez Pidal, I-V (Madrid 1953). Angebinde John Meier zum 85. Geburtstag am 14. Juni 1949, ed. by Fr. Maurer (Lahr 1949). 5. Proceedings of Recent

Congresses

Slavic Folklore, a symposium ed. by A. B. Lord (Philadelphia 1956). Internationaler Kongress der Volkserzählungsforscher in Kiel und Kopenhagen (19.8. - 29.8.1959), Vorträge und Referate, ed. by K. Ranke (Berlin 1961). Tagung der International Society for Folk-Narrative Research, Antwerpen (6.-8. September 1962); Bericht und Referate, ed. by K. Ranke (Antwerpen 1963). Tagung der Sagenkommission der International Society for Folk-Narrative Research, in Budapest (14.-16. Oktober 1963), ed. by G. Ortutay (Budapest 1964). IV. International Congress for Folk-Narrative Research in Athens (1.9—6.9.1964), lectures and reports, ed. by G. A. Megas (Athens 1965). Poetics, Poetyka, Poètika, I-II (The Hague-Warszawa 1961-66). Vile Congrès international des sciences anthropologiques et ethnologiques (Moscou 3 août—10 août 1964) [Vol. VI, section 13: "Poème orale populaire"]. Proceedings of VIII International Congress of Anthropological and Ethnological Sciences 1968 (Tokyo and Kyoto) II. Ethnology [Section B-6, Religion and folklore, pp. 266-318].

B. Major Historical Theories and Results (pp 746-49) EHRENREICH,

P. M. A. 1905a. Mythen der südamerikanischen Indianer. Berlin.

794

K. HORÁLEK

J. G. 1919. Folklore in the Old Testament: Studies in comparative religion, legend and law. London. POLÍVKA, J. 1904. Pohádkoslovné studie. Praha.

FRAZER,

C. Reconstruction as an Historical Method (pp. 749-51) ANDERSON,

W. 1934-40. Handwörterbuch des deutschen Märchens 11.508-22.

Berlin. I. 1967. Seal-folk in the East and West. Some comments on fascinating groups of folk tales. Folklore International, essays in traditional literature, belief and customs, in honor of Wayland Debs Hand. Hatboro. BOLTE, J . , and G. POLÍVKA. 1 9 3 2 . Anmerkung zu den Kinder- und Hausmärchen der Brüder Grimm, Vol. III. Leipzig. DUMÉZIL, G . 1 9 5 8 . L'idéologie tripartie des Indo-Européens. Brussels. . 1968. Mythe et épopée.. . . L'idéologie des trois fonctions dans les épopées des peuples indo-européens. Paris. HORÁLEK, K . 1 9 6 2 . Studie o slovanské lidové poezii. Praha. IVANOV, V . V . , and V . N . TOPOROV. 1 9 6 5 . Slavianskiie iazykovyie modeliruiuscie semioticeskie sistemy. Moscow. JAKOBSON, R. 1952. Studies in comparative Slavic metrics. Oxford Slavonic Papers 3. [Reprinted 1968, Selected Writings IV, The Hague.] JENNES, D. 1924. Report of the Canadian Arctic Expedition 1913-1918. Vol. XIII. Eskimo folklore (Northern Alaska, etc.) Ottawa. [German translation 1969: Eskimo Märchen, by H. Bariiske, 90-8, Köln-Düsseldorf.] KARBUSICKÍ, V . 1 9 6 4 . Stfedoveká epika a pocátky ceské hudby. Hudební vëda ANDERSSON,

3.

367-449.

LIUNGMAN, W .

1961.

[See I A . l , p. 7 9 1 . ]

1 9 2 3 . Les origines indo-européennes des mètres grecs. Paris. POP, M. 1967. Aspects actuels des recherches sur la structure des contes. Fabula MEILLET, A .

9.70-7.

K. Kiel.

RANKE,

1955-58.

Schleswig-Holsteinische Volksmärchen (Ath.

300^102).

1 9 2 5 . Dívky - ptáci. Sborník práci vënovany Janu Mácihalovi. Praha. V. N. 1963. Semiótica. Trudy po znakovym sistemam 4.9-43. Tartu. TRUBETSKOY, N. S. 1937. W sprawie wiersza byliny rosyjskiej. Prace ofiarowane Kazimierzowi Wóycickiemu. Vilnus. TILLE, V .

TOPOROV,

D. Folk Narrative (pp. 751-56) 1. General AARNE, A .

Hamina.

1913. Leitfaden der vergleichenden Märchenforschung. FFC 13.

FOLK POETRY

795

and S . THOMPSON. 1961. The types of the folktale. A classification and bibliography. [Antii Aarne's Verzeichnis der Märchentypen, FFC 3, translated and enlarged by Stith Thompson. Second revision, FFC 184.] Helsinki. ANDERSON, W. 1923. Kaiser und Abt; die Geschichte eines Schwanks. FFC 42. Helsinki. ANIKIN, V. P . 1 9 5 9 . Russkaia narodnaia skazka. Moskva. ASADOWSKI, M. K. (See Azadovskii). 1926. Eine sibirische Märchenzählerin. FFC 68. Helsinki. AZADOVSKII, M. K. 1932. Russkaia skazka. [An anthology of storytellers with characteristics and comments] I.—II. Moskva. AZADOVSKII, M. K . and L. E . ELIASOV, eds. 1 9 4 0 . Skazki Maglaia. Leningrad. BARTHES, R. 1 9 5 7 . Mythologies. Paris. BASCOM, W. 1965. The forms of folklore prose narratives. JAF 78.3-20. BEDIER, J. 1893. Les fabliaux. Paris. BEIT, H. VON. 1965. [See Rogues, H. von Beit.] BO§KOVIC-STULLI, M. 1967. Narodna predaja o vladarevoj tajni. Zagreb. BRINKMANN, O. 1933. Das Erzählen in einer Dorfgemeinschaft. Münster. CHRISTIANSEN, R. TH. 1959b. Studies in Irish and Scandinavian folktales. Copenhagen. . 1959a. The migratory legends; a proposed list of types with a systematic catalogue of the Norwegian variants. FCC 175. Helsinki. CISTOV, K . V . 1 9 6 7 . Russkie narodnie sotsial'no utopiceskie legendy. Moskva. DEGH, L . 1 9 6 2 . Märchen, Erzähler und Erzählgemeinschaft, dargestellt an der ungarischen Volksüberlieferung [with an extensive bibliography]. Berlin. DELARGY, J . H . 1 9 4 7 . The Gaelic story-teller with some notes on Gaelic folktales. London. DUNDES, A. 1964. The morphology of North American Indian folktales. FCC 195. Helsinki. . 1965. (See IA.l, p. 790.)

AARNE, A .

FRAZER, J . G .

1919.

(See IB, p . 794.)

. 1953. The Golden Bough; a study in magic and religion. 4th ed. London. GERHARDT, M. I. 1963. The art of story-telling; a literary study of the thousand and one nights. Leiden. GRIMM, W., and J. GRIMM. 1822. Kinder- und Hausmärchen. Vol. 3. [Expanded 1856.] GUNKEL, H. 1921. Das Märchen im Alten Testament. Tübingen. HAIDING, K . 1 9 5 5 . Von der Gebärdensprache der Märchenzähler. FCC 1 5 5 . Handwörterbuch des deutschen Aberglaubens, I-X. Berlin-Leipzig. Handwörterbuch des deutschen Märchens, ed. by J. Bolte and L. Mackensen, I-II. 1930-40. [A-Gyges.] Berlin. HEILFURTH, G. 1967. Bergbau und Bergmann in der deutschsprachigen SagenÜberlieferung Mitteleuropas. Marburg.

796

K. HORALEK

G. 1 9 5 1 . Uberlieferung und Persönlichkeit; die Erzählungen und Lieder des Egbert Gerrits. Münster. HERMES, E. 1964. Die drei Ringe; aus der Frühzeit der Novelle. Göttingen. HEYDEN, F. 1922. Volksmärchenerzähler; zur literarischen Gestaltung des deutschen Volksmärchens. Hamburg. HORALEK, K . 1 9 6 4 . Pohâdkoslovné Studie. Praha. . 1966. Studie ze srovnâvaci folkloristiky. Praha. JACKSON, K . H . 1 9 6 1 . The international populär tale and early Welsh tradition. Cardiff. JECH, J . 1959. Lidovâ vypravëni z Kladska. Praha. JOLLES, A. 1958. Einfache Formen. Legende, Sage, Mythe, Rätsel, Spruch, Kasus, Memorabile, Märchen, Witz. 2. unveränderte Aufl. Halle. KAPELUS, H. 1968. Bajka ludowa v dawnej Polsce. Warszawa. KARNAUKHOVA, I. V. 1941. Skazki I. F. Kovaleva. Moskva. [Collection and study.] KOSKO, M. 1966. Le Fils assassiné (AT 939 A); étude d'une thème légendaire. FFC 198. Helsinki. KROHN, K . 1926. Die folkloristische Arbeitsmethode. Oslo. LEIBFRIED, E. 1 9 6 7 . Fabel. Stuttgart. LÉVI-STRAUSS, CLAUDE. 1 9 5 8 . Anthropologie structurale. Paris. [See "La structure des mythes", p. 227 ff.] . 1962. La pensée sauvage. Paris. . 1964-68. Mythologiques. I: Le cru et le cuit (1964); II: Du miel aux cendres (1966); III: L'origine des manières de table (1968). Paris. LEYEN, F . VON DER. 1 9 5 3 - 5 4 . Die Welt der Märchen, I - I I . Düsseldorf. . 1958. Das Märchen. Ein Versuch. 4. erneuerte Aufl. (with K. Schier). Heidelberg. LINTUR, P. V. 1953. Zakarpatskij skazochnik Andrej Kaiin. Problema traditsii i licnogo nacala v tvorcestve skazochnika. Uzgorod. HENSSEN,

LIUNGMAN, W .

1961.

[See I A . l , p. 7 9 1 . ]

M. 1960. Das europäische Volksmärchen 2. Aufl. Bern. . 1962. Es war einmal; vom Wesen des Volksmärchens. Göttingen. . 1966. Volksmärchen und Volkssage. 2. Aufl. Bern. MATICETOV, M. 1961. Sezgani in prerojeni clovek. Ljubljana. [Der verbrannte und wiedergeborene Mensch.] MELETINSKII, E. M. 1958. Geroï volSebnoï skazki. Proikhozenie obraza. Moskva. NOVIKOV, N . V . 1 9 4 1 . Skazki F . P. Gospodareva. Petrozavodsk. OBENAUER, K . J . 1 9 5 9 . Das Märchen, Dichtung und Deutung. Frankfurt a.M. LÜTHI,

PENZER, N . M . PETZOLDT, L .

Helsinki.

1952.

[See I A . l , p. 7 9 1 . ]

1968. Der Tote als Gast. Volkssage und Exempel. FFC 200.

FOLK POETRY

797

W. P . 1 9 5 8 . Deutsches Volkstum in Märchen und Sage, Schwank und Rätsel. POLÌVKA, J. 1932. Slovanské pohädky. I. Üvod. Vychodoslovanské pohädky. Praha. POMORANTSEVA, E. V. 1952. Skazki Abrama Novopol'tseva. Kujbyäev. . 1965. Stud'by russkoi skazki. Moskva. PROPP, V. J. 1928a. Morfologiia skazki. Izd. 2nd ed. 1969. Moskva. [English translation 1956: Morphology of the folktale. 2nd ed. 1968.] RANKE, K., ed. 1955. Beiträge zur vergleichenden Erzählforschung. FFC 151— 163. Helsinki. . 1965. Einfache Formen. Das Fischer Lexikon, Literatur 2/1. 184-200. Frankfurt a.M. [Another version in Journal of the Folklore Institute 4.17-31, 1967.] ROBERTS, W. E. 1958. The tale of the kind and the unkind girls. Berlin. ROGUES, H . VON BEIT. 1 9 6 5 . Das Märchen, sein Ort in der geistigen Entwicklung. Bern, München. RÖHRICH, L . 1956. Märchen und Wirklichkeit; eine volkskundliche Untersuchung. Wiesbaden. . 1962, 1967. Erzählungen des späten Mittelalters und ihr Weiterleben in Literatur und Volksdichtung bis zur Gegenwart. Sagen, Märchen, Exempel und Schwanke mit einem Kommentar, I—II. Bern, München. . 1963. Märchen und Wirklichkeit. 2. Aufl. ROOTH, A . B . 1 9 5 1 . The Cinderella Cycle. Lund. RUMPF, M. 1951. Rotkäppchen. Eine vergleichende Märchenuntersuchung. Diss. Göttingen. Sagen und ihre Deutung; Beiträge von M. Lüthi, L. Röhrich und G. Fohrer, mit einem Geleitwort von W. E. Peukert. 1965. Göttingen. SAREYKO, H . U . 1 9 5 4 . Das Weltbild eines ostpreussischen Volkserzälers. Marburg. SATKE, A . 1 9 5 8 . Hlucinsky pohädkar Josef Smolka. Ostrava. [Study, collection and commentary.] SCHMIDT, L. 1963. Die Volkserzählung; Märchen, Sage, Legende, Schwank. Berlin. SIMONIS, YVAN. 1 9 6 8 . Claude Lévi-Strauss ou la 'passion de l'inceste'; introduction au structuralisme. Paris. SORIANO, M . 1 9 6 8 . Les Contes de Perrault, culture savante et traditions populates. Paris. STENDER-PETERSEN, A. 1 9 3 4 . Die Varägersage als Quelle der altrussischen Chronik. Aarhus-Leipzig. . 1953. Varangica. Aarhus. STRASSNER, E . 1 9 6 8 . Schwank. Stuttgart. SWAHN, J. ö . 1955. The tale of Cupid and Psyche. Lund. SYDOW, C. W. VON. 1948. Selected papers on folklore, published on the occasion of his 70th birthday. Copenhagen. PEUKERT,

798

K. HORÁLEK

J. 1957. Irisches Erzählgut im Abendland; Studien zur vergleichenden Volkskunde und Mittelalterforschung. Berlin. THOMPSON, S. 1 9 1 9 . European tales among the North American Indians; a study in the migration of folk-tales. Colorado Springs. . 1946 [SeeIA.l,p. 791.] WESSELSKI, A. 1931. Versuch einer Theorie des Märchens. Reichenberg. WOSIEN, M. G . 1 9 6 9 . The Russian folk-tale, some structural and thematic aspects. München. ZENKER, M., and E. STARZACHER. 1 9 4 1 . Eine deutsche Märchenerzählerin aus Ungarn. München.

SZÖVERFFY,

2. Folklore diffusion and methodology of historical investigation 1 9 3 4 - 4 0 . Geographisch-Methode. Handwörterbuch des deutschen Märchens II, pp. 5 0 8 - 2 2 . Berlin. The anthropologist looks at myth. 1966. Compiled by Melville Jacobs, ed. by John Greenway. JAF 79, No. 311. BOLTE, J . , and G. POLÍVKA. 1 9 3 2 . Anmerkungen zu den Kinder- und Hausmärchen der Brüder Grimm, V. Leipzig. [Theorien über Entstehung und Verbreitung der Märchen, pp. 2 3 9 - 6 4 . ] EHRENREICH, P. M. A. 1905b. Die Mythen und Legenden der südamerikanischen Urvölker. ZEthn 37, Supplement. HORÁLEK, K . 1969a. Märchen aus Tausend und einer Nacht bei den Slaven. Fabula 10.155-95. . 1969b. Asiatische Mythen in Europa. Zur Frage der Vermittlerrolle der Magyaren. Folklórica Pragensia 1.125-72. KRAPPE, A . H . 1 9 4 9 . Folklore and mythology. Funk and Wagnalls standard dictionary of folklore, mythology, and legend, 1 . 4 0 3 - 7 . New York. ANDERSON, W .

KROHN, K .

1926.

[See I D . l , p . 7 9 6 . ]

F. VON DER. 1953-54. [See ID.l, p. 796.] . 1958. [See ID.l, p. 796.] SPIES, O. 1952. Orientalische Stoffe in den Kinder- und Hausmärchen der Brüder Grimm. Walldorf-Hessen. VESELOVSKI!, A. N. 1 9 4 0 . Istoriceskaia poetika, 50Iff. Leningrad. VRIES, JAN DE. 1954. Betrachtungen zum Märchen, besonders in seinem Verhältnis zu Heldensage und Mythos. FFC 150. Helsinki. ZIRMUNSKII, V. M. 1961. Vergleichende Epenforschung, I. Berlin. . 1962. Narodnyi geroiceskii epos. Sravnitelno istoriceskie ocerki. Leningrad. LEYEN,

E. Folk Song (pp. 756-64) 1. General and musicological BOWRA, C. M. 1962. Primitive song. London.

FOLK POETRY

799

1 9 5 3 . Historické piesne. Bratislava. [With extensive commentary.] W. 1966. Das Volkslied im Abendland. Bern, München. DEAN-SMITH, M. 1954. A guide to English folk song collections, 1822-1952. Liverpool. KATONA, I. 1964. Historische Schichten der ungarischen Volksdichtung. FFC 194. Helsinki.

BRTÀN, R .

DANCKERT,

KRAPPE, A . H .

1930.

[See I A . l , p. 790.]

M. 1958. Die Volksepik der Albaner. Halle. 1949. Griechische Volksdichtung. Archiv für Literatur und Volksdichtung 1.196-250. Lahr. [See pp. 251-4 for comparative notes by E. Seemann.] SALMEN, W. 1960. Der fahrende Musiker im europäischen Mittelalter. Kassel. SEEMANN, E. 1949. Wolfdietrichepos und Volksballade; ein Beitrag zur Geschichte der mittelalterlichen Balladendichtung. Archiv für Literatur und Volksdichtung 1.119-76. Lahr. . 1951. Deutsch-litauische Volksliedbeziehungen. Jahrbuch für Volksliedforschung 8.142-211. SEEMANN, E . , and W . WIORA. 1 9 6 0 . Volkslied. Deutsche Philologie im Aufriss, ed. by W. Stammler, II. col. 349-96. 2nd ed. Berlin. Sowjetische Volkslied- und Volksmusikforschung. Ausgewählte Studien, ed. by E. Stockmann and H. Strobach, in collaboration with K. Cistov and E. Hippius. Berlin. SUPPAN, W. 1966. Volkslied; eine Sammlung und Erforschung. Stuttgart. SYDOW, A. 1962. Das Lied. Ursprung, Wesen, Wandel. Göttingen. VÄCLAVEK, B., and R. SMETANA. 1950. O ceské pisni lidové a zlidovélé. Praha. WILGS, D . K . 1 9 5 9 . Anglo-American folksong scholarship since 1 8 9 8 . New Brunswick. WIORA, W . 1953. Europäischer Volksgesang; gemeinsame Formen in charakteristischen Abwandlungen. Köln.

LAMBERTZ,

LÜDEKE, H .

2. Heroic epic

Russkiì bogatyrskiì èpos. Moskva. M. 1938-51. Byliny. Itogi i problemy inzucenija. BRAUN, M . 1 9 6 1 . Das serbokroatische Heldenlied. Göttingen. BURKHART, D. 1968. Untersuchungen zur Stratiographie und Chronologie der südslavischen Volksepik. München. CHADWICK, N. K., and V. 2IRMUNSKY. 1969. Oral epics of Central Asia. Cambridge. CHIKOVANI, M. IA. 1 9 6 6 . Narodnyj gruzinskij epos o prikovannom Amirani. Moskva. JAKOBSON, R. 1966. Slavic epic studies. Selected Writings IV. The Hague-Paris. KNAPPERT, J. 1 9 6 7 . The epic in Africa. Journal of the Folklore Institute 4 . 1 7 1 - 9 0 . ANIKIN, V . P .

ASTAKHOVA, A .

1964.

800

K. HORALEK

KOZIN, S . A . LAMBERTZ,

Èpos mongol'skikh narodov. Moskva. M. 1958. Die Volkscpik der Albaner. Halle. [Moskva-Leningrad, 1948.

1966.] E. M . 1963. Proizkhozdenie geroiceskogo eposa; rannie formy i arkhaiceskie pamiatniki. Moskva. ORBELI, I . A. 1 9 5 6 . Armianskij geroiceskii èpos. Erevan. PADLOV, V . 1 9 6 8 . Manas, geroiceskij èpos kirgizskogo narodna. Frunze. Philologiae Turcicae Fundamenta II. 1964. Wiesbaden. (See L'épopée et la Hikâye, by P. N. Boratav, pp. 411-40). REES, A. P., and B. REES. 1961. Celtic heritage; ancient tradition in Ireland and Wales. London. Skazaniia a nartakh — èpos narodov Kavkaza. 1969. Moskva. STEIN, R.-A. 1959. Recherches sur l'épopée et le barde au Tibet. Paris. [Includes items concerning Mongolian epics.] TEODOROV, E. K. 1963. Sushtnost i proizkhod na bulgarskite iunaski i khaidushki pesni. . . Sofia. VRIES, J . DE. 1 9 6 1 . Heldenlied und Heldensage. Bern. WAIS, K. 1958. An der Grenzen der Nationalliteraturen; vergleichende Aufsätze. Berlin. [Pp. 341-405: Zur Berührung der altorientalischen und europäischen Erzählungsdichtung: Ullikumini, Hrungnir, Armilu und Verwandte.] MELETINSKII,

3.

Folk

Narrative

Song

Editions

Bälgarsko narodno tvorcestvo. 1961-63. Vols. I-XII. Sofia. [I: Janaski pesni, II: Khajduski pesni, III: Istoriceski pesni, IV: Miticeski pesni. 1961. With commentary.] Bengâlské lidové balady, ed. and tr. by D. Zbavitel. 1956. Praha. BURLASOVÂ, S. 1969. L'udové balady na Horehroni. Bratislava. (With commentary.) CERNYSEV, V. I. ed. 1936. Russkaia balada. Leningrad. (Includes study by N. P. Andreev.) CZERNIK, S. 1958. Polska epika ludova. Wroclaw-Krakow. Deutsches Volksliederarchiv. Deutsche Volkslieder mit ihren Melodien... ed. by J. Meier et al. 1935-67. Berlin-Leipzig. Deutsche Volkslieder; Texte und Melodien, ed. by L. Röhrich and R. W. Brednich. 1965-67. I: Erzählende Lieder. Düsseldorf. ENTWHISTLE, W. J. 1939. European balladry. Oxford. Epiceskieska skazanija narodov iuzhnogo Kitaia, ed. and tr. by B. B. Vakhtin and R. F. Itsa. 1956. Moskva. FRIEDMAN, A . B . 1 9 6 1 . The ballad revival: Studies in the influence of popular and sophisticated poetry. Chicago. HORÂK, J . 1 9 5 8 . Slovenské l'udové balady. Bratislava. KALANDADZE, G. A. 1965. Gruzinskaia narodnoia ballada. Tbilisi.

FOLK POETRY

801

B. P . 1965. Ukrainskii narodyi èpos. Moskva. LAWS, G. M. 1964. Native American balladry, a descriptive study and a biographical syllabus. Rev. ed. Philadelphia. LEACH, M. 1965. The ballad book. Rev. ed. London. PUTILOV, B . N . 1 9 6 5 . Slavianskaia istoriceskaia bailada. Moskva. VRABIE, G. 1966. Balada populara romàna. Bucuresti. ZGANEC, V., ed. 1950. Hrvatske narodne pjesme kajkavske. Zagreb. KIRDAN,

4. Slavic ballads and Heroic Song M. P. 1 9 6 4 . Baladni motivi v narodnata poeziia I . [Pesenta za delba na dvama bratia.] Sofia. GRAFENAUER, I. 1 9 4 3 . Lepa Vida, Studija o izvoru, razvoju in razkroju narodne balade o Lepo Vidi. Ljubljana. . 1966. Slovensko-hrvaska ljudska pesem Marija in Brodnik. Ljubljana. KEMPPINEN, I. 1954. The ballad of Lady Isabel and the False Knight. Helsinki. KUMER, Z. 1963. Balada o nevesti detomorilki. Ljubljana. MEDENICA, R. 1965. Banovic Strahinja. Beograd. MURKO, M . 1937. Das Original von Goethes 'Klanggesang von der edlen Frau des Asan Aga' (Asanaginica) in der Literatur und im Volksmunde durch 150 Jahre. Brunn-Prag-Leipzig-Wien. NYGARD, H. O. 1958. The ballad of Heer Halewijn; its forms and variations in Western Europe; a study of the history and nature of a ballad tradition. Helsinki. POHL, E. 1934. Die deutsche Volksballade von der 'Losgekauften'; ein Versuch zur Erforschung des Ursprungs und Werdeganges einer Volksballade von europäischen Vertretung. FFC 105. Helsinki. [German folk ballads compared, Slavic parallels considered only a propos older collections.] VARGYAS, L. 1967. The survival of the heroic epic of the Hungarian Conquest period in our ballads. Researches into the mediaeval history of the folk ballad, 129-56. Budapest. [See The enticed wife — Child 4.] ARNAUDOV,

5. Folk song poetics R. 1958. Ob-Ugric metrics; the metrical structure of Ostyak and Vogul folkpoetry. Helsinki. BRAILOIU, G . 1 9 5 6 . Le vers populaire roumain chanté. Paris. BREDNIKOV, V . M . , and E . A . TUDOROVSKAJA. 1 9 4 5 . Poètika marijskix narodnyx pesen. Joskav-Ola. CIOBANU, G. 1969. La structure du système de versification populaire roumaine, sa relation avec la versification latine. VII Congrès international des sciences anthropologiques et ethnologiques, 1964, 6.303-8. Moskva. DINEKOV, P. N. 1 9 6 1 . Nekotorye osobennosti poètiki sovremennoj narodnoj pesni. Poetics-Poetyka-Poètika, 4 8 7 - 9 9 . Warszawa and The Hague. GÖDEKE, H . 1 9 6 9 . Motivübergänge vom mythologisch-astralen zum pflanzlichen

AUSTERLITZ,

802

K. HORÄLEK

Bereich im balkanslavischen lyrischen Volkslied. Diss. Hamburg. [From an historical perspective.] HORALEK, K. 1962. Studie o slovanske lidove poezii. Praha. JONES, J . H . 1 9 6 1 . Commonplace and memorization in the oral tradition of the English and Scottish popular ballads. J A F 7 4 . 9 7 - 1 1 2 . LORD, A. B. 1960. The singer of tales. Cambridge. PEUKERT, H . 1961a. Serbokroatische und makedonische Volkslyrik. Gestaltuntersuchungen. Berlin. . 1961b. Die Funktion der Formel im Volkslied. Poetics-Poetyka-Poetika, 487-99. Warszawa and The Hague. POLLOK, K. H. 1964. Studien zur Poetik und Komposition des balkanslavischen lyrischen Volksliedes. I. Das Liebeslied Göttingen. SIDEL'NIKOV, V. M. 1959. Poetika russkoi narodnoi liriki. Moskva. STEINITZ, W. 1934. Der Parallelismus in der finnisch-karelischen Volksdichtung, untersucht an den Liedern des karelischen Sängers Arhippa Perttunen. FFC 115. Helsinki. TROST, P. 1 9 6 1 . Das Metram der litauischen Volkslieder. Poetics-Poetyka-Poetika, 1 1 9 - 2 5 . Warszawa and The Hague. UKHOV, P. D. 1970. Atributsii russkikh bylin. Moskva.

F. Proverbs and Riddles (pp. 764-68) 1918-20. Vergleichende Rätselforschungen I - I I I . FFC 26, 27, 28. Copenhagen. Bälgarski narodno tvorcestvo, 1961-62. [See IE.3, p. 800.] [Vol. 12: Poslovici, pogovorki, gatanki, ed. by Cv. Minkov, 1963.] AARNE, A .

BOGATYREV, P . G . , e d . 1 9 4 3 , 1 9 5 6 .

[See I A . l , p. 7 9 0 . ]

Clareti Enigmata. The Latin riddles of Claret, ed. with introduction by Frederic Peacly. 1957. Berkeley-Los Angeles. FLAJSHANS, V . 1 9 1 1 - 1 3 . Ceska prislovi; sbirka prislovi prupovedia po rekadel lidu ceskeho v cechäch, na Morave av slezsku. Praha. GEORGES, R . A . , and A . DUNDES. 1 9 6 3 . Toward a structural definition of the riddle. J A F 7 6 . 1 1 1 - 1 8 . GORDON, E. I . 1 9 5 9 . Sumerian proverbs; glimpses of everyday life in ancient Mesopotamia. Philadelphia. GOTTSCHALK, W . 1 9 3 5 - 3 8 . Die bildhaften Sprichwörter der Romanen. I - I I I . Heidelberg. HAIN, M. 1951. Sprichwort und Volkssprache; eine volkskundlich-soziologische Dorfuntersuchung. Giessen. . 1966. Rätsel. Stuttgart. KRUMBACHER, K . 1895. Geschichte der byzantinischen Literatur. Handbuch der

FOLK POETRY

803

klassischen Altertumswissenschaft, ed. by I. P. E. von Müller, 9/1.527-1453. 1 9 6 0 . M^drej glowie dose dwie slowie, 1 - 2 2 . Warszawa. . 1965. [SeeIA.l,p. 790.] Kuusi, MATTI. 1957. Regen bei Sonnenschein; zur Weltgeschichte einer Redensart. FFC 171. Helsinki. . 1960. Kurt Ranke. Fabula 3.313-15. LEACH, E. M., ed. 1949-50. [See IA.l, p. 8 4 . ] MANSIKKA, V. J . 1 9 0 9 . Über russische Zauberformeln, mit Berücksichtigung der Blut- und Verrennungssegen. Helsinki. NEUMANN, S. 1 9 6 8 . Sagwörter im Schwank, Schwankstoffe im Sagwort. Volksüberlieferung. Festschrift für Kurt Ranke zur Vollendung des 60. Lebensjahres. 2 4 9 - 6 6 . Göttingen. NOVIKOVA, A. M., and A. V. KOKOREV, ed. 1969. Russkoe narodnoe poèticeskoe tvorcestvo. Moscow. [Bibliography, pp. 138-9, 154.] PANZER, F. 1934. Das Volksrätsel. Die deutsche Volkskunde, ed. by A. Spamer, 1.263-82. SADNIK, L. 1953. Sudosteuropäische Rätselstudien. Graz-Koln. SINGER, S . 1944-47. Sprichwörter des Mittelalters, I—III. Bern. SMITH, W. G. 1948. The Oxford dictionary of English proverbs. Rev. ed. by P. Harvey. Oxford. STEPHENS, T. A. 1930. Proverb literature; a bibliography of works relating to proverbs, ed. by Wilfril Bonser. London. TAYLOR, A. 1931. The proverb. Cambridge, Mass. [Reprinted 1962.] . 1943. The riddle. California Folklore Quarterly 2.129-47. . 1948. The literary riddle before 1600. Berkeley. . 1958. Proverbs and proverbial phrases in the writings of Mary N. Murfree (Ch. E. Craddock). Tennessee Folklore Society Bulletin 24.11-50. WALTHER, H . 1 9 6 3 , 1 9 6 6 . Proverbia sententiaeque latinitatis medii aevi; lateinische Sprichwörter und Sentenzen des Mittelalters in alphabetischer Anordnung. Vol. I: 1963, Vol. IV: 1966. Göttingen. KRZYZANOWSKI, J .

G. Folk Drama (pp. 768-74) D. 1959. Proiskhozhdenie teatra. Leningrad. 1 9 5 3 . Russkaia narodnaia drama X V I I - X X vekov. Moskva. BOGATYREV, P. G . 1 9 4 0 . Lidové divadlo ceské a slovenské. Praha. CHADWICK, H. M., and N. K. CHADWICK. 1940. The growth of literature III. Cambridge. [Part II: The oral literature of Polynesia, pp. 352-91: Dramatic and ritual poetry.] CUBELIp

(2)

since if /?>h2 it is impossible to express a certain idea in the given poetic form. Thus the inequality relation (2) is the necessary condition of any effective poetic creation understood as the distribution of expenses of entropy in the limits of the given language (Ivanov 1973). The 'anagrammatic capacity' of the text in Lévi-Strauss's sense is created by the equivalence (1) and inequality relation (2) and by the possibility of augmenting h2. The lower limit of h2 is defined by the number of all meaning-preserving synonymic transformations of phrases as studied in modern generative semantics. But if the quantity of formal constraints (Ì (e.g. of the anagrammatic repetitions used in the text) is augmented, then it is necessary to augment h2. It can be done because most of the words in the language may be used in the new meaning (referring to some other denotata) with the only exception: words belonging to the same semantic field can not be mixed up (thus a violin and a piano should be kept apart, but it is possible to use violin metaphorically as a designation of a loving human being as in Annensky's poem). From these elementary relations one might deduce the correlation between the intricacy of the form, for instance, of Dante's poems, and their metaphoric style, on the one hand, between Brecht's vers libre and his everyday vocabulary, on the other hand. To return to the object of Saussure's studies, the metaphoric style of Vedic hymns and Old Germanic verses may be deduced from the formal laws of the structure of these texts composed according to anagrammatic Indo-European traditions. The Vedic kavis were able to compose hymns based on anagrammatic principles because they used most of the words in metaphoric meanings. 1.6 Kolmogorov's studies of the informational capacity of the language and of what was called by Lévi-Strauss its 'anagrammatic capacity' have made it possible to distinguish between the statistical definition of the information transmitted in the given message and the non-statistical definition of information that is necessary in mathematical poetics.

840

VYACHESLAV V. IVANOV

It is possible to speak about the statistical aggregate of English telegrams that could help to define statistically the quantity of information contained in the given message. But it is no use applying the same method let us say to War and Peace since this text does not belong to an aggregate or a set of similar texts: the number of the elements of the set is 1 in such a case. Many objections against the use of exact methods in poetics are caused by the fact that these two cases were not kept apart. As Kolmogorov has pointed out, in mathematical poetics and linguistics one should deal with the concept of information applied to individual objects and not to statistical aggregates (Kolmogorov 1965). Then the main notion is that of a conditioned entropy H (x/y) of the object x if the object y is given. The information contained in the object y in its reference to the object x is defined by subtraction I(x/y)

= H(x)-H(x/y)

(3)

As an explanatory example the information contained in the grammatical rules and vocabulary in reference to the text may be mentioned. The entropy H (x/y) is the minimum length of the program P which helps to build up the object x if the object y is given: H (x/y) = min I (P) (4) From this point of view the probabilistic approach is necessary only if H(x)^l(x)

(5)

Such long sequences can not be defined by a program that is shorter than the length of the whole sequence. The absence of the law that could have helped to build a shorter program is called the chance (Kolmogorov 1969). This may be understood as an explanation of the image of the God playing the dice (to use Einstein's words) according to the stochastic model of the universe. Since linguistics and poetics often deal with such extremely complex programs and objects the approach to them becomes stochastic, although this does not mean that these objects cease to be individual in the above given sense. The probabilistic approach might be used to show the extremely rare and exceptional character of certain poetical compositions. Thus Kolmogorov succeeded in showing that the repetition of the same rhythmical pattern in several contiguous iambic lines of Pushkin's poem included in 'Egyptian nights' by the same author is by no means accidental. As a model of the artistic creation seen in the light of the successive expenditure of entropy, an ideal automaton might be built that should choose the preferable text among the astronomically great set of all the possible texts of a given length. It can be supposed that the extremely large work consisting of selecting all the possible variants might be reduced in part because of the existence of the multitude

THE THEORETICAL FRAMEWORK OF MODERN POETICS

841

of authors who are preparing the coming into the world of the great writer. The selection is made by the whole society reading its writers. That is why the obstacles that may prevent the society from making this choice may stop literary development. The evolution of the complex structures in stochastic poetics, as in modern biology (Monod 1970: 128-129), is seen from the point of view of chance. The idea of selection is connected directly with the breaking of tradition due to inaccurate transmission of the original pattern. Literary genres might be considered as equivalent to the stabilized species. According to Bachtin's apt formula, 'the genre is a representative of creative memory in the process of literary development' (Bachtin 1963: 142). This idea can be understood in accordance with the cybernetic concept of memory forming a part of the communication channel. The literary context is formed by a synchronic literary situation and the 'row of development' (Miko 1970: 121, where Tynyanov's ideas of literary evolution are combined with the Czech and Slovak theories).

2. MORPHOLOGY AND TRANSFORMATIONS

2.1 The first attempt to approach the work of verbal art from the point of view of morphological classification was made in the Indian treatises Natyaiastra and DaSarupa. In these works the traditional theory of the dramatic form (riipa 'form' from which the name of classical Indian drama rupaka is derived) is explained. The main feature of the theory, like that of Propp, is its syntagmatic character. It is constructed according to the linear order of the motives of the action (arthaprakrti), stages of the action (avastha) and connections (samdhi; the same term was used in the Indian grammatical treatises). The concept of a connection was the central one in the Indian theory (Erman 1961: 50), five main connections being used to distinguish five main genres of dramatic works: the combination of the first and last connections — the knot (mukha) and the final one (nirvahara) — and the absence of the three intermediary connections is characteristic of the oneact plays such as those belonging to the genre of the farce (prahasara). Such a definition of a genre coincides exactly with the scheme of description characteristic of modern linguistics and poetics. The whole syntagmatic scheme is realized only in some cases, while other objects may be described as having the incomplete set of all the possible elements. The same syntagmatic approach, similar to the study of distribution in taxonomic descriptive linguistics, is seen in the rules of the compatibility of different types of characters: for instance, the woman of a 'common' type (sadharanastri, a courtesan) is incompatible with a hero (netr), a king (raja) or a god (deva). The same approach to the scheme of a dramatic work has been developed recently by Solomon Marcus who has shown that the study of combinations of a given character with all other personages might give the exact cue to differentiation of all the heroes in a play (Marcus 1970); a similar idea was

842

VYACHESLAV V. IVANOV

suggested more than thirty years ago by Iarho in his treatise on mathematical poetics that remains unpublished (Gasparov 1969); a half-formalized approach to the distribution of characters in a play was present already in Baluhatyi's study (Baluhatyi 1927) published simultaneously with Propp's first monograph. 2.2 Propp's ideas on morphology were the direct continuation of Goethe's conception of morphology. As mottoes to different chapters of his Morphology of the fairy-tale, several quotations from Goethe's works on morphology were taken by Propp. In this respect Propp was ahead of his time even more than in those technical schemes that have become fashionable nowadays. Only now after Goethe's scientific genius has been fully appreciated by the physicists (Heisenberg 1967) do biologists pay attention to the importance of his morphology in which he had anticipated the modern ideas of the unity of all the organic world: Il y avait une ambition 'platonicienne' dans la recherche systématique des invariants anatomiques à laquelle consacrèrent les grands naturalistes du XIX e siècle après Cuvier (et Goethe). Peut-être les biologistes modernes ne rendent-ils pas toujours justice au génie des hommes qui, sous la stupéfiante variété des morphologies et des modes de vie des êtres vivants, ont su reconnaître si non une "forme" unique, du moins un nombre fini de plans anatomiques, chacun d'entre eux invariant au sein du groupe qu'il caractérise (Monod 1970: 117).

Instead of searching for similitude between Goethe's morphology and the theory of Darwin, who himself had called Goethe a path-maker (Viëtor 1950: 32), scientists of the present day can find in Goethe's morphology the first approach to the understanding of the discrete transformations that help to describe the invariant preserved under all transformations. Goethe saw such an invariant in his 'primal plant' (Urpflanze) to which Propp compared his own scheme of the fairy-tale. This conceptual link between Goethe's morphology and modern structural approach to the composition might be seen also in Lévi-Strauss's last monograph. The results of his studies on transformation of American Indian myths Lévi-Strauss compares to the important results of D'Arcy Wentworth Thompson's work in which one might see the exact continuation of the thought of Goethe who had suggested that different parts of the organism might be arranged in the general scheme 'nach gewissen Zahlen und Maßen' (Goethe 1963: 43). As Lévi-Strauss himself puts it, chaque changement discontinu impose la réorganisation de l'ensemble: il ne se produit jamais seul, mais en corrélation avec d'autres changements. En ce sens, on peut dire que l'analyse mythique est symétrique et inverse de l'analyse statistique: elle substitue la riguer qualitative à la rigueur quantitative, mais l'une et l'autre ne peuvent prétendre à la rigueur que parce qu'elles disposent d'une multiplicité de cas qui manifestent la même tendance à s'organiser spontanément dans l'espace et le temps. Ce qui précède aide à comprendre pourquoi les speculations de Dürer dans ses Livres . . . des portraits des corps humains, celles de Goethe dans la Métamorphose des plantes, reprises et généralisées par D'Arcy Wentworth Thompson qui leur a donné un statut scientifique, conservent aujourd'huis leur portée (Lévi-Strauss 1971: 604).

THE THEORETICAL FRAMEWORK OF MODERN POETICS

843

To Propp's morphological primal fairy-tale (Urmarchen) corresponds the morphologically primal myth at which Lévi-Strauss arrives in his study: 'aucun mythe n'est semblable. Pourtant, pris dans leur ensemble, ils reviennent tout à la même chose, et comme Goethe l'affirme des plantes, "leur choeur guide vers une loi cachée'" (Lévi-Strauss 1971: 620). The scheme reconstructed as a final result of morphological classification of myths has at the same time a historical and areal reality. Thus Lévi-Strauss's 'mythe de référence' (Mi, M 7 - 1 2 ) — 'dénicheur d'oiseaux' has been established as the most archaic myth of the Southern and Northern American Indian mythologies and at the same time has important Eurasian correspondences in the Ket myth with the same structure. The use of the transformations to find out the invariant that might be seen as the oldest source of all the transforms is similar to the procedures of modern linguistics where morphonemic internal reconstruction is rendered possible by transformational grammar 2.3 Propp himself was aware of the evolutionary implications of his morphological approach when he wrote about the theory according to which 'morphological similarity does indeed result from a known genetic tie — the theory of differentiation owing to subsequent metamorphoses or transformations of varying cause and occurence' (Propp 1971: 94). Comparison between organic evolution and the history of fairy-tales serves to establish the main laws of development (Propp 1971: 95). The same morphological approach based on the similarities between comparative anatomy and the typological study of cultural phenomena may be found in Hocart's works in which one may see the most interesting example of a pre-structuralist attempt to understand the evolution of social facts. It is by no means accidental that Hocart's main book, first published in 1936, has the subtitle An essay in the comparative anatomy of human society (Hocart 1970a). He begins this work by analyzing the methods of comparative anatomy which was one of the main objects of Goethe's studies; at the same time he quotes Cuvier. Just as the scientists of the Russian formalist sohool (called 'morphological' in the early writings of its representatives) and of the German morphological school in literary criticism, Hocart may be called a morphologist. His similarity to Propp is revealed in the most striking manner in his book on kingship and in the short essay "Saviours" (first published in 1936) in which he analyzes the plot of "Popul Vuh" according to Lord Raglan's pattern of a hero. After a formal analysis Hocart suggests a sociological interpretation of the pattern according to which (in the oldest form of it) the hero slays and defeats not his father (as in the Oedipus myth analyzed by Lord Raglan) but his mother's brother. This deduction is confirmed by the rules of diachronic typology according to which there is only one possible way of development: 'uncles' > 'fathers' but not vice versa; the validity of Hocart's reconstruction is proved by the recent studies of Indo-European kinship terms (Gates 1971: 52, 56, 61, 65). This methodologically faultless reconstruction is used by Hocart to illustrate the need of a comparative approach to history:

844

VYACHESLAV V. IVANOV

The point is important, because the psycho-analysts have made great capital out of the Oedipus myth. It would serve them right if it turned out that originally it was not the father but the mother's brother who was killed, for they would rush into history without ever consulting the comparative historian (Hocart 1970b: 148).

2.4 This polemic remark is quite to the point since Freud's theory in its first stages was a direct continuation of those evolutionary tendencies of the humanitarian sciences of the 19th century that had no relation to comparative morphology. That is the reason why not the orthodox Freudian approach but 'the archetypal criticism' (Frye 1969: 131) proved to be important for poetics. Although Frye was influenced by Jung's concept of archetype, he did not try to apply it directly to literature. His starting point (as that of Bachtin in his epoch-making studies) was rather the traditional theory of genres based on Aristotle. As Frye himself said, Aristotle's Poetics 'remain as good an introduction to the subject as ever' (Frye 1969: 14). It is a rewarding task to compare Bachtin's studies (published twenty years after they had been finished) with Frye's book. Some results of their work are identical, particularly the comparison of the work of such authors as Rabelais with Menippean satire and the concept of temporal cycles (Hayman 1973: 82). Both Bachtin and Frye have undergone a strong influence of Freudian (and Jungian) writings. Bachtin was the author of the best Russian book on Freud, published in 1927 under the name of his pupil (Voloshinov 1927), in which he tried to give what might be called a semiotic reinterpretation of psychoanalysis similar to the attempts of Sapir, Benveniste, Lacan, Shands. In this early work Bachtin had introduced already the distinction of the official and unofficial conscience which was later developed in his study on unofficial medieval culture and Rabelais. For instance, his deep remarks on the unpublished spheres of the vocabulary (Bachtin 1965a: 459) are close to the theory suggested recently by a prominent anthropologist (Leach 1964: 38): the ambivalent character of the meaning of such words can be explained by the neutralization of binary opposition between the 'self' (T — the subject) and the universe. Bachtin's work anticipated the idea according to which Rabelais' world contained the archaic system of binary oppositions (Paris 1970). Another important case of neutralization found in the popular unofficial culture by Bachtin was also confirmed by recent anthropological studies (Leach 1961: 135): unofficial behaviour during the carnival fests was characterized by the inversion of social roles. The structural analysis of this feature of archaic rituals was given already in O. M. Freudenberg's early work (written in 1925) where the origin of different genres was studied in the light of comparative ethnology (Freudenberg 1936), similar to the archetypal criticism of Frye. The same problem was dealt with in the yet unpublished works of Eisenstein where he analyzed the inversion of the roles of the king and his subject and of a man and a woman in his own film "Ivan the Terrible" (Ivanov 1970). His analysis is particularly close to that of modern structural anthropology (Turner 1969).

THE THEORETICAL FRAMEWORK OF MODERN POETICS

845

2.5 Bachtin, Eisenstein, Freudenberg (whose work was appreciated by both the former), and Propp in his later work might be considered as structuralists in the sense in which Lévi-Strauss in his well-known article on Propp's Morphology of the fairy-tale understands structuralism as opposed to formalism (Lévi-Strauss 1960). As for Propp's works of that later period one should particularly stress the importance of his book on the historical origin of the fairy-tale (Propp 1946). As Freudenberg and Bachtin, Propp tried to find out the historical explanation of the genre in the light of the ethnological evidence. He thought that the contents of the pattern of the fairy-tale can be reconstructed in the light of rites of initiation. Again the analogy with Hocart's work seems striking. After establishing the exactly determined linear order of the coronation rites Hocart sought its origins in an invariant set of rituals common to the initiation ritual as well (Hocart 1969). The concrete rituals were derived from that primal rite. Thus after a purely morphological work on the syntagmatic linear pattern its sociological or ethnological interpretation is attempted. Many details in Propp's book on the diachronic study of the fairy-tale have proved to be particularly revealing in the light of the comparative evidence that was found after he had finished his work. As an example one might cite his analysis of Gogol's Viy whose mysterious character was later on stressed in a special study by Driessen (Driessen 1965: 152 and ff.). According to Propp's explanation, 'in Gogol's "Viy" the devils do not see the Cossack. Those devils who might see the living persons are like shamans among the devils because the only living human beings who can see the dead are shamans: average humans do not see the dead. The devils call for such a shaman. He is Viy' (Propp 1946: 60). Typological verification of this explanation is given by the myths of the American Indians where a terrible giant or spirits of the darkness are characterized as having their eyes shut (Lévi-Strauss 1971: 267 and 274). Propp's ideas were in part provoked by Potebnya's brilliant analysis of a Common Slavic children's game in which a special personage appears whom Potebnya interpreted as a blind representative of death (Potebnya 1865: 92-94). Here as in many other genial conjectures Potebnya is close to modern semioticians, who are discovering again and again the proofs of the ritual origin of such children's games. Developing Potebnya's analysis Propp established the fact of the 'reciprocity of the blindness' which helps to explain the image of Viy. Potebnya's idea may be supported by Siberian mythological parallels. The importance of mythological archetypes in Gogol's work is proved by other facts (Ivanov 1971). Among other important discoveries of Propp one should highly evaluate his article on ritual laughter (Propp 1939); Propp's last book on the theory of laughter is yet unpublished. Some of the ideas of the article were developed in the most suggestive way by Roman Jakobson in his remarkable study of the Old Czech play 'Unguentarius' (Jakobson 1958). Both the articles deal with the same problem as that studied in Bachtin's book on Rabelais, where the author discusses the aspects

846

VYACHESLAV V. IVANOV

of popular culture connected with the laughter. The results of Bachtin's approach to 'laugh doubles' (smekhovyie dublyory in Russian) are quite similar to that of Benveniste's theory according to which the game is an 'inverted image' of the sacral rite (Benveniste 1947). Unfortunately the results of these works and of the most important Eisenstein study (1966) on the theory of laughter are not known sufficiently as can be seen from the absence of all these writings in the bibliography of the latest article on "Homo Ridens" (Milner 1972). As yet it is difficult to write the complete history of this almost unknown trend in the history of the poetics of the grotesque and comic genres since not only Propp's posthumous book on laughter but, also important, Freudenberg's writings of the later period remain unpublished. Freudenberg used the development of the tragedy leading to transformation into comedy to illustrate her conception of evolution, having nothing to do with the ideas of the 19th century: 'evolution' here meaning transformation into something opposite (one may cite as another example the transformation of the myths on eysight into myths of blindness studied in an important unpublished book on Logics of the antique myth by I. E. Golosovker). As an interesting analogue of this conception one may refer to Roman Jakobson's analysis of the tragical and grotesque elements in Mayakovsky's works. To demonstrate the inner connections between different variants of the morphological approach it is worth quoting Hocart's ideas on the role of the grotesque elements in the older forms of the rites. As Hocart has suggested, these elements had been very early omitted from the rituals of coronation and consecration but they have been preserved and even amplified in the rituals of initiation because of their popular character (Hocart 1969: 158). Thus the popular trend in ritual practice is seen from the point of view similar to that of Bachtin's study of the carnival. In Eisenstein's commentary on his own unfinished film ¡Que viva México! this combination of the grotesque and the tragical themes in Mexican popular customs is discussed. 2.6 Eisenstein's own morphological and evolutionary theories have developed out of his biological interests. In the beginning of his scientific activity he was influenced by the ideas of biomechanics. According to his own words the main problem of his studies at the time when he was writing (in cooperation with the wellknown futurist writer S. M. Tretyakov) an unpublished essay on expressive gestures was the relation between central and peripheral movements. In modern terms this relation was stated in Kleist's brilliant essay on the theatre of marionettes which had anticipated recent mathematical analysis of the art. Discussing Kleist's idea Eisenstein stressed the fact that in organic movement, as opposed to that of a marionette, there is a certain conflict between the central impulse and the behaviour of the peripheral parts of the organism. The same differentiation of higher and lower levels and centres was later formulated in mathematical terms by one of the predecessors of modern cybernetic biology, N. A. Bernstein, who (as

THE THEORETICAL FRAMEWORK OF MODERN POETICS

847

Eisenstein himself) had started with biomechanic studies but afterwards built a new theory of movements based on the existence of different levels depending on the evolutionary chronology of the nervous centers (Bernstein 1947, 1966). These ideas were developed by such cyberneticians and mathematicians as I. M. Gelfand and M. L. Zetlin, according to whose theory 'the complex multilevel system of control is considered as a set of subsystems having relative autonomy. . . . Complex systems of control may consist of several levels; into each of them a row of such subsystems may be included' (Zetlin 1969). This generalized formula helps to understand that the relation of high and low levels is not a problem only of biology: it is a general scientific problem as all morphological questions are (that is why, for instance, Gelfand used language as a model for the study of some biological questions). Wider sociological implications of the theory discussed by the late Zetlin in his last works make it possible to compare this theory to that of Hocart who had tried to show (using as early as 1936 cybernetic notions of control and information) that the role of the center is not to be exaggerated from the evolutionary point of view: For there is a striking analogy between this evolution of human society and that of the body. In the history of the body we get at one end invertebrates and chordates, such as Amphioxus, which consist of segments placed in a line. In some worms the middle segments are so much alike that, if they were cut out, the anatomist could not put them back in the right order with any certainty. Each segment, though not independent, is self-contained: it may have its own muscles, nervous ganglions and dependent nerves, its own blood vessels, its own gonads; but they all are linked up, so that every segment regulates itself in harmony with the rest. There is a slight differentiation due to position. The most differentiated part is the head. It is only primus inter pares; but, because it goes first and so is the best placed to receive information, it exercises a general control over the body. Other parts are adapted to their position; the rear part has to close the procession. At the other end of the scale are highly centralized animals like ourselves. They no longer consist of almost identical and autonomous parts, but everything is under the control of a nervous system which radiates through the body. The head is no longer a mere primus inter pares, but the coordinating center. All information has to be passed on through it to lower or higher centres, which give out the appropriate impulse to the periphery, except in purely routine work which the organs themselves can deal with. The food supply too all goes up to the heart, and is distributed from it. We are far removed from the organism in which every part takes in its own nourishment, and responds directly to any stimulus. In the same way, out of a society of autonomous, but cooperative, parts there has evolved a society of specialized parts controlled from a centre. Each part used to be like each other and self-sufficient, except in the presence of common obstacles. Such common obstacles are droughts, disease, enemies that threaten the life of the whole. There are, in consequence, joint arrangements to overcome them. These arrangements are the communal rites. They first appear as rites independent, but to some degree autonomous; they grow into an elaborate organization centralized and controlled from above. Centralization means specialization (Hocart 1970: 38-39). Eisenstein was particularly interested in the oldest unspecialized stages of evolution. In his aesthetic studies he used to return to the problem of the undifferentiated

848

VYACHESLAV V. IVANOV

senses of the primitive organisms where there are no specialized organs (eyes for instance) and of the corresponding levels of evolution of society as reflected in rites and mythologies. He was particularly interested in the Aranda myth of the origin of human beings from the rella intarinja 'undifferentiated humans'. The main idea of Wagner's performance put on the stage in 1940 was the reconstruction of primitive undifferentiated society. Anticipating Levi-Strauss's definition of Wagner as the precursor of the structural study of myth, Eisenstein tried to understand and develop Wagner's main mythological image — the world tree which he made the central image of the performance (Eisenstein 1968). His investigations of this image are close to modern semiotic approach that succeeded in reconstructing the schema of the world tree common to all the mythologies of the world (Toporov 1964,1973). 2.7 Eisenstein's interest in myths and rituals as cues to the archetypes that he studied in his last aesthetic works (that remain unpublished) was connected with his conception of the 'Grundproblem' of the art. This 'main problem' consisted in the necessity of the conflict of at least two evolutionary levels in the structure of each work of art. According to Eisenstein's theory, developed already in 1934 in his address at the Congress of Filmmakers (Eisenstein 1964), the aesthetic form should be based on lower archaic levels cognate to those reflected in myths, rituals, symbols of the unconscious, etc.; at the same time it is possible to use art to express modern intellectual conceptions. As can be seen from Eisenstein's early writings in the first period of his aesthetic development he supposed that art can be understood on the primitive stimulus-reaction base formulated in terms of conditioned reflexes. Here again the analogy with the theories of Bernstein (who was Pavlov's former pupil) is evident. Just as Bernstein developed a new theory of the physiology of activity, stressing the importance of higher levels in the human central nervous system, Eisenstein too moved from primitive conceptions of a behaviorist type to the understanding of the multilevel structure of the art creation. He was closely associated with the great psychologist Vygotsky, who after he had written his important study of the psychology of art (Vygotsky 1971), in which he had given analysis of Freudian and formalist points of view, became interested in the semiotic approach to the highest forms of psychic activity. In the last years of his life (before his death in 1934) L. S. Vygotsky met with S. M. Eisenstein, A. R. Luria and N. I. Marr to discuss problems of inner speech and the archaic lower levels of consciousness studied in Vygotsky's important article on schizophrenia (Vygotsky 1956). In a series of semiotic works published posthumously (Vygotsky 1960) Vygotsky tried to show that in the elementary acts of common day behaviour certain relics of archaic sign systems can be found. The same approach seemed important for Eisenstein who used this diachronic analysis to undertand the oldest psychic levels vital for the form of the art. His conception of the 'Grundproblem' was connected with the diachronically motivated understanding of the role of regressive movements in art and in history. It can be by no means accidental that a series of Eisen-

THE THEORETICAL FRAMEWORK OF MODERN POETICS

849

stein's writings on the archaic evolutionary stages when 'there were no eyes' coincides chronologically with Mandelshtam's poem on Lamarck (where one finds the line 'There is no eye-sight') and with the development of the same theme of evolutionary regress in Raymond Quenaux's Gueule de Pierre (first written in 1934): Peut-être une lente et longue descente eut-elle été preferable. J'aurais recherché chez le singe tout ce qui y survit d'humain, puis dans le chien, le chat, l'éléphant, le raton laveur, jusqu'à l'ornithorynque; puis chez les oiseaux. Avec les reptiles déjà, j'aurais pressenti les premieres fissures. Les poissons, bien que toujours vertebres, inquietent difinitivement. Avec les invertebres, commence l'angoisse. Mais ce chemin aurait été trop long. Je ne recherche pas la déperdition de l'humain à travers les éspèces, mais l'aube de l'unhumanité (Quenaux 1948: 2).

According to J. Queval (1960: 192), this 'dawn of inhumanity' is the main theme of the book of Quenaux and, one may add, of Mandelshtam's "Lamarck". In Mandelshtam's recently published notes on naturalists he wrote: 'Lamarck feels the gaps between the classes. These are intervals of the evolutionary row. The abysses are yawning. He is listening to the syncopes and pauses of the evolutionary row' (Mandelshtam 1968: 180). Here again one is impressed by the apparent analogy with such modern morphological conceptions of evolution as that of D'Arcy Wentworth Thompson to whom the 'principle of the discontinuous' compared with physicists' quanta seemed the central point of the evolutionary theory (Thompson 1952: 1094). A year after the publication of Mandelshtam's "Lamarck", B. M. Eichenbaum wrote in his report "On Mandelshtam" (March the 14th, 1933): 'his biologism is not without relation to history and that is why it is not asocial', he saw this 'biologism as visionarism an das a wide semantic sphere for the poetic "self" ' (Eichenbaum 1967: 168). The vision of regress towards the earliest stages of evolution was determined by the social atmosphere of the time. Thus it became the main theme of Eisenstein's unpublished aesthetic works "Method" and "Grundproblem". 2.8 The main difference between Eisenstein's approach towards the ethnological (archetypal) study of poetics and that of such scientists as Freudenberg may be illustrated by Eisenstein's remark on Freudenberg's "Poetics of the plot and the genre". He felt dissatisfied with the fact that in this work Freudenberg had studied only the contents of the plot (trying to find, for instance, the archetypal themes in Calderon's and Shakespeare's plays) but not the structure of the situation itself. In this respect Eisenstein is much nearer to Propp and Hocart in their strictly morphological approach. The presupposition for the archetypal analysis was the reconstruction of the formal pattern. Only afterwards one might attempt to give it a semantic (or diachronic) interpretation. Critical remarks of this kind are possible, for instance in regard to Frye's interpretation of Kafka's Trial. In Frye's analysis (Frye 1969: 42) the archetypal scheme of the novel is suggested without any comparison with other of Kafka's

850

VYACHESLAV V. IVANOV

writings. But the analogies in formal patterning of Trial and Castle are too great to be left without consideration. Not only the general scheme of the judicial organization (in Trial) or of the 'Castle' (in the novel with the same title) being unattainable for the hero is the same but some important implications in the relations of secondary characters to the main hero (Joseph K.) can be identified. Thus Joseph K. is related to the mistress of the lawyer in Trial in the same way as the hero of the Castle is related to Klamm's mistress. Two persons who came to Joseph K. in the beginning of the Trial are similar to his two assistants in the Castle. This point is made clear by the text of the second novel: ' "Frau Wirtin", sagte K. . . . "es sind meine Gehilfen, Sie aber behandeln sie so, wie wenn es Ihre Gehilfen, aber meine Wächter wären".' It seems that in his analysis of Kafka's novel Frye did not apply the method which he used with success while discovering 'a central unifying myth' (Frye 1969: 192) that is invariant (Urmyth or primal myth) defined by all possible transformations. It can be shown that the same method may help to establish the general pattern of Dostoevsky's later novels. To such a pattern belong the situations of mass scandals and the spatio-temporal properties described in Bachtin's study (Bachtin 1963) which succeeded in deriving these characteristics from the total polyphonic principle. In modern structural or generative descriptions of the works of verbal art based on the principles of Eisenstein's poetics, his idea of recoding the main invariant theme by means of different expressive means is exposed (Zholkovsky and Scheglov 1971, 1972). But it is necessary to stress several important differences between Eisenstein and his followers. First of all for Eisenstein as for Propp the analysis of the formal pattern was only the first step before its diachronical interpretation could be applied. Secondly Eisenstein supposed that the form (or structure as he usually put it) itself was in iconic relation to the theme that was expressed in the given work. That is why he was far from generalizing the set of separate expressive means that he explained in his aesthetic treatises. To his mind not the paradigmatic set of such means but their interrelationship with other components of the aesthetic whole was the main problem. 2.9 This structuralist point of view may be illustrated on the example of the problem of inner speech. Eisenstein saw in the inner monologue the most evident manifestation of the archaic level of consciousness. He was particularly impressed by Joyce's inner monologue in Ulysses. As he explained in a special chapter on Joyce and Dujardin in his "Method" the difference between Joyce and those authors who had tried to use inner monologue in their novels before Joyce lies in the fact that Joyce succeeded in rendering the structure of inner speech. Later Eisenstein thought that this structure reflects those lower levels which define the formal characteristics of the art. To the English translation of his article "A course in treatment" (Eisenstein 1957: 57) Eisenstein gave the following motto from Ulysses: 'So that gesture, not music, not odours, would be a universal language, the

THE THEORETICAL FRAMEWORK OF MODERN POETICS

851

gift of tongues rendering visible not the lay sense, but the first entelechy, the structural rhythm'. This aesthetic conception of Joyce's hero Stephen Dedalus that appeared so near to Eisenstein's own understanding of the role of structural rhythm in the art is fully developed in Joyce's first novel The portrait of an artist as a young man. In this novel Stephen Dedalus is working on aesthetics 'by the light of one or two ideas of Aristotle and Aquinas' (Joyce 1934: 212). A hint of Aristotle is hidden already in the above-given quotation from Ulysses where entelechy is used in the Aristotelian sense familiar to other scientists and theologists of the 20th century, for instance to P. A. Florensky (Rosanov 1918). The structural aesthetics of Joyce's early writings was based also on Aquinas' saying: 'Ad pulcritudinem tris requintur integritas, consonantia, claritas', which the author himself renders as 'three things are needed for beauty, wholeness, harmony and radiance' (Joyce 1934: 241). The structural concept of the whole was based on the notion of rhythm: 'Rhythm, said Stephen, is the first formal aesthetic relation of part to part in any aesthetic whole or of an aesthetic whole to its parts' (Joyce 1934: 234). Structural rhythm is seen as the method to perceive the whole: 'You pass from point to point, led by its formal lines; you apprehend it as balanced part against part within its limits; you feel the rhythm of its structures' (Joyce 1934: 241). Since the last formula coincides completely with that cited by Eisenstein from Ulysses it can be established that a whole conception of structural rhythm organizing the aesthetic whole can be found in Joyce's novel. Joyce's own composition was understood by Eisenstein as the practical application of these aesthetic theories. Eisenstein considered Joyce to be an experimental writer in the sense in which one may speak about experimental aesthetics. The most important experiment according to Eisenstein consisted in the fact that Joyce mastered the structural rhythm of inner speech recreating thus the 'physiology' of human affects. Eisenstein himself tried to do the same thing in his early films. As early as 1927-1928 he wrote in his diary about his projected intellectual film "Das Kapital" that the formal side of it would be dedicated to Joyce. Later he explained that all his study of the 'Grundproblem' grew out of his interest for inner speech. Here again one might see interesting parallels to his friend Vygotsky's scientific preoccupation. Vygotsky was one of the first to find the genetic link between children's early egocentric speech (in Piaget's sense) and inner speech. The same diachronic (not ontogenetic, but phylogenetic) approach was characteristic of Eisenstein who studied inner speech in connection with the archaic strata of consciousness. Eisenstein appreciated Dostoevsky's short story "The meek" (included in The writer's diary) because of the use of the inner monologue in this work. But he stressed the fact that only in some parts of the story did Dostoevsky recreate the syntax of inner speech. Similar remarks were made by Shklovsky about young Tolstoy's History of yesterday (Shklovsky 1939: 148). Thus the comparison with Joyce helped to understand the creative possibility hidden in classical Russian prose.

852

VYACHESLAV V. IVANOV

2.10 Eisenstein's example is extremely important because it serves to explicate the relation between diachronic and synchronic approaches in modern morphological studies in poetics. Eisenstein's conception belonged to that line of typological diachrony that was characteristic of the most interesting Russian humanitarian studies of the first half of the century. Its precursor might be seen in Veselovsky who tried to construct a huge building of historical poetics (Shishmarev 1972) and influenced Propp's morphological studies (for instance, in his understanding of the motif in its distinction from the plot). Early formalist writings, particularly those of Shklovsky, were directed against Veselovsky and Potebnya, but it seems now that in the diachronic poetics of both the great preformalist scholars one might find more impulses for serious structural understanding of the literary patterns than in Shklovsky's brilliant but chaotic games of aphorisms. Those structural works on poetics that were done afterwards, both in Russia (as Bachtin and Eisenstein) and in Czechoslovakia, where Mukafovsky tried to construct historical poetics (Wellek 1969: 13), were characterized by the combination of synchronic and diachronic studies. The approach to the theory of genres in the works by Propp, Bachtin, Freudenberg, Eisenstein can be compared to that in modern astrophysics. The scientists begin by synchronic comparison of different types of objects and their morphological classification. But afterwards they have to build a diachronic theory of evolution in which every type finds its explanation. The distinction between synchrony and diachrony is not taken away, but both the approaches are united, as in modern transformational grammar, to help reconstruct the previous stages of linguistic development.

3. GENERATIVE POETICS AND LITERARY CREATIVITY

3.1 The first approach to generative metrics was made in the Classical Arabic theory of *arud as was shown independently by Morris Halle (1966: 113-116) and A. Sanchez (1968: 86-95), the latter stressing important links between Arabic metrics and what might be called the transformational approach in Arabic linguistics. The main principle of generative poetics is formulated in the Arab statement according to which *artid is the science studying the main concrete rhythms of the verses out of which by the rules of deduction all the concrete multivarious rhythms are deduced that are declared to be correct Qauzan assfr-far") this formulation being implicitly the criterion for discarding other rhythms as faulty' (Sanchez 1968: 93-94). It might be supposed that in some other old systems of versification similar generative rules might be applied. Thus the system of classical Chinese verse distinguished some sequences of tones as correct from faulty sequences which helped to build symmetrical patterns of tones studied recently by Roman Jakobson from the point of view of the general theory of symmetry (Jakobson 1970: 597605). But it is not yet sufficiently clear whether this procedure in early Chinese

THE THEORETICAL FRAMEWORK OF MODERN POETICS

853

metrics was already formalized to a degree comparable with the achievement of the Arab grammarians It seems obvious that metrics and the study of other formal sides of poetry has been the field were the most exact methods can be applied. Thus Zhirmunsky's remarkable books on metrics (Halle 1968), rhyme, and composition and Jakobson's study of Czech and Russian verse (later developed in his collective work with Mukarovsky) remain as most solid results of the formalist activities of the twenties. Tomashevsky's writings on verse, showing sufficient knowledge of mathematics, have laid foundations for modern mathematical poetics. Other representatives of the Russian formalist school such as Tynyanov and Eichenbaum were most closely related to the structural linguistic approach in their study of verse. In the last two decades the use of modern mathematical tools and some ideas of generative grammar have transformed metrics into the model for other parts of poetics which are still far from being exact to the same degree. 3.2 A s Chomsky has shown with success in his book on the history of the generative approach in grammar, the French linguist of the 18th century Du Marsais was a forerunner of the modern search for linguistic universals (Chomsky 1966: 45-53). It is by no means accidental that Du Marsais's theory of tropes is developed in some of the recent structuralist works on poetics (Todorov 1967: 97-118). Particularly close to modern combinatorial semantics was Du Marsais in his statement that 'ce n'est que par une nouvelle union des termes, que les mots se donnent le sens méthaphorique' (Du Marsais 1818: 161). But it is necessary to state that Todorov in his study of Du Marsais, as most of the authors of modern literary and linguistic works trying to reformulate old rhetoric classification (Hendricks 1970: 165-177), still adheres to the word (or to the sentence) as the main object of study. Meanwhile even recent research on linguistic generative semantics has shown that it is by no means adequate to formulate semantic rules as referring to a word and not to a larger unit. Thus Bachtin's conception of 'metalinguistic' ('translinguistic' in R. Barthes's terminology) science studying the literary context larger than a sentence (Bachtin 1963) becomes more and more actual. After Mathesius' pioneering linguistic study and Harris' work on discourse analysis the problem of the text as a syntagmatic whole was investigated by several linguists and literary scholars (Mayenowa 1971; Lotman 1970: 371, n.7; Grimes and Glock 1970; Loriot and Hollenbach 1970; etc.) particularly in many works on 'Textlinguistik' by W . Dressier and other scientists. But the most important ideas on the subject were expressed by Bachtin. His attitude in his early writings towards metalinguistics as opposed to linguistics was very close to the point of view expressed in Benveniste's recent articles, where the sharp distinction between the science of discourse and that of signs is suggested (Benveniste 1969). The difference between the approach suggested by Bachtin and that of other modern authors lies in the fact that according to him metalinguistics studies the communicative aspect of a text. That is why it may study even a sign - a word if it is taken in the context of concrete communica-

854

VYACHESLAV V. IVANOV

tion. Thus the distinction is not restricted to the length of the objects, it refers rather to the relation between a part (the sign) and the whole (the discourse). The program of the metalinguistic study of discourse was fully developed in Bachtin's early writings particularly in reference to the problem of the 'word of the other person'. 2.3 To use Bachtin's own terminology the dialogue is the dominant of all his creative work, the central notion with which all his principal themes and achievements can be associated. The dialogical relations are discovered in the structures of a whole text and in those types of speech that presuppose an attitude directed towards 'the word of the other person'. In the classification suggested by Bachtin (1963: 266-267) and consisting in an analogue of the periodic system of elements for describing prose the place is given to different kinds of narration such as skaz, studied before him by formalists (Erlich 1965: 75 and 238). In Bachtin's early writings (published under the name of his former student Voloshinov) a new interpretation was given to the problem of reported speech which he understood as a 'discourse about discourse' or 'message referring to message' as Roman Jakobson wrote later using the terms of information theory to render Bachtin's ideas (Jakobson 1971). Later work on reported speech in avantgarde prose (Dolezel 1960) was influenced by Bachtin's ideas. Bachtin was the first to stress the role of language contacts for the origin of the novel. 'Every novel in a more or less degree is a dialogized system of "languages", styles, concrete consciousness that cannot be set apart from their languages. The language in a novel is not so much the means to describe, it is itself the object of description' (Bachtin 1965b: 89). If the metalanguage of linguistic (especially that of Old Indian grammar) was created for the interpretation of 'the word of another language' (of Sanskrit that was coming out of the everyday usage) the novel and its multivariance of languages helped to establish the function of a word not as the means of description but as an object of description. That is why different genres of parody were necessary to form the ground for the novel, since in parodies the word was freed from its primary functions. The parody was understood by Bachtin as the result of creolization of languages (Bachtin 1967: 21). In a series of lectures read before World War II, but published in some fragments only several years ago, Bachtin developed the conception of the novel arising from the creolization of languages. It can be proved that a multitude of languages is typical of the novel. It refers not only to such classical works as Anna Karenina (Jakobson and Halle 1956: 18) but also to the best novels of the 'third world'. The use of the French language in one of the best novels in Modern Vietnamese literature — S6 do — by Vu Trong Phung (first published in 1937) — deserves a special study from this point of view. In his critiques of the early formalist writings Bachtin stressed the fact that the methods used by this school were not sufficient for the study of the novel. It is worth

THE THEORETICAL FRAMEWORK OF MODERN POETICS

855

noticing that the problem of the morphological evolution of the novel was studied by such precursors of the German morphological school as Dibelius (Dibelius 1922). For Bachtin the novel was the main object of all his studies. He investigated the novel using classical examples (such as Rabelais' and Dostoevsky's novels) or even descending down to the origins of the genre. But for him it is first of all the genre that appears before our eyes, a creative experiment and not the dead legacy of the past. One can not but be amazed by the deep coincidence between the laws of the genre discovered by Bachtin and the structural characteristics of Joyce's Ulysses, written in the time of Bachtin's youth. One might note in parentheses that Joyce's compatriot and contemporary J. M. Synge was the author of "A Rabelaisian rapsody", in which Rabelais is made equal 'to any of the Saints'. In Ulysses one may find the fullest development of all of the main features of the genre of the novel according to Bachtin: the point-counter-point interlacing and ideological comparison of different genres of speech, their conflict inside the novel, the use of parodies and travestied displaced words. The parallelism between the aesthetic theories and literary practice typical of avantgarde literature sometimes may be understood by the causal link between them (Pomorska 1968). But in the case of Ulysses and Bachtin's theory one is confronted by the striking parallelism that should find its explanation not in the reciprocal influence of the theories on the artistic creative experiment, but in the certain common spiritual scheme that should be explained by the typological history of culture. 2.4 Particularly illuminating for the study of the novel is the analysis of its structure in reference to different points of view on which the narration is based. As pointed out by Tzvetan Todorov (1968: 158) this parameter was introduced by Lubbock in 1921 in the description of the novel in his study of Henry James (Lubbock 1965; Booth 1961) that was continued by Todorov himself (1971: 187-188). Pouillon in his important study of time in the novel that had partly anticipated later work on the subject (Lammert 1955; Weinrich 1964) distinguishes between three main points of view in the novel (Pouillon 1946). Particularly important for the comparison with Bachtin's poetics of the novel seems the point of view that coincides with that of a hero (Pouillon's 'vision avec'). The study of classical Russian prose in the light of Bachtin's ideas concerning different points of view was developed technically in recent years (Uspensky 1970; Chudakov 1971). But it is necessary to stress that to Bachtin the problem of the point of view has never become a technical question. His formulation of the problem was rather philosophical than technical. His description of the consciousness of one man against the background of another, 'I-formyself' on the background for 'I-for-another' (Bachtin 1963: 277) coincides word for word with the formula in an early work by Sartre (1943: 277) who has expressed later similar views on the modern novel. To Bachtin the main point was the distinction between the monological philosophy of the writers of the last

856

VYACHESLAV V. IVANOV

centuries and the possibilities of dialogical understanding of the world which he, as his contemporary Chicherin (1970: 95, 134, 251 and others), compared to polyphonic structures in music. From this point of view Dostoevsky (in opposition to Goethe) is always synchronic: 'Dostoevsky as opposed to Goethe tended to understand different periods in their synchronic coexistence, to compare and oppose them to each other, and not to extend them in a row of becoming' (Bachtin 1963: 38). Thus paradoxically enough the morphological analysis of transformations in the history of literature has led to the discovery of an opposite procedure in the object of the study itself. 2.5 Recent metalinguistic research on the microstructure of words and contexts larger than a sentence in an artistic text have not yet approached the limit where they can meet with the results of the improved macrostructural analysis of narration based on Propp's methods (Ihwe 1972; Hendricks 1973; etc.). But some hints as to the possibilities of the future joints between these two fields of research may be found in Bachtin's study of the word in the novel and in the new understanding of metaphoric and metonymic styles due to Roman Jakobson (Jakobson and Halle 1956). Jakobson saw in the distinction between metaphor and metonymy the result of a general semiotic distinction between paradigmatic and syntagmatic structures. In the light of this analogy he distinguishes the metaphoric style typical of poetry and the metonymic one characteristic of prose (Jakobson 1973). Modern investigations on the essense of the metaphor make possible the comparison with similar notions in the theory of film where the montage metaphors of the early cinema were opposed to later metonymic style by Eisenstein and other theoreticians (Ivanov 1970). The metaphor and metonymy cease to be objects of lexical studies and become involved in the complex studies of the text and of its informational characteristics (/?, hi, I12). The approach of the language of modern poetry to everyday speech noted by great poets (such as Mandelshtam and T. S. Eliot) themselves may be compared to the prominence given to other metonymic means in modern art (especially in prose where the stress on detail as the main formal device unites different authors from Boll and Salinger to the writers of detective stories). In the Russian tradition of poetics the structure of the sign-image as its main object was perceived by Potebnya, whose prestructuralist formulation of the problem was influenced by Humboldt's creative view of language. Later on some of the early formalists such as Shklovsky tried to oppose this evidently sound approach to poetics. G. Shpet, who was the first to suggest a multilevel approach to the study of literary text, was also against Potebnya's psychological approach to the inner form. Poetics was understood by him as 'the grammar of poetical language and poetical thought' (Shpet 1923). His system of poetics is rationalistic since the upper level of a text is logical, as in modern linguistic theory. The whole trend of the development of the most fruitful ideas of Moscow, Petrograd and Prague formalists

THE THEORETICAL FRAMEWORK OF MODERN POETICS

857

and structuralists seem to confirm the fact that the ways of modern poetics and linguistics are the same. Their interest lies in the establishment of those laws that can explain the creative activity of a speaker or of a writer. The complexity of the task is not to be underestimated. But one may hope that some of the above mentioned scientists and many others wo~king by similar methods have opened the path for future understanding of human spiritual achievements.

REFERENCES

M. 1963. Problems of Dostoevsky's poetics. 2nd ed. Moscow. (In Russian.) . 1965a. Rabelais's work and the popular culture of the Middle Ages and of the Renaissance. Moscow. (In Russian.) English translation: Rabelais and his world, Cambridge, Mass., 1968. . 1965b. The word in a novel. Problems of literature 1965/8. (In Russian.) . 1967. From the prehistory of the word in the novel. Russian and Western literature, Scientific papers of the Mordovian University 1967/61. Saransk. (In Russian.) BALUHATYI, S. D. 1927. Problems of the analysis of dramaturgy. Leningrad. (In Russian.) BENVENISTE, É . 1 9 4 7 . Le jeu comme structures. Deukalion 1 9 4 7 / 2 . . 1964. Lettres de Ferdinand de Saussure à Antoine Meillet. CFS 21. . 1969. Sémiologie de la langue I—II. Semiotica 1.1-12, 127-135. BERNSTEIN, N. A. 1947. On construction of movements. Moscow. (In Russian.) . 1966. Essays on the physiology of movements and physiology of activity. Moscow. (In Russian.) BOOTH, W. C. 1961. The rhetoric of fiction. Chicago. CAMPANILE, E. 1963. Note sur Saturnio. Annuale Scuola Normale Superiore de Pisa 32.191-197. CHICHERIN, G . 1 9 7 0 . Mozart. Leningrad. (In Russian.) CHOMSKY, N. 1 9 6 6 . Cartesian linguistics. New York and London. CHUDAKOV, A . P . 1 9 7 1 . Chekhov's poetics. Moscow. (In Russian.) COLE, T. 1969. The Saturnian verse. Studies in Latin poetry. YCS 21.1-75. DANDIN. 1 9 5 2 . Dandin's Kâvyâdarsa, with the commentary of Jeebananda Vidyasagara Bhattacharya and an introduction and Indian translation by V. Narayana Iyer. Madras. DIBELIUS, W. 1922. Englische Romankunst, Bd. I—II. 2nd ed. Berlin. DOLEZEL, L . 1 9 6 0 . O stylu moderni ceské prozy. Praha. DRIESSEN, F. C. 1965. Gogol as a short-story writer. The Hague. EICHENBAUM, B . M. 1 9 6 7 . From the unpublished works. On Mandelshtam. The day of poetry, 1967, Leningrad. Leningrad. (In Russian.) BACHTIN,

858

VYACHESLAV V. IVANOV

S. M. 1957. Film form. Film sense. New York. . 1964. Selected writings, vol. II. Moscow. (In Russian.) . 1966. Selected writings, vol. IV. Moscow. (In Russian.) . 1968. Selected writings, vol. V. Moscow. (In Russian.) ERLICH, V . 1965. Russian formalism: History, doctrine. 2nd ed. The Hague. ERMAN, V. G. 1961. The theory of drama in Old Indian classical literature. Drama and theatre of India. Moscow. (In Russian.) FREUDENBERG, O. M. 1936. Poetics of the plot and the genre. Moscow-Leningrad. (In Russian.) FRYE, N. 1969. Anatomy of criticism. New York. GASPAROV, M. L. 1969. B. Iarcho's work on exact poetics. ERMSICOTIKII IV. Tartu. (In Russian.) GATES, H. P. 1971. The kinship terminology of Homeric Greek. IJ AL Supplement 37/4. GERZENBERG, L. 1972. Morphological structure of the word in Old Indo-Iranian languages. Leningrad. (In Russian.) GOETHE, J. W . 1963. Schriften zur Botanik und Wissenschaftslehre. München. GRIMES, J. E., and N. GLOCK. 1970. A Saramaccan narrative pattern. Lg 46/2, Part 1. HALLE, M. 1966. On the metrics of Pre-Islamic Arabic poetry. QPR-RLE 83. . 1968. Zirmunsky's theory of verse: A review article. SEER 12. HAYMAN, D . 1973. Au-delà de Bakhtine. Poétique 13. HEISENBERG, W. 1967. Die Goetesche und Newtonsche Farbenlehre im Lichte der modernen Physik. Goethe im XX Jahrhundert. Hamburg. HENDRICKS, W. O. 1970. Review of G. N. Leech, A linguistic guide to English poetry. Lingua 25/2. . 1973. Methodology of narrative structural analysis. Semiotica 7/3. HOCART, A. M. 1969. Kingship. 2nd ed. Oxford. . 1970a. Kings and councillors. 2nd ed. Chicago. . 1970b. The life-giving myth and other essays. 2nd ed. London. IHWE, J. 1972. On the foundations of a general theory of narrative structure. Poetics 3.5-14. IVANOV, V. V. 1970. Eisenstein et la linguistique structurale moderne. Cahiers du cinéma 1970. . 1971. On one typological parallel to Gogol's Viy. XrineicotiKf) V. Tartu. (In Russian.) . 1972. Two examples of anagrammatic compositions in the later verses of Mandelshtam. Russian Literature 3.81-87. (In Russian.) . 1973. La traduction poétique à la lumière de la linguistique. Change 14: Transformer, Traduire. Paris. IVANOV, V. V., and V. N. TOPOROV. 1973. Etymological study of semantically restricted lexical groups in connection with the problem of reconstruction of EISENSTEIN,

THE THEORETICAL FRAMEWORK OF MODERN POETICS

859

Proto-Slavic texts. Slavic Linguistics. Moscow. (In Russian.) R. 1958. Medieval mock mystery. Studia Philologica et Litteraria in Honorem L. Spitzer. Bern. . 1966. Slavic epic verse. Selected Writings, vol. IV. The Hague. . 1970. The modular design of Chinese regulated verse. Echanges et communications: Mélanges offert à C. Lévi-Strauss. The Hague. . 1971. Shifters, verbal categories, and the Russian verb. Selected Writings, II. The Hague. . 1973. Questions de la poétique. Paris. JAKOBSON, R . , and M. HALLE. 1956. Fundamentals of language. The Hague. JOYCE, J . 1934. A portrait of the artist as a young man. London. KOLMOGOROV, A. N. 1965. Three approaches to the definition of a concept "quantity of information". Problems of the Transmission of Information 1/1. (In Russian.) . 1969. Towards logical foundations of the theory of information and the theory of probability. Problems of the Transmission of Information 5/3. (In Russian.) LÄMMERT, E. 1 9 5 5 . Die Bauformen des Erzählens. Stuttgart. LEACH, E. R. 1961. Rethinking anthropology. London. . 1964. Anthropological aspects of language: Animal categories and verbal abuse. New directions in the study of language, ed. by E. H. Lenneberg. Cambridge, Mass. LÉVI-STRAUSS, C. 1960. Analyse morphologique des contes russes. I J S L P 3.12249. . 1971. L'homme nu. Paris. LORIOT, J . , and B . HOLLENBACH. 1 9 7 0 . Shipilo paragraph structure. F L 6 . LOTMAN, J . M. 1 9 7 0 . Structure of the text of the art. Moscow. (In Russian.) LUBBOCK, P. 1965. The craft of fiction. London. MANDELSHTAM, O . E . 1 9 6 8 . Notebooks: Notes. Problems of Literature 1 9 6 8 / 4 . (In Russian.) MARCUS, S . 1 9 7 0 . Poetica matematicä. Bucureçti. Dü MARSAIS = FONTANIER, P. 1818. Les tropes de Dumarais, avec un commentaire r a i s o n n é . . . . Paris. MAYENOWA, M. R . , ed. 1 9 7 1 . O spôjnoéci tekstu. Praca zbiorowa pod redakciq M. R. Mayenowej. Wroclaw-Warszawa. MEILLET, A . 1 9 2 3 . Origines indo-européennes des mètres grecs. Paris. MIKO, F . 1 9 7 0 . Text a styl. K problematike literârnej komunikacie. Smena. MILLER, D. G. 1968. Traces of Indo-European metre in Lydian. Studies presented to Professor Roman Jakobson by his students. Cambridge, Mass. MILNER, G . B . 1 9 7 2 . Homo Ridens. Semiotica 5 . 1 - 3 0 . M O N O D , J. 1970. Le hasard et la nécessité. Paris. PARIS, J . 1 9 7 0 . Rabelais au futur. Paris.

JAKOBSON,

860

VYACHESLAV V. IVANOV

B. L. 1965. The translator's notes. Literary Russia 1965/13.18. (In Russian.) POMORSKA, K . 1 9 6 8 . Russian formalist theories and its poetic ambiance. The Hague. POTEBNYA, A. A. 1865. On mythological meaning of some rites and beliefs II. Readings in the Emperor's Society of the Russian history and antiquities, Book 3, July-September 1865. Moscow. (In Russian.) POUILLON, J. 1946. Temps et roman. Paris. PROPP, V. J. 1939. Ritual laughter in folk-lore. Scientific Papers of Leningrad University 46. Leningrad. (In Russian.) . 1946. Historical roots of the fairy-tale. Leningrad. (In Russian.) . 1971. Fairy-tale transformations. Readings in Russian poetics: Formalist and structuralist views, ed. by L. Matejka and K. Pomorska. Cambridge, Mass. QUENAUX, R . 1 9 4 8 . Saint Glingin. Paris. QUEVAL, J. 1960. Essai sur Raymond Queneau. Poètes d'aujourd'hui 72. Paris. RASTIER, F. 1970. A propos du Saturnien. Latomus 29/1.3-24. ROSANOV, V. V. 1918. The Apocalypse of our time. Sergiev Posad. (In Russian.) SANCHEZ, A. A. 1968. On the problem of the essentials of the Arabic system of metrics. Arabic philology, ed. by A. A. Kovalev and G. M. Gabuchan. Moscow. (In Russian.) SARTRE, J.-P. 1943. L'être et le néant. Paris. SCHMITT, R. 1967. Indogermanische Dichtersprache. Wiesbaden. SHISHMAREV, V. F. 1972. Alexandre Nikolaevich Veselovsky. Selected essays. Leningrad. (In Russian.) SHKLOVSKY, V . B. 1 9 3 9 . Diary. Moscow. (In Russian.) SHPET, G. G. 1 9 2 3 . Aesthetic fragments. 1 - 3 3 3 . Petrograd. (In Russian.) STAROBINSKI, J. 1 9 7 1 . Les mots sous les mots: Les anagrammes de Ferdinand de Saussure. Paris. THOMPSON, D'ARCY WENTWORTH. 1 9 5 2 . On growth and form, vol. I I . 2nd ed. Cambridge. TODOROV, T . 1967. Littérature et signification. Paris. . 1968. Poétique. Qu'est-ce que le structuralisme? Paris. . 1971. Poétique de la prose. Paris. TOPOROV, V . N . 1 9 6 4 . Towards the reconstruction of some mythological images (on the material of the Buddhist visual art). People of Asia and Africa 1 9 6 4 / 3 . (In Russian.) . 1973. L'«albero universale». Saggio d'interpretazion semiotica. Ricerche semiotiche. Torino. TURNER, V. 1969. The ritual process. Chicago. USPENSKY, B . A. 1970. Poetics of composition. Moscow. (In Russian.) VIËTOR, K. 1950. Goethe the thinker. Cambridge, Mass. VOLOSHINOV, V . N . 1927. Freudism. Moscow-Leningrad. (In Russian.)

PASTERNAK,

THE THEORETICAL FRAMEWORK OF MODERN POETICS

861

S. 1956. Selected psychological papers. Moscow. (In Russian.) . 1970. Development of highest psychic functions. Moscow. (In Russian.) . 1971. Psychology of the art. Cambridge, Mass. WATKINS, C . 1 9 6 3 . Indo-European metrics and archaic Irish verse. Celtica 6 . . 1970. Language of gods and language of men: Remarks on some IndoEuropean metalinguistic traditions. Myth and law among the Indo-Europeans: Studies in Indo-European comparative mythology, ed. by J. Puhvel. Berkeley and Los Angeles. . 1972. A Palaic carmen. Department of Linguistics of Harvard University, Indo-European Studies, special report to the National Science Foundation. Cambridge, Mass. WEINRICH, H. 1964. Tempus: Besprochene und erzählte Welt. Stuttgart. WELLEK, R. 1969. The literary theory and aesthetics of the Prague School. Michigan Slavic Contributions. Ann Arbor. WEST, M . L . 1 9 7 3 . Indo-European metre. Glotta 5 1 / 3 ^ 1 . er WUNDERLI, P. 1972a. F. de Saussure: "1 Cahier ä lire preliminairement". Zeitschrift für französische Sprache und Literatur 82/3. . 1972b. Ferdinand de Saussure und die Anagramme. Tübingen. ZETLIN, M. L. 1969. Studies on the automata theory and the models of biological systems. Moscow. (In Russian.) ZHOLKOVSKY, K. A., and J. K. SCHEGLOV. 1971, 1972. Towards the description of the sense of a contiguous text I—II. Institute of Russian Language, Research Group of Experimental and Applied Linguistics, Preprints nos. 22 and 33. Moscow. (In Russian.) VYGOTSKY, L .

T H E O R E T I C A L POETICS IN THE TWENTIETH CENTURY F. SVEJKOVSKY

0.

INTRODUCTION

Since the turn of a century does not necessarily coincide with a turning point in the evolution of a science or art, it would be in order to question whether the limitation of our theme to the twentieth century is not too mechanical. It may, furthermore, be appropriate to ask whether it is not premature to draw conclusions concerning a development that began only seventy years ago. And yet, it appears that the new developments in art and in the theory of art which are often designated by the not too accurate label 'modern' came to the fore and began to assert themselves precisely at the turn of the present century. Similar considerations also apply to the development of modern poetics, which emerged as a new kind of discipline. It was neither a descriptive poetics, dealing with the 'description of form', rules or the technique of a work; nor a poetics which revolved around esthetic and philosophical notions; nor a poetics that emphasized individual features of an author who creates his own unrepeatable forms. It was a poetics which imparted a new sense to the tendencies implicit in older theories, placing them, as it were, in a new context. But only from the standpoint of immediate antecedents does something appear to be completely new, for what is new as a negating force is in broader context merely a variation of a general attempt to grasp the essence of certain phenomena. The historical process is thus not circular, for each historical step contributes features of its own to the spiral of our knowledge. Twentieth century poetics is charged with the energy of the new and with the courage to overcome the old, from which it draws at the same time much of its support. The abandonment of the tenets of the second half of the nineteenth and the beginning of the twentieth century has become the starting point for a development which resounds, like modern linguistics (e.g. in the works of Jakobson and Chomsky) with reminiscences of traditional poetics and its ties to grammar, rhetoric and logic, along with references to contemporary philosophy, logic and linguistics. In the works of the Russian Formalists as well as in those of the French group centered around R. Barthes, we find both references to these old roots of scholarship and remarkable innovating convergences. It is perhaps enough to mention as an example a single chapter from G. Morpurgo-Tagliabue's recently

864

F. SVEJKOVSKY

published book Linguistica e stilistica di Aristotele (1967: 352f.). Here the author is able (after the publication of Todorov's 1965 anthology of translation from the Russian Formalists) to call attention to the 'classical' tradition in one of the most outstanding trends of twentieth century poetics and to relate it to the broader context of recent literary and linguistic structuralism. And if N. Frye speculates about different types of interpretation in his Anatomy of criticism (1957), he is consciously reviving the tradition of medieval hermeneutics, while his conception of the poet recalls the notion of 'the efficient cause' of the work. Heidegger's suggestions, which see the essence of a work in the relationship 'Wahrheit-Seiende-Schonheif, recalls the medieval conception of the unity of good-truth-beauty. It would be easy to multiply the examples by referring, for instance, to the 'Chicago School of Criticism'. It is not, however, our intention to enumerate the reversion and references to the past, but rather to point to the existence of coalescences and the reappearance of problems and methods in the different turns of the historical spiral. Such coalescences impart to the present the dimension of the past.

1.

POETICS A N D ITS PLACE IN TWENTIETH CENTURY THEORETICAL STUDIES 1

1.1 Today's dictionary definitions originate from a long tradition, and they examine this tradition in various ways, in part according to the different usages of a word in particular countries or cultural spheres. Generally they concentrate on several points. (1) Definition from existing treatises designated by the title 'poetica'; categories and means are described with the aim of classifying the principles of the creative process or of codifying them. (2) Theories concerning the concept of a literary work; this encompasses more than the first case, since problems of historical poetics, for example, also take into account the evolutionary dynamics of single categories. (3) In a stricter sense characterization of the creative principle of a certain author, group, school and the like. (4) Less common today is the conception of poetics as the theory of poetry. (5) Sometimes poetics is equated with literary theory. This is because more general aspects of study, which exceed the limits of theories pertaining to the intrinsic composition of the work, to its realization by linguistic means, are included under the term 'poetics'. Or because, on the contrary, the focal point of literary theory is seen in the study of the work itself, in its shaping; because the autonomy of the act of creation is stressed; and because questions of the connections of the literary, artistic work with the remaining cultural and social context are relegated to other disciplines. Each of the above points could be further specified. Similarly the addition of historical aspects would introduce other distinguishing details into our classification. However, to trace the history of definitions is not the aim of this essay. In 1

This survey was essentially completed by 1969-70. Therefore it was possible to take later works into account only in isolated cases.

THEORETICAL POETICS IN THE TWENTIETH CENTURY

865

the same way we shall not concern ourselves with a study of the usus of the term 'poetics' in different cultural and linguistic spheres. R. Wellek (1969: 2) has called attention to the difficulties connected with the usage of this word in the realm of the Anglo-Saxon tradition. Not long ago V. Strada considered it necessary to comment on the use of this word in the Italian specialized terminology of recent years.2 A close connection between the word 'poetics' and works devoted to questions of 'I'arts poétique has prevailed in the Romance countries. Both in Slavic lands and in the German cultural sphere, on the contrary, the term has been used profusely and with many meanings. Dictionaries of the terms used in literary studies or even special works usually provide information about the details and the differences. In general, however, it appears that the very application of this word in the theory of literature has been increasing during the twentieth century, and it is perhaps one of the symptoms of a typical evolution in the field of literarytheoretical studies in our age. 1.2 The purpose of our essay is to trace the theory of the constitution of the literary work, the evolution of the study devoted to the devices and the means which realize a literary work. We are using the traditional term 'poetics' in this sense. We realize that it is a matter of approaching a work from a specific angle, that the problems of poetics conceived in this way are linked to a wider range of problems included under the concept 'theory of literature'. In a certain sense poetics is directly pervaded by these more general aspects of literary creation, for, as the Polish scholar J. Krzyzanowski says, ' . . . theory of literature . . . has . . . a considerably wider range: after all, it elucidates among other things the meaning of "poetics" itself' (1966: 31). On the other hand our conception delimits a specific area within the generalized literary work, for it deals with a fundamental aspect of creation - the concretization of the work. Its significance in theoretical study has been long since confirmed, because theory specifically comes into contact with artistic practice in this aspect. A retrospective analysis informs us that today we are far removed from that time when the discipline of poetics was sharply delineated and was completely isolated — a condition that was further strengthened by normative concepts and by the notion of the permanence of its categories. But the evolution of science, philosophy and, above all, the theory of art has paved the way for diverse theories concerning the work of art and has introduced new elements, new aspects into poetics and its object. Despite all the differences in these views, despite all the changes which historical evolution has brought, there remain stable elements that comprise the kernel of the field of poetics: certain principles, means and devices operate in creation, and the concretizations of a work consciously preserve and emphasize them, now modify them in various ways or even consciously violate them (in these 2

In a note to the translation of J. Lotman's article 'Metodi esatti nella scienza letteraria sovietica' (1967a : 128).

866

F. SVEJKOVSKÎ

cases, however, they are the background for this 'breaking with tradition'). This awareness of the existence of the permanent in creation is expressed wherever emphasis is placed on normative or regulatory categories, on objective principles, but it also appears indirectly in conceptions which primarily pursue the changeability, the individuality of expression, for even in these cases the standard categories of poetics are used for description and analysis. Moreover, this very problem of the relations between the permanent and the changeable is the energy that invigorates not only creation itself but also its theory, and an emphasis on one or the other is only one side of the coin—the literary work. Both sides will always attract attention, and their study will continue to reveal new aspects, just as literature will always continue to be new. These two sides will be an eternal problem; only the planes upon which they appear and the vector from which study approaches them will be different. They will constantly return in connection with the questions which Aristotle inscribed in the first sentences of his Poetics: 'how plots ought to be constructed if the making is to be done beautifully'.3 Thus those who maintain that poetics is in crisis, that it does not exist, or who simply end their considerations of it at some earlier terminus are both right and wrong. They are right insofar as they proceed from some model of poetics derived from its previous tradition.4 A previous tradition or a certain well-defined conception must not, however, be taken as an absolute standard - in this they are ill-advised. It is necessary of course to respect change and to see that in it the essence of the problems of poetics has survived, albeit in new specific formulations which are just as dependent upon theoretical formulations as upon artistic practice. And consequently, poetics as a field of enquiry has also survived. Finally, if we collect the evidence of the last decades, we find a multitude of variants: from the German terms Werkspoetik, Schaffungspoetik, immanente Poetik' to the international terms 'linguistic, structural, generative . . . poetics'. The word itself has also appeared in the titles of books and journals in various countries. So it is possible to ascertain that these questions not only exist, but that they have also been the object of considerable interest. We can even say that they have often been of the foremost interest to twentieth-century literary theory. 1.3 3

The initial impulses that were decisive both for the orientation and for the

Aristotle's Poetics, translation and analysis by Kenneth A. Telford (Chicago, 1961). Cf. e.g. Philippe van Tieghem and Pierre Josserand, Dictionnaire des littératures, t. I (Paris, 1968), p. 241: 'Depuis 1930, les études d'esthétique littéraire se sont multipliées, mais le triomphe, d'une part, du point de vue historique et de la notion de devenir, d'autre part, de l'idée de la liberté absolue en art et de la souveraineté du tempérament individuel a, presque partout, éliminé toute tentative de codification technique de la poésie.' Or cf. Hermann Pongs, Das kleine Lexicon der Weltliteratur, 5th ed. (Stuttgart, 1963), pp. 1341-2: 'Die Gegenwart, über Trümmern zweier Weltkriege, steht im Chaos, auch der Poetik. Der Aufspaltung von Form und Gegenstand entspricht die tiefere Aufspaltung des Dichters (bewusst-unbewusste Ambivalenz). Zeitausdruck: das Experiment. Muster der Dichtung gibt es nicht mehr. Wie soll man Urmuster des Schaffens gewinnen? — Der Begriff 'Poetik' schwindet ' 4

THEORETICAL POETICS IN THE TWENTIETH CENTURY

867

focus of the dominant tendencies in twentieth-century poetics can be summarized in a simple and seemingly self-evident premise: to approach the work as a phenomenon sui generis, as a specific expression of human creativity. This then led further to questions of its specificity. The roots of the problems that were revived in this way are not, however, only in the soil of literature itself; they penetrate beyond the limits of artistic practice and its theory. A s a matter of fact, they extend to questions of philosophy, science and art in their totality. A t the turn of the century many fundamental social values and ideas, including those of art, were in conflict. Wherever new ideas opposed the increasing one-sidedness of the empirical sciences, an effort to differentiate and specify many-faceted humanistic activity was launched. Thus it is no accident that it was precisely art which received due attention. In this connection it is sufficient to mention only two names: B. Croce, with his concept of artistic activity, and W . Dilthey, with his effort to define clearly the sphere of Geisteswissenschaften and to elucidate art as the creative transformation of life experiences. In opposition to a one-sided interpretation of art as the product of factors lying somewhere outside the area of artistic activity, in opposition to a mere reduction of the work of art to the sum of the elements relating to its content and to its form, in opposition to a mere description of details, it is precisely the autonomy of art, the complexity of the work, that is emphasized. Along with this, more general considerations appear: the relationship between the work of art in its uniqueness and the cultural, social context. Having thus been disturbed, the surface of art as a whole naturally found the specific response in the realm of literature. Today it is possible to refer to several informed accounts of the entire situation of literary theory in the twentieth century, particularly to R . Wellek's studies. Our theme leads us to view problems of literature from a narrow angle. A s it turns out, the poetics angle is such that especially important problems of twentieth-century literary theory are in the foreground, since the effort to obtain objective criteria for identifying the essence of literature and the specificity of its individual works was essential to the theoretist. 1.31 On the other hand, however, it is not possible, as mentioned above, to view this special position of poetics in twentieth-century literary theory without broader connections, because poetics has not been detached from the whole of theory of literature, as used to be the case up to about the eighteenth century. Historically it merged or emerged in various ways from the general field of literary theory, philosophy and aesthetics. Its practitioners' attention oscillated between general aspects of literature and the particular act of creation. Plato, with his philosophical considerations, and the Sophists with their interest in the technique of creation, had already opened up this extensive field, and the wise Aristotle had been seeking a balance between the two tendencies. What J. J. Murphy (1965) has traced in the tradition of rhetoric from antiquity to the fifteenth century can be applied mutatis mutandis with respect to the evolution of poetics up to the modern period: there have been conceptions of the Aristotelian type, just as there have been

868

F. SVEJKOVSKtf

programmatically oriented works which follow Cicero; there have been conceptions close to the grammatical tradition (Horace, Donatus and Priscian stand at the beginning of it), and finally, not even the Sophistic tradition has died out. Thus in our examination we, too, are aware of the fact that the problems of poetics are not only objects of special study but that precisely under the influence of the evolution of modern scholarship they constitute a component of differently conceived theories. An oscillation between poetics and literary theory is a characteristic feature of the period that we are examining; moreover, it is not only a matter of terminological questions but of conceptual questions as well. Poetics has become an important integrating component of literary-theoretical studies, but sometimes it has been assigned the mere role of summarizing 'formal means' in the spirit of older traditions. 1.32 Besides these relations to more general literary theory it is necessary to mention still others of particular consequence in the twentieth century: emphasis on the work, on its artistic quality, on its specificity has made prominent a particular notion of the criticism, analysis and interpretation of the work. Here questions of theory — and also of poetics — have more than once come to the fore. Although interpretation of the concrete work is the main aim, this activity has also frequently had a stimulating impact upon the field of poetics. 1.33 Definitions of poetics do more than point out the relationship between poetics and literary theory or criticism. Above all there is the question of norms, rules or the description of individual categories or means of literary creation. In the light of twentieth-century theories they have ceased to be constants or empirically gained characteristics. Their basic trait has turned out to be the dynamism derived from the conception of the work: a literary work cannot be understood as ergon but should be considered as energeia (allowing ourselves to employ Humboldt's contrast, which precisely in this period would often be revived in the works of literary theoreticians and linguists, and elsewhere too, as Croce's Aesthetics, e.g., indicates). The sources of this dynamism have been sought — in accord with individual conceptions — either in the creative act that concretizes the work of art or in the work itself as a specific form of communication. 1.34 This dynamic conception of creation as the realization of certain possibilities, tendencies and models rooted in the essence of the creative act has paved the way further for a new conception of the relationship between general poetics and an author's poetics (a distinction that is likewise often mentioned in definitions of the term). Theoretical poetics of the twentieth century spans the connection between these two spheres: an individual poetics (or the poetics of a group, of a movement) is nothing more than one of the concretizations of the potential that is offered to creation. Trubetzkoy's call for investigating not only what distinguishes phenomena but also what connects them could stand as a motto here. The relations between contemporary art and the evolution of literary theory or poetics directly support this. We can refer both to the Symbolist poets' attitudes toward questions

THEORETICAL POETICS IN THE TWENTIETH CENTURY

869

of theoretical poetics, for instance in France or in Russia, and to the relations between the Futurists and the theoreticians of Russian Formalism, which K. Pomorska (1968) has recently emphasized again. We can mention the role of modern art in the evolution of Czechoslovak Structuralism or the relations between contemporary French literature and the advancement of literary-theoretical studies in France today. And finally, let us go back to the turn of the century and substantiate the importance of these relations by referring to a completely different plane, to the work of W. Dilthey, who sees sure support for scientific theories in contact with the theories of writers and artists: Man darf die Ästhetik unseres Jahrhunderts nur nicht in den Kompendien und den dickleibigen Lehrbüchern suchen. Wie einst Lionardo und Dürer über ihre Kunst sannen und schrieben, wie Lessing, Schiller und Goethe eine neue Technik des Dramas und der epischen Dichtung durch die Verbindung von theoretischem Räsonnement mit Kunstübung schufen, so haben in unseren Tagen Semper, Schumann, Richard Wagner, die modernen Poeten der Franzosen wie der Deutschen, bei uns zumal Otto Ludwig, der tiefgründige Pfadfinder unserer neuen Richtung, sowie Hebbel, und bei den Franzosen von den Goncourts ab das ganze neue Dichtergeschlecht, ästhetische Spekulation mit dem künstlerischen Schaffen verknüpft. 5

1.35 The relationship between general poetics and different types of individual poetics does not mean, however, that a poetics linked only to a certain kind of creation, primarily contemporary creation, was thus being established. This initial relationship is only a basis for understanding certain general principles of creation in a synchronic and diachronic direction. The problems that stand out against the background of modern art provide the possibility of a confrontation with prior creation, and it is therefore possible to elaborate a historical poetics and to broaden again the basis for modelling and for typological study at the level of general theoretical works. Hence historical poetics will appear in a new light, while, on the other hand, the study of so-called immanent poetics, as B. Markward (1961: 233f.) has formulated it, will become actual.6 Today interest in an individual formulation of the theoretical aspects of creation is finding a response precisely in the theory of immanent poetics. 1.4 Therefore, it again appears that twentieth-century poetics is not new or modern in its object but in its methods and results, in which it has sought and has found its way both to the literary work and to its theory. In this manner we arrive at a historical view of this search and its results. Our aim in this essay is primarily to specify tendencies, not to provide an exhaustive enumeration of works and authors — to indicate the basic trends orienting poetics toward its new directions. 5

Wilhelm Dilthey, "Die drei Epochen der modernen Ästhetik und ihre heutige Aufgabe", Gesammelte Schriften, VI, p. 246. 8 On other aspects cf. e.g. Hans Blumenberg, "Sprachsituation und immanente Poetik" in the book Immanente Ästhetik, Ästhetische Reflexion, Poetik und Hermeneutik (München, 1966), pp. 145-155.

870

F. SVEJKOVSKY 2.

A S U R V E Y OF THE SITUATION U P TO T H E FORTIES

2.1 We have designated an orientation toward the work itself — rather, often toward the work in itself — understood as a specific manifestation of creative activity — as the initial gesture of poetics in the twentieth century. This, however, characterizes only the most general tendency which has attained various realizations. Above all, the emphasis on the work has allowed its two basic components to emerge again: the creator (who in a certain sense turns out to be not only the author but also the receiver, the reader, the listener) and the text. By means of these the very foundations of creation and the phenomena of its objectification have been sought in close relation with contemporary conceptions of literature and art. On this basis a differentiation of theories according to their relations to the two fundamental poles has resulted. At the beginning of the century theories that originated in Germany and Russia represent these two poles most conspicuously, whereas, on the contrary, in Italy B. Croce by his theory of 'intuitionexpression' abandoned the traditional questions of poetical categories from theoretical investigation. 2.2 An attempt to deal with older concepts characterized the beginnings of a new interest on the part of German scholarship in questions of poetics during the first decades of the twentieth century. This occurred primarily on those planes where the theory of poetics had developed during the nineteenth century: philosophy, aesthetics and philology. The new basis was found chiefly in psychology: 'Die Poetik hatte zuerst einen festen Punkt in dem Mustergültigen, aus dem sie abstrahierte, dann in irgendeinem metaphysischen Begriff des Schönen: und nun muss sie diesen im Seelenleben suchen.'7 The aim was to create a new, complete system. Meanwhile other conceptions tended for the most part toward partial solutions of the problems; a systematically constructed theory of poetics became a typical feature of German scholarship in the first half of the twentieth century. It is symptomatic that Wilhelm Dilthey, who participated in the formation of the foundations of twentieth-century philosophy and Geisteswissenschaften, chose poetics as a special object of his study. His returns to the problems of poetics from his first version called Die Einbildungskraft des Dichters: Bausteine für eine Poetik (1887) to his last formulations revealing his plans to rework the original book (1907 and 1908) belong almost symbolically to the turn of the century; they reflect a whole scale of problems which were raised for other studies of poetics. In accord with his philosophy he constructed a poetics at the author's pole. The objectification of the poet's experience (Erlebnis) is the main subject of his poetics, whereas problems of the concrete elaboration of the work, conceived as problems of 'poetic technique' (poetische Technik), recede into the background in the face 7 Wilhelm Dilthey, "Die Einbildungskraft des Dichters. Bausteine für eine Poetik", Schriften, VI, p. 126.

Gesammelte

THEORETICAL POETICS IN THE TWENTIETH CENTURY

871

of this. When Dilthey did deal with these in more detail, he approached them primarily from their relation to the poet. At the same time he clearly formulated the aim of poetics: '. . . die Poetik [soll] dem schaffenden Dichter nützen, das Urteil des Publikums leiten, der ästhetischen Kritik und Philologie einen festen Halt gewären.'8 From the viewpoint of the further evolution of poetics it is necessary to note his special emphasis on the role of the audience as an active component which participates in the concretization of the work. These are some of the main features of Dilthey's published ideas. If we have designated his work as symptomatic for the turn of the century, we have done so for yet another reason. In his unrealized plans of 1907/1908 for reworking his poetics, elements appear which indicate not only a substantial change in regard to Dilthey's previous ideas, but which also anticipate other important changes in twentieth-century poetics, primarily, with regard to its psychological conception which this great philosopher participated in creating and profoundly influenced. The publisher of Dilthey's writings comments on this plan that was never realized in the following words: ' . . . die Schrift [sollte] radikal umgestalltet werden,'9 and R. Wellek (1957: 121f.) has designated it as a rejection of his former psychological poetics. The plan of the new poetics testifies to a detachment from a one-sided dependency on the creator and to a transference of attention to the other pole — the work: 'Lösung vom Persönlichen', 'eine Terminologie, die nichts mit Empfindung, Gefühl usw. hat' — these are the significant signals of the unrealized change. Of course, these changes, just as the older poetics, are not without connections to the more general development of Dilthey's ideas. They are linked to an effort to overcome the difficulties which he himself found in his conceptions and which in his last creative period he sought to surmount by inclining toward hermeneutics. H. Diwald (1963: 123) says about Dilthey's progress on the groundwork of his theory: 'Deshalb hat auch Dilthey in seiner letzten Schaffungsperiode ausdrücklich versucht, von der Psychologie loszukommen und an ihre Stelle seine Lehre von der Hermeneutik zu setzen. Man wird darin die Einsicht erkennen müssen, dass das Werkzeug der Psychologie allein nicht die Aufgaben lösen konnte . . .'. So much from the standpoint of the whole system of Dilthey-the-thinker's works. The comparison to a landmark in the evolution of German scholarship is, we believe, justified from the viewpoint of poetics. It points from the older genetic and deterministic conception of poetics to the poetics of creation itself. At the end of his life this German philosopher was seeking an equilibrium between the poles creator-text together with an emphasis on the objective phenomena of the literary work, and thus he was paving the way for the study of an entire fundamental scheme of communication: author-work-audience. 8

Ibid., p. 126. Ibid., p. 310; the following citations are also to this work (pp. 310-313). The emphasis on a structural conception (Strukturpsychologie, Strukturzusammenhang) is characteristic; cf. R. Wellek, "Wilhelm Dilthey's poetics and literary theory" (1957 :121f.). 9

872

F. SVEJKOVSICf'

In order to define clearly the main tendencies of the German studies of poetics at the beginning of the century one has to mention several other works. The turn toward a psychological basis and, together with this, toward the author, resp. the audience, brought these studies close to Dilthey. H. Roetteken (1902; 1924) attempted to attain ideal models, e.g. precisely by means of the reception of the work. It is symptomatic how emphatically most of these German works of this period called attention to their polemical relation to prior tradition and at the same time set as their goal a 'scientific poetics' (a formulation greatly used by these authors). In particular, Richard Müller-Freinfels with his Poetik (1914) demonstrated the possibility of approaching the fundamental principles of creation on the basis of contemporary psychology. In his conception poetics deals with the psychological preconditions of the existence and functioning of the work, and hence the center of gravity of creative activity lies therein. The abandonment of the former descriptive or regulatory poetics is so emphatic that the reality of the text has a place in this system only through the mediation of its highest unifying element — style. He understands this as 'Gesamtheit [or elsewhere, Einheit] der Eigenschaften des Kunstwerkes' (1914: III). Another feature of his 'scientific conception of poetics' is his renunciation of 'Normen aufzustellen und ein ästhetisches Gesetzbuch zu liefern' (1914: III). In German theoretical poetics of this period the problems of style had become the widest bridge connecting different speculative systems with the work in its concrete linguistic appearance. Here we also encounter one of the roots of stylistics which will have a substantial impact upon the solution of the problems of poetics, especially whenever consideration for the work as a linguistic expression comes to the fore. In this area of German scholarship two men played a particularly significant role — Theodor A. Meyer with the book Das Stilgesetz der Poesie (1901) and Karl Vossler with his broadly based studies. Meyer's views comprised a counterbalance to those tendencies that had in various ways estranged the problems of poetics from its very object — the work of art. Response to his work in the context of German theory was great. Even in 1928 (p. 119) R. Bosch wrote: 'Meyer bezeichnet die Dichtkunst als die Kunst der sprachlichen Vorstellung. Etwas Allgemeineres über das Wesen der Dichtkunst ist wohl niemals gesagt worden . . .' (which could, of course, have applied to the German situation at that time). For Meyer style becomes the central theme: '. . . ich möchte den Stil der Poesie, d.h. die besondere Art, in der sie in ihrem Teil der gemeinsamen Absicht der Kunst gerecht wird, ableiten und verständlich machen aus der Natur der Sprache und der sprachlichen Vorstellungstätigkeit als dem dafür massgebenden Gesetz' (Meyer 1901: IV). The author's polemical direction of his work results in his handling the question that he has posed primarily in relation to aesthetic and psychological theories (he develops his theory against the background of the Lessingian tradition). Thus he does not probe more deeply into the structure of the literary work either. K. Vossler's presentation reveals how scholarship of that time was approaching

THEORETICAL POETICS IN THE TWENTIETH CENTURY

873

the problems of stylistics in various ways — in this case, directly on a linguistic basis. His position along side of the Saussurian tradition, represented by Bally, has often been elucidated. In the field of poetics at the beginning of the century this position no doubt played a significant role that extended beyond the borders of Germany into the Romance countries. Vossler's relation to Croce led him to emphasize aesthetic aspects and at the same time to promote consideration for the individual features of the work. Therefore, his theory, too, created preconditions for reconciling the spheres of poetics and interpretation of individual works. He remained faithful to the German tradition by virtue of the fact that he proceeded from a psychological basis, and from that basis he sought a way to discern more general national features through the mediation of an individual style. Still another German work that cannot be overlooked — Rudolf Lehmann's Deutsche Poetik (1908) — indicates the intensity of the search for a new 'objective' or 'scientific' poetics. This book also has its source in previously published works; and immediately in the introduction it renounces both the former 'classificatory poetics' and the contemporary psychological orientation. However, it does derive from impulses of psychological poetics as far as the author's role and the functioning of the work concerned (cf. the German terms Schafjungs- and Wirkungspoetik). It also addresses itself to the stimuli of aesthetics, for the literary work interests him primarily as a work of art concretized by means of language. Consequently this book represents the point of view that directs attention to the work itself. Its approach to poetics acquires significance chiefly in light of later trends. Let us quote its characterization of poetics: 'Die praktische Regel wird für sie zum theoretischen Gesetz in dem gleichen Sinne, wie die wissenschaftliche Betrachtung der Sprache hinter der grammatischen Regel das wirkende Gesetz erkennt. Diese wissenschaftliche Poetik als solche hat kein unmittelbar praktisches Ziel mehr, so wenig wie die wissenschaftliche Erkenntnis überhaupt: sie will weder dem Dichter, noch dem Kritiker dienen; sie will nur Einsicht in das Wesen der Dichtkunst sein' (1908:2). His effort to disengage the work from a one-sided dependency on the author and to emphasize that the work lives an independent life as soon as it has been written is decisive. Thus the object of poetics is not 'das Subjekt des Künstlers, sondern das objektiv vorliegende dichterische Kunstwerk' (1908:42). We cite yet another passage which demonstrates how Lehmann was able to discern problems that still exist in present-day studies devoted to theory of literature and poetics. On the subject of poetics he remarks: ' . . . es handelt sich nicht um eine blosse Technik, um eine Systematisierung der äusserlichen Gattungen und Formen, nicht um Einteilungen und Aufzählungen, sondern um die Festellungen der Gesichtspunkte nach denen die Dichtung als Kunstwerk von innen heraus und ihren eigenen immanenten Gesetzen gemäss zu erfassen ist, und um das Verständniss der dichterischen Form, soweit sie organische, d.h. lebendige, von innen bedingte Gestaltung ist' (1908:42). Characteristic are his evaluation of T. A. Meyer's book and his reference to W. v. Humboldt. In a new way he raises —

874

F. SVEJKOVSKf

even if he does not solve — questions of the relationship content-form. He arrives at a functional conception ('die Poetik [wird] die Einheit des dichterischen Kunstwerks immer nur als eine gewollte und beabsichtige begreifen können' (Lehmann 1908:48)). He poses the question of the permanent and the changeable in the work of art and finds the point from which it is possible to elucidate the impact and the value of the work: it is the author's intention which meets with the reader's experience and affects him. The author has an opportunity to acknowledge the influence of Dilthey's theory of experience (Erlebnis). Emphasis on the artistic aspect of the work, then, orients Lehmann's poetics toward aesthetics: poetics is a special sphere of aesthetics. The aesthetic aspects and mainly a consideration for the linguistic realization of the literary work substantially distinguish this book from other contemporary types of German poetics. The paths which the further evolution of German theory followed were not, however, to be found in Lehmann's work but in the stimuli of psychological poetics and especially in Dilthey's work from which the characteristic German Geisteswissenschaft developed. This, in turn, influenced in various ways particular fields of the theory of art. 10 The Geisteswissenschaft orientation led to speculative literary theory, on which questions of poetics acquired more the appearance of a descriptive and classificatory cataloguing of means or forms. Emil Ermatinger's work Das dichterische Kunstwerk (1921 and republished in a revised edition at the end of the thirties) above all is evidence of further progress in the attempt at a systematic elaboration of problems in the spirit of the Diltheyan tradition. Besides this there was Julius Petersen's book Die Wissenschaft von der Dichtung (1939), which in its conception and method testifies to the crisis of systems stemming from the old Geisteswissenschaft traditions of German scholarship. Rudolf Bosche's Die Problemstellung der Poetik: Eine historische Untersuchung über die Methoden und Grenzen wissenschaftlicher Wertbestimmung (1928) is a continuation of the attempts at a psychological poetics of the 'scientific type' (which the author distinguishes from 'wertsetzende Poetik'). This book is above all evidence of the penetration of experimental and statistical methods into studies of poetics. Oskar Walzel introduced new motivations, not entirely traditional in the field of German poetics. Under the influence of studies in the history of art he revived the theory of styles (a reference to H. Wölfflin's works) and restored equilibrium to poetics by formulating in a striking way the question of the relationship contentform. Such was the case especially in the work Gehalt und Gestalt im Kunstwerk der Dichtung (1925). He proceeded to reevaluate the traditional dichotomy contentform and substituted the opposition Gehalt-Gestalt for it, whereby he revived a number of questions of poetics traditionally linked to the notion of form. 11 10 On the whole set of problems concerning Dilthey's views on literature and the response to them cf. Kurt Müller-Vollmer, Towards a phenomenological theory of literature (The Hague, 1963). 11 Cf. the critical comments on Walzel's conception of Gehalt-Gestalt in R. Wellek, "Concepts of form and structure in twentieth-century criticism" (1963 : 63).

THEORETICAL POETICS IN THE TWENTIETH CENTURY

875

Emphasis on the expressive means of literature found a response in his 'höhere Mathematik der Form'. These means do not appear to him as a priori, ready-made forms realizing content, but as possibilities that acquire shape only in contact with a concrete 'content'. Walzel's theory converges with other theories of twentiethcentury poetics precisely in this changed attitude toward the means and the categories of the older poetics. He himself, however, did not work out a systematic poetics. He did not elaborate the entire system; he only touched upon more than one problem along with other questions of literary theory. Similarly, he participated in the expansion of the aspects of literary stylistics, for next to linguistically oriented studies and interest in individual style he placed the study of the typology of styles, which stems from a comparative view of various artistic fields. His theories had an influence not only in the German cultural sphere but also in other countries; they elicited other works and critical comments. From the viewpoint of poetics one has to state that his conception was an important counterbalance to the Geisteswissenschaft orientation, because it stressed the role of formal means, the role of the 'mathematics of form'. Walzel accordingly revived the tradition of German aesthetic formalism (Herbart,12 Dessoir), which at that time was also finding responses elsewhere, especially in Czechoslovakia where Herbartian aesthetics had a long tradition. At the end of the thirties and the beginning of the forties Robert Petsch developed an approach to the literary work in general and, above all, to questions of poetics as the basis of literary studies. He did not view the problems of poetics either in description or in rules but in the detection of the essence of categories and forms and in the ways in which this essence is realized; hence a poetics tending toward typology. Petsch sought the essence of these forms in the spiritual domain of human life, in certain Grundeinstellungen, and these, if they are projected onto the level of literature, are the starting point for its division into genres. Petsch did not work out a total Literaturwissenschaft but essentially a poetics of the epic, the lyric and the drama — 'die Lehre von den dichterischen Formen', as his pivotal works Wesen und Formen der deutschen Erzählkunst (1934) and Wesen und Formen des Dramas (1945) indicate. Thus he too studied the individual components realizing 'form' in relation to the basic genres, to their essence (Wesen). Consequently, forms ceased to be a mere instrument and poetics a set of technical means. He formulated the significance of this approach to the literary work and to the tasks of its theory as follows: 'Die Aufgabe der allgemeinen Literaturwissenschaft ist es, diese verborgenen Wege zwischen dichterischer Schau und Auffassung, Aufbau und Stil immer wieder zu beschreiben, um das Einzelne aus dem Allgemeinen, die Erscheinung aus dem Wesen zu erklären' (1940:27). At the same time a direct theoretical substantiation for dividing poetics into single fields can be seen in Petsch's works. He himself realized the study of literary 12

Walzel published a study on Herbart: "Herbart über dichterische Form", Zeitschrift Ästhetik und allgemeine Kunstwissenschaft 10 (1915), pp. 435-454.

f.

876

F. SVEJKOVSKÎ

genres (genology). Aside from stylistics it was precisely this area of studies that would be the most productive, especially when it came to allocating interest to a separate study of genres. As early as the twenties works appeared in Germany which focused attention toward these questions by means of psychological theory and which attempted to anchor individual genres of creation in the spiritual domain of life on a psychological basis. For example, there had been E. Winkler's work Das dichterische Kunstwerk (1924) and R. Hartl's Versuch einer psychologischen Grundlegung der Dichtungsgattungen (1924). In this connection a special place in the studies of genres fell to André Jolies' work Einjache Formen, Legende, Sage, Mythe, Rätsel, Spruch, Kasus, Memorabile, Märchen Witz (1930), for, on the one hand, it dealt with problems and tendencies of its time, but, on the other hand, it followed a solitary course in the German sphere of research at that time. Jolies intentionally turned to the simplest forms of verbal expression in order to attain an understanding of its essence. Through his method he paved the way for an anthropological interpretation of a literary work. With his typology of genres he simultaneously refers to the ties with linguistics. He himself acknowledged Goethe's Morphologie, which just at that time was producing possibilities for formalist tendencies in German literary studies and was also providing the impetus for a more extensive incorporation of these ideas into the process of human and natural activity. The studies of the German theoreticians and historians of literature G. Müller and H. Oppel were further proof of the inclination toward the Goethean tradition. As we have already discussed special aspects in studies concerning poetics, we must once again turn to stylistics. Since the twenties Leo Spitzer was the leading personality in this area. He recapitulated his attitude toward the older, Vosslerian tradition in the lecture "Les études de style et les différents pays" (1961: 23f.) which he gave at the end of his life. In it he once again distinguished his conception from that of de Saussure and Bally. In the studies of stylistics which were developing he represented the fusion of a linguistic and a literary interest, and at the same time he emphasized a consideration for the function of the individual and his expression. Certainly the basic aspects of Spitzer's theory are a significant contribution of a psychologically-oriented stylistics to the study of the literary work. He converged with the development of poetics in that he placed the work itself at the center of attention, in that he revealed the dynamism of the individual means and categories of poetics; but his works gravitated toward practice, toward interpretation rather than toward the solution of theoretical problems. Moreover, he asserted most of his influence — from European Romance studies to American stylistic criticism — in these areas. After all, even his theoretical deliberations head primarily in this direction, as is apparent, for example, from the essays in Linguistics and literary history (1948) or from A method of interpreting literature (1949).

THEORETICAL POETICS IN THE TWENTIETH CENTURY

877

2.3 The Russian Formalists approached questions of poetics in substantially different ways than German theoreticians did, and they introduced an orientation that would become determinative for a wide range — truly world-wide in scale — of studies. We can find a characterization and evaluation of Russian Formalism in the development of literary theory as well as its detailed history in more than one work of recent times, especially in V. Erlich's (1965=1969) and K. Pomorska's (1968) books or in J. Striedter's (1969) comprehensive analytical study which introduces the edition and translation of Russian Formalist texts. We can merely make reference to these as well as to some other studies, for we are not concerned with the analysis of the whole formalistic theory but with seeing this component in the context of the global development of poetics. As the subsequent state of research has shown, the theories of the Russian Formalists were consequential not only for the development of the theory of the literary work in Russia itself but substantially influenced — and are still influencing — other work in this area. Their basic significance lay especially in the fact that they consistently stressed the autonomy of literary creation in opposition to traditional views and further in the fact that they introduced a poetics based on linguistics. Stemming from the evolution of Russian linguistics and its relation to poetics (A. Potebnja), they combined this interest with an inclination toward contemporary artistic tendencies, especially literary tendencies (Futurism). And in a favorable climate of modern linguistics (Baudouin de Courtenay, L. Scerba, F. de Saussure, E. Sievers' school), science and philosophy (esp. responses to Husserl's phenomenology) on Russian soil the members of the new generation constructed their theory, their approach to literature. They created nothing more than thenown special variant of an approach to art as a specific field of creativity which was being revived at that time. It was a rational approach that sought to rid artistic creation of the haze of Romantic mythicality, but at the same time it desired to restore to art the individuality that it was losing in the 'scientific interpretations of the nineteenth-century'. For this reason was the work itself — specifically the work in itself — promoted so emphatically. Hence the stress on the autonomy of literature, on its intrinsic immanent laws. The literary work was conceived as a specific kind of linguistic communication that is examined in its concrete appearance — the text of the work. Therefore, its language, poetic language, became the center of study. Jakobson's (1910:21): 'Poetry is nothing but an utterance oriented toward expression' stands out here beside the formulations of V. Sklovskij and L. Jakubinskij. It became the starting point for his conception of linguistic poetics. In his own poetics Tomasevskij sums up his view of literature with the words: 'Literature is a particular, fixed utterance' (1927:4). And Tynjanov says: 'Literature is the dynamic construction of an utterance' (1929 = 1971:87). The essence of the artistic act was sought in a linguistic realization based on specific means and devices (priemy). Here concentration on the work was consistent to such a degree that the aspects of both the author and the audience disappeared; in other words

878

F. SVEJKOVSKY

they were absorbed by the essence of the work itself, by its 'literariness'. 'The object of literary studies is not, therefore, literature but literariness, that is, what makes a given work a literary work' (Jakobson 1921:11). 'OPOJAZ declares that there are no poets and literati — there is poetry and literature. Everything that a poet writes has value as a part of his contribution for the common cause, and it completely loses its significance as an expression of his "ego" ' (O. Brik 1929:213f.). If in the Russian Formalists' theory the work of art was understood from the viewpoint of the text, from the viewpoint of a linguistic realization, poetics above all found its justification and a large field in this domain. V. Zirmunskij formulated this as follows: 'Poetics is a science concerned with poetry as art. If we accept this word, which has been illuminated by tradition, we can boldly claim that in recent years literary studies have been developing in the name of poetics' (1928:17). Thus poetics encompassed the total literary work, the thematic level as well regarded from the viewpoint of expression. Sklovskij's statement 'the literary work is form' was the motto of this theory, and it bridged the traditional dichotomy of content and form. In the same way questions of aesthetics and the historical development of literature were solved on this basis by references to the constant overcoming of the automatization of means through their transformation. Yet, it is necessary to say that both the traditional problems of content and of aesthetics of the literary work were not elaborated. The following statements by Èjxenbaum help clarify the question of historical interpretations: 'Time in history is a fiction. We do not trace evolution in time but evolution as such, a dynamic process' (1924:8). The main interest was concentrated on the text, on poetic language, and therefore on poetics. 'The task of general or theoretical poetics is the systematic study of poetic devices [priemy], their comparative description and classification . . . . Inasmuch as the material of poetics is the word, the basis of a systematic organization of poetics must be the classification of linguistic facts, which is what linguistics offers us' (1928:39). Since attempts to develop a general poetics were made, Zirmunskij's conclusion was realized. His essay "The tasks of Poetics", divided into Stylistics, Thematics and Composition, indicated this as early as 1921. In working out individual means and devices, he came to conclusions about the interpénétration of thematics and composition, e.g. in his exposition of the pair plot-subject [jabula-sjuzet]: 'the problem of the selection of means belongs to the sphere of thematics', whereas their organization is the object of composition. Genres likewise belong to the area of composition, for Zirmunskij sought their essence primarily in an individual, fixed compositional frame. B. Tomasevskij's book with the characteristic title Teorija literatury: Poètika (1927) was another general work. In the introduction the author defines and characterizes poetics: 'The object of poetics (in other words, the theory of verbal art or literature) is the doctrine of the ways of composing literary works. . . . Literature or verbal art, as the designation that has just been introduced indicates, is a component of man's verbal or linguistic activity. From this it follows that the theory of

THEORETICAL POETICS IN THE TWENTIETH CENTURY

879

literature adheres to the doctrine dealing with language, that is, to linguistics' ( 1 9 2 7 : 3). Tomasevskij divides poetics into three parts: between Stylistics and Thematics (which includes the questions of subject [sjuzet] and literary genres) he places Comparative Metrics together with the fundamental questions of prose as well as the problems of verse. The difference between 'fixed' and oral expression or folklore is especially emphasized both in his definition of literature and in his further deliberations. We find the differentiation between poetic language and the spoken language expressed particularly in the parts devoted to stylistics, where he considers the specificity of the function of language and distinguishes 'poetic lexicon', 'poetic stylistics' and 'euphony'. In general, works of such a synthetic character are isolated in the Russian Formalists' writings, and it is therefore no accident that Russian Formalism has been treated as a specific trend (even if internally differentiated) from which no complete systems but only works dealing with separate questions came forth. The need to elaborate the results achieved by returning to individual problems was juxtaposed with the effort to synthesize. The other aspect of this situation that differs from that in Germany is the reconciliation of the theory and the concrete analysis and interpretation of the work. We may quote B. Bjxenbaum's two articles on the problem of skaz as an example of such an oscillation: he dealt with this question first in his study of Gogol's "The overcoat", then in a special theoretical paper. He wrote: 'A program is a very handy thing for critics, but not at all characteristic of our method. Our scientific approach has had no such prefabricated program or doctrine, and has none. In our studies we value a theory only as a working hypothesis to help us discover and interpret facts; that is, we determine the validity of the facts and use them as the material of our research. . . . There is no ready-made science; science lives not by settling on truth, but by overcoming error' (1927 = 1965, 102-103, 139). Perhaps these statements clarify the situation sufficiently and also explain why the most significant research is found in collections of essays or anthologies with titles like Poetika. Here, too, we find an answer to the question of the relationship between the theory and the analysis of the work: in order to discover the former the Russian Formalists dealt with the latter. From the outset they were concerned with the study of the individual levels of the literary work, above all with the rhythmic, sound aspect, and in some cases with the relationship between certain levels, e.g. between verse and syntax. Although interest in poetry prevailed, studies devoted to prose also appeared (Sklovskij's essays O teorii prozy, 1925 and 1929; fijxenbaum's or Tynjanov's works are important evidences of this fact). Sklovskij's formulation, the work of art is 'the sum of its means', soared above the Russian Formalists' initial studies and reflected a conception of poetics. It was a polemical gesture introducing valuable new impulses into previous study: yet at the same time it was the basis for certain simplifications and one-sided tendencies that subsequent development and practice itself overcame. Accordingly, poetics was the field of study of the individual components of the literary work, of the individual

880

F. SVEJKOVSKi

means of its linguistic realization with a specific function — the poetic function. From about 1920 to 1925 a further elaboration of the initial theories and a differentiation of views appeared, f o r — a s has been emphasized more than once — Russian Formalism was not a uniform movement. It began to move from individual components toward the whole, to reveal the inner dynamics of the work and thereby to overcome a mere parallelism of these components. The perspective of the modern notion of a linguistically oriented poetics was also included in this. Unifying aspects of the work began to be sought. Sklovskij and Ejxenbaum introduced the theory of the 'dominantä that is formed in relation to the individual components and expressive means of a work. Zirmunskij discussed the unifying principle of the work as a whole, which he expressed by means of his theory of style. At the same time Vinogradov formulated a theory of style from a different angle — on a linguistic basis. A functional conception, developed especially by Tynjanov, emerged. Finally, it was important for the evolution of the theory of the work and for the development of poetics that the question of the role of linguistics with regard to the literary work was raised. Specific aspects of literary studies (fijxenbaum, Zirmunskij) that focus attention on phenomena extending beyond the framework of a linguistic conception were simultaneously juxtaposed to these linguistic notions. When Zirmunskij summarized this study in his report on the Russian Formalists' works in Zeitschrift für slavische Philologie (1925) he placed emphasis on problems connected with the interpretation of the level of plot. In this manner the Russian Formalists' theory was worked out and simultaneously differentiated; their special contribution in the areas of versification, theory of genres, theory of prose and stylistics appeared along with their general expositions of poetics. The elaboration of stylistics was not, however, of major concern in Sklovskij's group; it proceeded rather from those scholars who sometimes travelled roads different from the representatives of the initial positions of Russian Formalism. We have already mentioned V. Zirmunskij and V. V. Vinogradov. These two scholars had in common an emphasis on the central position of stylistics in the interpretation of the work as a whole. Whereas Vinogradov worked out a stylistics primarily in relation to the role of language in the literary work (among other things his stimulating approach to the work from the viewpoint of the relationship monolog-dialog), Zirmunskij paved the way for viewing the work in a broader context even beyond its boundaries with his notion of style as the unifying principle in art. It is indicative that this is exactly where Zirmunskij demonstrated the distinctiveness of his views from certain theses of Russian Formalism. He also expressed this attitude in his introduction to the Russian translation of a representative work of German 'Formalism' and of stylic studies in the broader context of art — O. Walzel's Die künstlerische Form des Dichtwerks (1923). Toward the end of the twenties it was apparent that the previous development of the theories of Russian Formalism was culminating and that important transformations of them were now taking place. These transformations were precipitated,

THEORETICAL POETICS IN THE TWENTIETH CENTURY

881

on the one hand, by the effort to reevaluate or elaborate previous results and, on the other, by the criticism that stemmed from the principles created by Russian Formalism which in many cases was already merging with critical expressions from outside sources, from different literary-theoretical positions or even from the more general positions of the new tendencies in Marxist scholarship. We can therefore make mention of psychologically or sociologically oriented theories. At that time primarily the latter had become the area into which the influences of Marxism were penetrating. In the twenties we find, e.g. the works of P. N. Sakulin (K voprosu o postroenij po&tiki, 1923; Teorija literaturnyx stilej, 1927), G. N. Pospelov(Problema teoreticeskoj poetiki, 1928) andV.M. Frice (Problemy sociologiceskoj poetiki, 1926). However, as we have already said, the development of the stimuli of prior research was occurring at the same time, and important results indicating the future development of theoretical poetics, in other words, directly paving the way for it, were appearing. Today such a role has already been accorded to the JakobsonTynjanov theses called "Problemy izucenija literatury i jazyka" (1928) which is considered the decisive step in the ingress of literary theory into the field of literary structuralism. In this sense they are the culmination of the theories which J. Tynjanov in particular had represented in the development of Russian Formalism but which had also been asserted elsewhere, e.g. in G. N. Pospelov. If we project the ideas of the Jakobson-Tynjanov theses into the domain of poetics, we can see a decisive rejection of any kind of 'cataloguing' of phenomena (cf. point 5) in referring to their conditioning by ties and functions in the structure of the work as a whole. In the same way they surpass the limits of the earlier theories of Russian Formalism by emphasizing two main aspects in their approach to each phenomenon: the relationship between a system of norms and an individual phenomenon (here the background is the Saussurian opposition langue-parole; cf. point 6). Similarly, the aspects of synchrony and diachrony would become consequential for poetics, especially from the viewpoint of historical poetics (cf. point 4). In general, the theories of Russian Formalism revealed many problems of twentieth century poetics which other contemporary theories reached only when they left the realm of the older regulatory or descriptive poetics. This is a conformity characteristic of the historical situation. It is a conformity in the raising of certain actual problems of creation and its theory; however, the solutions entail differences and differentiations in the theories, as has already been shown and will be shown again. Besides the Jakobson-Tynjanov theses the works of Zirmunskij and Vinogradov represented further important steps, above all in the new aspects introduced into stylistic studies and in the effort to view a work in a context broader than that of the traditional interpretation of the work as a specific linguistic phenomenon. Scholars, who are today designated belonging to 'Baxtin's school'1S and G. O. Vinokur, among others, represented both the surmounting of this framework of literary 13

On this subject cf. A. A. Leont'ev, Psixo-lingvistika

(Leningrad, 1967).

882

F. SVEJKOVSKtf

theory, in some cases a static conception of poetics, and the further developing of previous theories in various directions. Recently Soviet scholars who acknowledge the traditions of Russian Formalism have even recognized the importance of S. M. Eisenstein's works.14 Whereas one can see certain connections with the development toward structural poetics in Eisenstein, G. O. Vinokur, on the other hand, devoted attention to the relations between linguistics and poetics and criticized Formalist conceptions which had narrowed down questions of poetics to the problem of the 'technology of poetic creation'. A consideration of the scholars belonging to 'Baxtin's school' shows how Russian Formalism developed in various ways. P. N. Medvedev attempted to incorporate theory of Russian Formalism in his work Formal'nyj metod v literaturo-vedenii: Kritiieskoe vvedenie v sociologiceskuju poetiku (1928). Moreover, he was establishing a basis for coalescing with the 'ideological' criticism of Russian Formalism. In his linguistic studies V. N. Volosinov contributed to the expansion of research, particularly in his study of the problematics of the sign. M. Baxtin's interest lay directly in the field of literature. His development of the theory against the background of Russian Formalism shows again that the move toward structuralism in Russian scholarship was significantly motivated by the tradition. The importance of Baxtin's book Problema tvorcestva Dostoevskogo (1929) is apparent not only from the viewpoint of the situation during the critical years of Russian Formalism but also in the context of current study which is again calling attention to it and which values its theoretical foundation as well as Baxtin's contribution to previous literature on Dostoevsky's works. The recent second edition of this book has a symptomatically modified title: Problemy poetiki Dostoevskogo. This immediately suggests a relationship with the theoretical problems we are concerned with here. In this study Baxtin's consideration of poetics and more general theoretical aspects pervade each other, and in this respect he approximates the practice of Russian Formalists who likewise did not begin from speculative premises but from problems which the analysis of concrete works revealed. He comes close to the approaches proceeding from the problems of the stylistic interpretation of the work, hence to the conceptions of the Leningrad scholars, Zirmunskij and Vinogradov. For Baxtin a linguistically oriented study is a springboard for reaching the specificity of literary expression, since he, too, distinguished linguistics from a poetics based on it. Characteristically he proves by applying certain unifying aspects of the work, basic structuring principles and devices which are defined by their relations to a context wider than the work itself: they are rooted in artistic tradition and in social situation. In the case of Dostoevsky's works Baxtin finds this unifying principle in 'polyphony' or in 'dialog'. He elucidates his conception of these terms from a historical viewpoint, on the one hand, and by a structural analysis of the text, on the other. As a result, he constantly moves into the field of theoretical poetics. He surmounts the older Formalist conceptions of the confrontations of the 14 Cf. A. K. Zolkovskij and Ju. K. Sieglov, "Iz predystorii sovetskix rabot po strukturnoj poetike" in Trudy po znakovym systemam, III (Tartu, 1967).

THEORETICAL POETICS IN THE TWENTIETH CENTURY

883

synchronic and diachronic approach to individual aspects of poetics, and at the same time he makes the concrete work the center of attention as a specific sign incorporated into the continuity of the process of communication and dependent on a historical context. Baxtin thus arrives at a concept of poetics as the theory of the concrete work. What interests him is precisely the process of concretization and the dynamics of both the permanent (tradition) and the changeable (the individual realization of the work) within this process. In many respects these ideas suggest an affinity with the development of Prague literary structuralism, especially since it culminates the end of the third and the beginning of the fourth decade of the twentieth century. In regard to other aspects of poetics it is necessary to mention in particular Baxtin's contribution to the study of genres. If we survey the Russian situation at the end of the twenties and the beginning of the thirties as well as the developmental stimuli found in prior studies, we must call special attention to one more trend — the investigation of oral verbal art. A special contribution of Russian Formalism was the incorporation of analysis of folklore, which at the time was more likely to be differentiated from the study of 'artificial' or 'artistic' literature. To the Formalists, on the contrary, folklore was closely related to literature and became the starting point for an elaboration of various questions of theory and interpretation. Some basic forms and devices of verbal expression were discovered in oral verbal art, and therefore further possibilities of penetrating to the essence of verbal expression were demonstrated. Such a study did not lack a tradition in Russian scholarship, especially after A. N. Veselovskij's appearance. In the succeeding development of folklore research the 'morphological' approach to creation crystallized in close relations with literarytheoretical problems. Its special significance for the study of folklore is substantiated above all by V. J. Propp's representative work Morfologija skazki (1928), which belongs to the period of the culmination of Russian Formalist traditions. Its inclusion in the series called Voprosy poetiki also shows its close relation to the literary-theoretical thought of that time. Its title suggests an inclination toward contemporary tendencies, to the 'morphological' approach which concentrates on the work in itself and strives to specify the individual components of a work.15 In addition to a Formalist analysis, a significant differentiation of the work in its linguistic concreteness from questions of its aesthetic value characterizes the conception of Propp's book from the viewpoint that we are pursuing in this essay. Here we are reminded of a problem that was often to come to the fore in literary theory from the thirties onward under the influence of aesthetic theories, especially phenomenological theories (Ingarden, Mukafovsky, Dufrenne). Sharp criticism of an ideological nature was decisive in interrupting the further development of the theories attained in the U.S.S.R. by the end of the third decade. 15

Let us mention Propp's introductory formulation in this work: 'The word "morphology" means the study of forms' (quoted according to the English translation, Morphology of the folktale, 2nd ed., Austin, 1970).

884

F. SVEJKOVSKtf

In the field of theoretical poetics this criticism resulted not only in the suppression of tendencies represented by the differentiated views thai had arisen from the basis of Russian Formalism but also in the total abandonment of problems of the 'form' of the literary work. The actual functions of literary criticism, conceived primarily as important components of a new, developing social ideology — and the striving for a Marxist theory of literature — gained prominence. In particular, sociological and ideological aspects were asserted in theory. V. F. Pereverzev and later I. A. Vinogradov represented these tendencies. Moreover, attention was directed rather to historical aspects of study. Interest in 'how the work is made' appeared primarily in the study of the cognitive functions of literature and their relation to reality. Specific problems of poetics most often appeared in connection with the study of an author's 'language' or style, and they were again introduced with regard to the ideological or social basis of a work. Thus 'theory of literature' took the place of poetics; in this framework a type of descriptive or regulatory poetics was revived (mostly in works of a schoolbook character). Only in some works that have been published since the end of the fifties has there been a revival and a new critical evaluation of the Russian Formalists' theories. Thus some tendencies which the expansion of research initiated in the period before the thirties received belated responses in the very country where they had originated, whereas investigations in other countries had already analyzed and elaborated them earlier. 2.4 The theories of Russian Formalism received their first international response in the Slavic lands of Czechoslovakia and Poland. We encounter an especially intense and early response in Czechoslovakia. There knowledge of Russian Formalism spread not only in response to its works but also through contact with some of its representatives — R. Jakobson, N. S. Trubetskoy, P. Bogatyrev, S. Karcevskij — who worked in Czechoslovakia for a shorter or longer time and who participated in the further development of linguistic, literary and folklore studies. A special base was soon established for this contact and for the expansion of research — the Prague Linguistic Circle (1926), an analogue of the Moscow and Leningrad associations (Vachek 1964, 1966). It suffices to mention only a few of the names of those who lectured at the meetings of the Prague Linguistic Circle to gain an understanding of its place and its impetus in the field of literary studies: in addition to the abovementioned scholars, E. Husserl, R. Carnap, E. Utitz and B. Tomalevskij, e.g., attended meetings. Of course, there were also local representatives of this effort to find the new paths of modern science — the science of literature and linguistics in particular. The Prague Circle's relations to Russian Formalism have already been sufficiently elucidated in specialized literature. Less known however, is the fact that the possibilities of coalescing with this trend and directly developing some of its tendencies in 'Czechoslovakian Structuralism' had been prepared by local developments and by contact with the rest of contemporary European scholarship. Furthermore, Czechoslovak Structuralism is known mostly in its initial phase,

THEORETICAL POETICS IN THE TWENTIETH CENTURY

885

in its development in the Prague Linguistic Circle. But from the end of the thirties a conspicious part in this movement was also played by ? Slovak center of study; it was connected with the Czech traditions primarily through the person of J. Mukafovsk^ who generated interest in the new trends in literary studies at the University of Bratislava.19 2.41 The Russian Formalists' theories also attracted the interest of Polish scholarship. In Poland it was not a question of a mere transference or an accidental development of foreign stimuli with which responses to the work of the Prague Linguistic Circle were soon associated. At that time Polish scholarship was keeping pace with European developments and with German studies especially. Husserl's works had a special impact upon literary studies in Poland. There is no doubt that precisely this basis paved the way for new currents and literary-theoretical conceptions. A very early testimony to this fact is Zygmunt Lempicki's programmatic essay 'W sprawie uzasadnienia poetyky czystej" (1921) which adresses itself directly to the new problems of poetics. Lempicki sought possibilities for surpassing the previous genetic and psychological conceptions of the work. He found support for his ideas precisely in phenomenology and in the aesthetics based on it. One of the most pivotal works of the twentieth century on the theoretical problems of literature originated somewhat later on the same basis. This was Roman Ingarden's Das literarische Kunstwerk: Eine Untersuchung aus dem Grenzgebiet der Ontotogie, Logik und Literaturwissenschaft (1931).17 The thirties were an important period for Polish inquiry into questions of literary theory and poetics especially. As can be seen in retrospect, the directions of development that essentially determined the manysided, modern expansion of studies beyond the borders of Poland were decided at that time. Above all, the publications in the series Archivum tiumaczen, the first volume of which was the translation of V. Zirmunskij's "Zadaci poetiki" (1934), remind us of the relationship between the ideas and the works that introduced new tendencies into Polish theory of literature and European studies, especially in the field of poetics. The second volume, a collection of works by K. Vossler, L. Spitzer and V. Vinogradov (Z zagadnien stylistyki, 1937) acquaints the readers with the contemporary trends in stylistics. Z. Lempicki wrote the introduction to this volume. Finally, a selection of Russian Formalist articles, Rosyjska szkola formalna, 1914-1934 was the third volume prepared for publication, but it could not be published after the war broke out in 1939. The translation of B. TomaSevskij's Teorija literatury was also prepared. Another characteristic series which collected works from the realm of literary theory was entitled Z zagadnien poetyki. In it appeared works directed also toward actual problems. Yet in it both Polish and foreign scholars acknowledged the local tradition: a Festschrift dedicated to K. 16

Anton Popovii's book Strukturalismus v slovenskej vede, 1931-1949 (Martin, 1970) provides a detailed history of Slovak developments. 17 Ingarden's extensive studies Vom Erkennen des literarischen Kunstwerkes (Tübingen, 1968) and Erlebnis. Kunstwerk und Wert (Tübingen, 1969) are the outcome of this line in his work.

886

F. SVEJKOVSKf

W6ycicki, a representative of this tradition as well as an initiator of the expansion Kazimierowi of the studies, published under the title Prace ofierowane Woycickiemu (1937). Among Woycicki's writing are works which stimulated theoretical investigation in various directions, as is demonstrated by his generally conceived study Historia literatury i poetyka (1914) or his works devoted to stylistics and, above all, to versification. Thus it appears that in addition to the local tradition, partly influenced by German theories, the stimuli of phenomenology and Russian Formalism were primarily decisive for the development of new theories. In the thirties, the period of the greatest expansion in research, interest in poetics was generated on two planes — on the one hand, in the field of general theory of literature, on the other, in the development of certain specific branches, especially versification and stylistics. Most of all the effort to apprehend and to elucidate the literary work in its specific features — literature as an autonomous realm of art realized by language — linked this first general plane of study to contemporary tendencies. An approach from the positions of philosophy and aesthetics rather than from those of linguistics was characteristic for Poland; the aim of this approach was to find an intrinsic literary basis for theory. M. Kridl's theoretical work was written toward this end. His ideas are summarized in the book Wstgp do badan nad dzielem literackim (published in 1936 as the first volume of the aforementioned Z zagadnien poetyky).18 Kridl was the first to formulate the tendencies of the phenomenologically conceived theory of literature by emphasizing the literary work as a particular reality and by defining the object of study as the 'description' of the literary work in the phenomenological sense. He viewed the work as a unified structure. On this basis he then constructed his 'integral method', the aim of which was the study of the individual components of this structure. Since the goal of the theory was the 'description' of the literary work in its objective principles, poetics above all came to the fore. Like the Russian Formalists Kridl identified poetics with theory of literature and defined it as a 'general discipline studying literature as art'. It was no accident that he made reference to Zirmunskij precisely in this conception. In Kridl's work as well as in works of other scholars in this area of study, poetics acquired a definition but not a direct elaboration. Thus, e.g., in her study "Przedmiot, metoda i zadania teorii literatury' (1938), S. Skwarczynska tried to make the field of literary theory much broader than Kridl had. Skwarczynska likewise drew support from phenomenological aesthetic theories and their particular formulation in Ingarden. She surpassed the theories of the Russian Formalists by means of the thesis: 'literary creation is not verbal technique, as it might follow, e.g., from what the Formalist Tomasevskij suggests' (1938 = 1960: 279f.), and she broadened her view of the work by incorporating it into its relation to the creator and into its other, suprapersonal ties. Conse18

M. Kridl has provided a brief summary of the basic theses of his work in the article "The integral method of literary scholarship: Theses for discussion", Comparative Literature, 3 (1951). pp. 18-31.

THEORETICAL POETICS IN THE TWENTIETH CENTURY

887

quently, poetics became a part of a more complex system of literary theory, and precisely under the influence of the connections between a work and a 'life', as the author says, was a consciousness of the two-sidedness of the individual categories of poetics attained: a consciousness of the permanent and the changeable in every element, in every means. The differentiation of the work in its generality and in its individual concreteness, which phenomenological aesthetics had accepted, corroborated this very conception. K. Troczynski's works were another important contribution to the field of general literary theory in Poland in the thirties. His starting point was similarly the new spirit in these studies and an approach to the definition of poetics from a general philosophical-aesthetic basis. He defined the literary work as 'a world of intentionally created fictions in the written word', as seen in his study Przedmiot i podzial nauki o literaturze (1929). For his overall conception of literature the analysis of the work in various graduated areas is characteristic: 'An element, a system, a change of system, a work, a complex of works, interrelations between works constitute a series of facts which are more and more complex, which complicate one another more and more and which comprise the object of study.' He then divides these areas into two spheres of study: the first three are the subject of 'theory of literature or poetics', the remaining three the subject of the history of literature. The identification of poetics with the theory of literature and the scope of its object link Troczytiski to contemporary formalist aspects of literary-theoretical studies, as follows from his statement: 'Poetics is the science of the configuration of the work exclusively, eliminating the problem of content as extra-aesthetic and not susceptible of being grasped theoretically as a whole; history of literature, on the contrary, does take into account problems of content, studying the relationship of configuration to the contents expressed by it' (1929 = 1960:33). In Troczynski's last theoretical book, Elementy form literackich (1936), special attention is given to the structure of the work itself. By distinguishing the individual levels of the work — the graphic, the linguistic (the symbolizing level), the level of artistic reality — he introduced his own conception, distinct from Ingarden's theory. He thus assigned an important place to poetics in the study of the literary work, as he had already suggested in his earlier study of 1929 when he wrote: 'Poetics is a generalizing, analytic, comparative, causal, nomothetic and therefore theoretical conception of literary elements, their typical systems as well as the changes occurring in them' (1929 = 1960:33). In his formalizing tendency and in his effort to generalize he approaches the most modern trends in theoretical poetics. A second important aspect for the study of problems of poetics in Poland derives from the opposite side of the theoretical work just discussed: from the study and the analysis of material. Versification, represented above all by F. Siedlecki, and the study of literary stylistics (J. D. Hopensztand, K. Budzyk) were main areas of interest for Polish research at that time. Furthermore, in comparison with the general theoretical branch of research on literature it is characteristic that it drew its support

888

F. SVEJKOVSKtf

more from linguistics. On this plane it found points of contact with Prague Structuralism. 2.42 To understand the significance and place of Czechoslovak Structuralism from the twenties to the forties one has to take into account not only the situation of the entire development of study in Europe but also the local preconditions for a coalescence with the whole spectrum of contemporary tendencies. It is necessary above all to mention the Czech traditions of Herbartian Formalism, which have their roots in the nineteenth century, — the particular elaboration of the tradition of psychological aesthetics in the works of O. Zich, J. Mukafovskf's teacher, and finally the close connection between theory of literature, on the one hand, and theory of art and aesthetics, on the other. J. Mukarovsk^ himself, the main representative of the emerging structuralism, emphasized the role of this tradition. About O. Zich he wrote that 'in his last works he was approaching structuralism' (current interest in the history of Czechoslovak Structuralism has elaborated this study of connections even further, mainly O. Sus 1964, 1966a, b, c and K. Chvatik 1966).1® This tradition represented forces that laid the foundation for future development and participated in the further course of structuralism since they raised problems which extended beyond the limits of the initial formalist studies influenced by Russian Formalism. In these forces we have one of the sources which in the work of Mukafovsk^ and his followers led from the initial stages to the full development of structuralism with the cooperation of linguistics and, above all, of R. Jakobson, whose work connected the two spheres, the linguistic and the literary, and at the same time represented a path to new levels of Russian Formalism. The road to the further developmental stages of structuralism entailed an increasingly wider encompassment of problems, in some cases of theory of art in general, since structuralism was dealing with a wide range of questions of literary theory at this time. Herein lay the answer to questions raised by some members of the beginning stages of structuralism in the Prague Linguistic Circle (esp. R. Wellek and A. Bern) and in a wider scope mainly by Marxist critics (K. Konrad, Z. Kalandra who likewise appeared at the meetings of the Prague Linguistic Circle).20 When Mukarovsk^ was working cn his article 'Structuralism' to appear in the Czech encyclopedia, and when he was devoting attention to the entire range of structuralism in aesthetics and literary studies, he emphasized three basic roots: aesthetics, philosophy, linguistics (1963a). If we confront this with what we have already learned from European developments, it is evident that in the merging of tradition with modern trends the further developmental stage of Western European 19

Cf. also R. Wellek, "The literary theory and aesthetics of the Prague School" in Discriminations, 2nd ed. 1971, pp. 275-303 (with a bibliography of 'Papers by Mukarovsk^ available in French, German and English'). 20 See F. VodiSka, "Celistvost literarniho procesu" in the book Struktura a smysl literdrniho dila (Praha, 1966), pp. 87-107.

THEORETICAL POETICS IN THE TWENTIETH CENTURY

889

studies coalesces with Slavic studies on the level of this structuralism. It is also necessary to relate this to poetics. Just as important is the relationship of Prague Structuralism to the role of theoretical principles found in the poetics of individual poets or directly in their theoretical statements. J. Mukafovsky embodies the aforementioned trends by combining a consideration for aesthetics and theory of art with a linguistically oriented study of literature. As early as the twenties his interest in new theoretical trends was accompanied by an effort to overcome traditional positivist conceptions or the one-sidedness of psychological interpretations. It is noteworthy that Husserl was not unlike him, that he dealt with neo-Kantism, that Dessoir interested him (Mukafovsky 1936a). The role which his encounter with Russian Formalism played is evident from his first works published in the second half of the twenties. Mukarovsk^'s interest was concentrated mainly on questions of poetic language and the sound level of the work. Poetics became his major field of interest, and — as further developments showed — it occupied a decisive position in the formation of his initial theoretical views. Therefore a study of the work led Mukarovsk^ to poetics, and only from there did his interest extend to the work in the total complexity of its conditions and relations — no longer just to the literary work but very often to the work of art in general. Mukafovsky himself has designated the study of questions of poetics as the first phase of his work. From here he saw a path to the study of the role of the individual in art, from a conception of the work as an autonomous whole to an examination of its connections with the extra-artistic sphere.21 At the same time the general questions of aesthetic function, norm and value are raised, questions to which he devoted special attention and which he included in his object of study — the historically concrete work of art. Thus his theory emerged at the end of the thirties and the beginning of the forties. If poetics stood at the beginning of this development, then his interest in it did not wane even in these wider views of the problems of art. His conception of poetics became more profound precisely as a result of them. The development and contribution of Czechoslovak Structuralism begins where the perspectives of a structural conception of the work expanded after the initial stages. Consideration of questions of structure was closely connected to a response to Hegelian dialectics which had already played an important role in the Russian school. In reference to this stage Mukafovsky's studies, developed in close association with the work of representatives of the Russian tradition — primarily R. Jakobson and P. Bogatyrev who were already permanent residents in Czechoslovakia — we can quote Jakobson's remark on Russian linguistics: 'The attitude toward the great epoch of German philosophy and especially to Hegelian dialectics was positive. I do not believe that it would be possible, for example, to explain certain aspects of the linguistic doctrine of Trubetskoy, one of the most prominent representatives of the linguistics of our century, without realizing that Hegel 21

F. Vodi2ka presents an analysis of this evolution and its sources in his introduction to the first edition of Mukafovsky's Studie z estetiky (Praha, 1966), p. 11.

890

F. SVEJKOVSKf

was the philosopher who had influenced him' (Jakobson 1970:31). Mukafovsky widened the view of the structure of the literary work by his notion of the 'semantic gesture' as a unifying principle of semantic character that is the foundation of the composition of the work, of the organization and the hierarchization of the individual components of its structure. This principle is part of a broader view of the literary work as a specific kind of sign. Linguistic starting points, on the one hand, and an orientation toward the problems of art as a whole, on the other, made it possible for Mukafovsky to develop a theory of the work of art as a sign. In 1934 he wrote in an article that accompanied the Czech translation of Sklovskij's Teorija prozy: 'Structuralism . . . insists on the requirement that scientific study not view its material as a static and disintegrated chaos of phenomena but apprehend each phenomenon as both a resultant and a source of dynamic impulses and the whole as a complex concert of forces', and he emphasized the need to value the 'semantic validity of those components which are usually called formal' (1934 = 1948, 344f.). The following formulations indicate further steps: 'Everything in the work of art and in its relation to the surroundings, then, appears to structural aesthetics as a sign and as meaning; in this sense structural aesthetics can be considered as part of the general science of the sign or semiology' (1936a = 1948:22). 'Without a semiological orientation the theoretician of art will always be inclined to regard the work of art as a purely formal construction or as the direct reflection either of the psychic, even physiological dispositions of the author, or of the distinct reality expressed by the work or of the ideological, economic, social or cultural situation of the given milieu'(1936b= 1966:87).Thus the evolution of Mukafovsky's ideas anticipated one of the main courses of literary theory in the future. It must be realized that he drew support from the Saussurian tradition and from Biihler; AngloAmerican semantic theories did not have a direct influence on the formation of these ideas. In the essay "Umeni jako semiologies fakt", from which we have just quoted, we find still another passage that has influenced the conception of poetics in a particularly significant way: 'Every work of art is an autonomous sign composed: 1. of a "work-thing" functioning as a sensory symbol; 2. of an "aesthetic object" lodged in the collective consciousness and functioning as "meaning"; 3. of a relation to the thing signified, a relation which does not aim toward distinct existence — because we are concerned with an autonomous sign — but the total context of social phenomena (science, philosophy, religion, politics, economy, etc.) of the given milieu' (1936 = 1966:88). Mukafovsky did not project questions of poetics onto the plane of general categories and unchangeable essences but elaborated them on the basis of 'the work-aesthetic object', hence the work in its concrete realization and concrete relations. He arrived at a model that attempts to encompass all the components of communicative relations: the author — the work — the receiver. At the same time this model paves the way for solving the problems of the aesthetic value of the work and the problems of evolution. Namely, it is

THEORETICAL POETICS IN THE TWENTIETH

CENTURY

891

necessary to see — and the quotation has drawn attention to this — that one component of the model that results from Mukafovsky's conclusions is intersubjective 'consciousness', a tradition which preserves a particular collective experience, certain norms and ideals. At the same time this experience is the background for the origin and the reception of the work. Lodged in it is the 'permanent' that older poetics tried to include in their descriptions or prescriptions, but we also find here the preconditions of development, of change, hence of the 'changeable' that determines an author's poetics and the variety of the interpretations of a work in each new concretization by a reader or a listener. This model also deals with the relationship of a literary work to a broader historical context (the 'context of social phenomena' in the quotation above). In such a way Mukafovsky dealt with the preceding stages of his ideas which evoked the criticism that he had lost sight of the creative individual and the historical situation.22 From our viewpoint possibilities of a historical poetics closely related to theoretical poetics are thus manifested here. Mukafovsky wrote: 'The work of art in contrast to other kinds of signs, e.g. linguistic, does not primarily place emphasis on results, an unambiguous relationship to reality, but on the process by which this relationship originates' (1966:111). The specific function of the literary work (of art) in relation to other functions of human activity is expressed here. The essence of a conception of poetics, of its main field which is already far removed from the initial ideas of the twenties, is also specified. Moreover, poetics as it is conceived here can no longer exhaust the entire problematics of the literary work, but it is only a part of the wider range of the aspects of literary theory. At the same time this 'emphasis . . . on process' includes another characteristic aspect of this structuralism — the constant bond between theory and the concrete work of art, in other words the formation of a theory of the concrete individual work. Those scholars, who brought with them the traditions of Russian studies to Czechoslovakia, particularly R. Jakobson, also participated prominently, of course, in the further development of this theory in the Prague Linguistic Circle. Jakobson remained a representative of a linguistically oriented poetics, and thus along with Mukafovsky's aesthetically directed works he maintained particular elements and aspects of the studies of that time. Together with P. Bogatyrev he made an impact on the field of folklore studies (1929, 1931). He also intervened influentially in a specific area of study, in versification, especially with the works O ceSskom stixe preimuSZestvenno v sopostavlenii s russkym (1923) and "Starocesky vers" (1924). A many-sided study of verse was also a subject of particular interest for J. Mukafovsky, who in the same volume of Ceskoslovenska vlastiveda in which Jakobson published "Starocesky vers" elaborated a history of modern Czech verse. Both Mukafovsky and Jakobson, later, aroused interest in the study of poetics — and particular interest in the problems of verse — among their pupils (J. Hrabak, 22

Cf. p. 888 above.

892

F. SVEJKOVSKY

M. Bakos). At the beginning of the forties the anthology Cteni o jazyce a poesii (1942), in which we encounter other scholars such as F. Vodicka and J. Veltrusky, originated from the circle of students around Mukarovsky and from the activity of the Prague Linguistic Circle. M. Bakog, Mukarovsky's Slovak pupil, soon became the leading figure in the further development of literary structuralism in Slovakia, and besides his study Vyvin slovenského versa (1939) he also produced an extensive selection from the Russian Formalists' writings, Teoria literaîury (1941). 2.5 We have already mentioned in our introduction that the approach to the work that crystallized during the twentieth century reconciled studies of poetics, in some cases of other questions of literary theory as well, with the theoretical expressions of artists themselves or, again, with the interpretation and analysis of the literary work. Therefore before we attempt to survey the situation in the last decades, we must return to those areas which usually did not have their center of gravity in theory but which nevertheless penetrated it in connection with the problems of creation, because the subject of poetics was immediately related to it. Moreover, these expressions often make palpable the global trend of the evolution of poetics, as it has begun to appear to us in the course of the preceding considerations. 2.51 The role of scholars concerned with the theory of creation, of those in France above all, is extraordinarily important. This classic land of literary theory did not produce any modern theory of poetics during the first half of the century.23 Instead the names of authors or critics who proceed from the problems of contemporary art (and frequently from polemics against the rigidity of theories and interpretations) have figured in surveys of twentieth century poetics. The names of S. Mallarmé, P. Valéry, A. Breton and the manifestos of Surrealism have accordingly recurred in this connection. It is no accident that even Mallarmé is included here, for this line truly has its source in the nineteenth century and leads, as H. Friedrich stresses in his work Die Struktur der modernen Lyrik (1961), up to modern times. In this sense not only the further evolution of 'Symbolism' but also the Surrealists — and the theoreticians of our time — acknowledged Mallarmé. Valéry's deliberations even resulted in a cycle of lectures on poetics and in the publication of these studies in Variété. Primarily a rationalistic, de-Romanticized attitude to creation and interest in the principles of literary production brought these poets close to the contemporary approach to questions of poetics. Hence the complete separation of literary creativity as a specific human activity. The Symbolists in Russia played a role similar to that of their counterparts in France. With their interest in linguistic means, mainly in the study of verse, rhythm and rhyme, they partly anticipated the theories of the Russian Formalists and partly comprised another important element in the 23

Michel Dragomirescu provides a comprehensive synthesis of the traditional approaches in the book La science de la littérature (Paris, 1929).

THEORETICAL POETICS IN THE TWENTIETH CENTURY

893

field of literary theory during the first decades of the twentieth century (esp. A. Belyj, V. Brjusov).24 2.52 It is also possible to find connections with French Symbolism or a generally analogous approach to creation in close relation to theory among poets in the Anglo-American cultural sphere. It is not accidental that in the study of modern literary criticism we often encounter references to the statements of T. S. Eliot or E. Pound and find the name of H. James next to theirs. The connections which appear with regard to global contemporary tendencies are demonstrated above all by the fact that the work itself, the work liberated from an excessive and one-sided dependency on external causes and influences, comes to the fore here, too. The work is no longer understood as the immediate expression or product of these causes and influences but as the expression of the objectification of the poet's experience, as its transposition into a new situation by means of literary media. It is necessary to accept and to study the work keeping this process in mind. Thus one of the pivotal problems of twentieth-century literary theory is revived in a specific modification. This problem had an equal impact upon the thinking of T. E. Hulme, E. Pound and T. S. Eliot, as it had had somewhat earlier, e.g. in the origin of T. A. Meyer's book (which we discussed above). Meyer said: . . . 'so gelangte ich zur Uberzeugung, dass nicht innere Sinnenbilder, wie man lehrt, sondern die Worte und Gedanken der Sprache selber das Darstellungsmittel der Poesie sind' (1901:IV). In 1910 Pound stressed that in his conception 'poetry is a sort of inspired mathematics, which gives us equations, not for abstract figures, triangles, spheres, and the like, but equations for the human emotions' (1910:5). Later he wrote: 'Great literature is simply language charged with meaning to the utmost possible degree' (1929 = 1960:28). He recommends that in the approach to the work one set out in the spirit of Dante; 'A canzone is a composition of words set to music' (1960:31). This consideration for formal means is not, however, dictated by mere formalism; on the contrary, emphasis on meaning, on what the work communicates, is essential, as his deliberation on the relationship of poetry and prose shows: 'Poetry must be as well written as prose. . . . Rhythm MUST have meaning. . . . Objectivity and again objectivity and expression: no hindside-beforeness, no straddled adjectives (as "addled mosses dank"), no Tennysonianness of speech; nothing—nothing that you couldn't, in some circumstance, in the stress of some emotion, actually say' (1950:48f.). The problem of objectivity in literary expression appeared even more conspicuously in Eliot's well-known formulation: 'The only way of expressing emotion in the form of art is by finding an "objective correlative"; in other words, a set of objects, a situation, a chain of events which shall be the formula of that particular emotion; such that when the external facts, which must terminate in sensory experience, are given, the emotion is immediately evoked' (1921:92). And let us add one more statement: ' . . . without a 24

Cf. Victor Erlich, "Russian poets in search of a poetics", Comparative Literature 4 (1952), 54-74; D . Tschiiewskij, "Wiedergeburt des Formalismus? In welcher Art?", in Immanente Ästhetik, Ästhetische Reflexion, Poetik und Hermeneutik (München, 1966), 297-305.

894

F. SVEJKOVSKY

labour which is largely a labour of the intelligence, we are unable to attain that stage of vision amor intellectualis Dei' (1921:13), and we surmise how the theory of creation is approaching the sphere of poetics; we surmise how many impulses are revealed here (in the approach W. K. Wimsatt and C. Brooks labeled the concept of 'an impersonal Art', 1957, chapter 29) for the study of its problems. However, a prevailing symptom of the development of the Anglo-American line of studies is the fact that consideration for these problems was predominantly linked to the concrete analysis of works. Interest in creation proper, in close reading, is primary, and only through their mediation are theoretical questions approached. Still another feature merits comment. This is the interest in the problem of 'objectification', not only in contemporary but also in older literature. T. S. Eliot directly ascertained and emphasized the continuity of creation against the background of his study of objectification (e.g. in the essays "Tradition and the individual talent", "The function of criticism" or "The metaphysical poets", 1932:3f., 12f., 241f.), addressed himself to both literatures as did Pound. In comparing different epochs, Eliot found 'a [artists'] method curiously similar' (1932:248-49). This regard for the literature of different periods and the search for points of contact between these poets and critics again coalesces with analogous tendencies in twentieth-century studies. 2.53 Besides Pound and Eliot, I. A. Richards shaped the traditon of English and American criticism. A primary regard for critical practice is stressed in the very titles of his major works: Principles of literary criticism (1925), Practical criticism (1929). He finds a methodological basis in linguistic semantics and in psychology. These starting points are also decisive for the selection of problems and the approach to them. The essence, the psychological essence of both the act of creation, and that of the reception of the work interest him much more than the structure of the work itself. He also analyzes its individual components from this perspective. His Philosophy of rhetoric (1936) marks the further development of his semantic approach to the work. Thus stimuli for the study of the problems of poetics are found above all in the solution of the problems of the designating act. In this direction, too, he very strongly influenced the further development of criticism as can be seen in the works of his Cambridge pupils and coworkers — particularly in W. Empson's Seven types of ambiguity (1930) or The structure of complex words (1952). Likewise F. R. Leavis' approaches to the analysis of the work and to its linguistic component direct interest toward the text; however, they do not lead to a special theoretical study but remain faithful to the study of the concrete work and its uniqueness. If it soon appeared that the course of the next strongest trend, New Criticism, had been foreshadowed in this line of Anglo-American studies, then it is possible to apply the same statement to the subsequent relationships of criticism to poetics. 2.6

Still another strong trend crystallized in Europe. It attempted to carry out

THEORETICAL POETICS IN THE TWENTIETH

CENTURY

895

the suggestions of the new attitudes toward the literary work — to examine it in its uniqueness, to concentrate primarily on it. This was the German Interpretation der Texte. As a matter of fact, a 'mirror image' of the relationship between the theory and the study of the work appears in both Anglo-American criticism and this interpretational tendency. Whereas the dominanta in theoretical studies had naturally been in the solution of general questions, here, on the contrary, the work was the center of attention, and a theory was worked out with respect to it, to 'close reading'. At the same time the Anglo-American starting points in the study of texts are closer to the linguistically oriented theoretical line, whereas ties with a philosophical or traditional Geisteswissenschaften basis are predominantly determinative in the German cultural sphere. We are leaving aside the French explication de texte which had no substantial impact upon the development of twentieth-century theories, since in its methodological foundations it remained tied to traditional descriptive analysis that was developed further in the spirit of "Les techniques de la critique et de l'histoire littéraires" (1923) summarized by G. Rudier. The evolution of the German Interpretation der Texte stimulated wide interest in questions of poetics by its approach to the work. This was apparent above all in the works of a group of Swiss scholars headed by E. Staiger. W. Kayser also is associated with them. That which is understood today as a well-defined sphere of German studies in the area of interpretation originated from conditions and connections quite different from those of Anglo-American criticism. Most important, the latter did not have an antipode in the intensive elaboration of a theory but created one itself — albeit to a limited degree. The role of poets themselves in the evolution of this criticism is also characteristic. In contrast to this, German interpretation proceeded from a theoretical and historical interest and carried out, rather perfected, a theory. Furthermore, German Interpretation, characterized by an emphatic concentration on the individual work and by an effort to interpret it exclusively on its own terms, developed up to the end of the thirties to a great extent as a counterbalance to the analyses and literary-historical explications stemming from a Geisteswissenschaft conception and turning the work once again into the product of an ideological atmosphere, or even of the ideology of a certain period. Accordingly the shift of attention toward the work also had a partly polemical impetus, the roots of which extended beyond the field of literature in the period of Hitler's Germany. It is not accidental, then, that even after 1945, in the era of the search for the paths of a new culture, interpretation played an important role in the study of literary works (K. May, H. O. Burger). At the outset Max Kommereil, E. Staiger and Th. Spoerri above all exerted an influence on the development of interpretational work. At the same time Kommerell and Staiger demonstrate in a striking fashion what role Heideggerian existentialism plays in German Interpretation. In his general and theoretical considerations Kommerell emphasized the abandonment of older methods in favor of close reading (cf., e.g., his Gedanken über Gedichte (1943) or Geist und Buchstabe der Dichtung

896

F. SVEJKOVSKf

(1956)). The other two scholars (Kommerell died in 1944) have had an important impact upon this work since the forties. The rejection of positivism and psychologism as a basis for a different approach to the literary work has also been emphasized more than once in these scholars' statements. This is already clear in Staiger's first works which are considered characteristic for the beginnings of the theory of interpretation, especially the introductory part of the book Die Zeit als Einbildungskraft des Dichters (1939) and the collection of essays Meisterwerke deutscher Sprache aus dem 19. Jahrhundert (1942). From these works we can also deduce the line determining Staiger's path to interpretation: Dilthey, Wolfflin, Walzel, Husserl, Heidegger. The studies from the years 1945-1955, published under the collective title Die Kunst der Interpretation, are the culmination of his works. Th. Spoerri proceeds from the traditions of German stylistics and, together with E. Staiger, finds a platform in the journal Trivium. Both scholars consider the style of the work as the basic unifying principle of interpretation. Spoerri says: 'Literary science has its real existence in the criticism of style; it is philology, a love for language in the highest sense, and all history, philosophy and psychology must be no more than ancillary sciences for it. The purpose and aim of any philological effort is the text i t s e l f . . ( 1 9 4 2 : 2 ) . Staiger's exposition of style again points to positions from which style is interpreted as permanent and independent of time, but neither as eternal nor as infinite. Whatever appears as characteristic for style changes from time to time (1945:192). A second leading motive, which derives from their conception of the work and interpretation, is closely connected with the preceding: this is the most general plane of the 'poet's world' to which it is necessary to relate the literary work. Here these anthropological aspects come much more to the fore in the two authors' interpretational procedures than does a consideration of the means and ways of concretizing the work. It is no accident that a result of this is the suppression of historicism in explications, as the above quotation from Spoerri's work has already indicated. At the same time it is necessary to seek here the starting points for Staiger's contribution to theoretical poetics, his Grundbegriffe der Poetik (1945), of which we shall speak later. Thus the affinity of German interpretation and poetics is evident; at the same time, however, the two spheres are distinguished on the basis of their different aims: interpretation is concerned with problems of poetics insofar as they appear in connection with the interpretation of concrete devices or means, as a part of the problematics of the style of a work and its evaluation. Yet the conception of style and its evaluation are determined primarily by philosophical aspects, as we have seen. In contrast to this, poetics extends beyond this framework of the individual work and its structure and tends in Staiger's conception toward generalization on the level of philosophical anthropology. 2.7 If we have concentrated our attention on certain works, authors or cultural spheres, we have done so primarily in order to sharpen the contours of the basic new trends in twentieth century poetics. Hence it has not been a question of exhaust-

THEORETICAL POETICS IN THE TWENTIETH CENTURY

897

ing the entire situation of poetics from the viewpoint of all countries or all works, insofar as older tendencies were dominant in them, e.g. in the countries of Northern Europe, or the focal point of study was elsewhere than in the theoretical sphere, e.g. in Italy, Spain, South America. It is not possible, however, to pass over in silence the works and ideas of someone who at the very beginning of the twentieth century rejected the possibility of theoretical study that we are examining here — Benedetto Croce. He proceeded from the same foundations and needs for a new formulation of the problems of art at the turn of the century as those who in their conclusions reached new conceptions of poetics. Croce, however, arrived at precisely an opposite view of the work and its theory. He joined other theorists of his time in emphasizing the specificity of the work of art, but, on the other hand, he directly and polemically countered tendencies to formulate a theoretical poetics. He rejected not only the older models of this poetics but also any kind of attempt to observe in art a certain tradition of technical principles or formal devices. Wherever others during the twentieth century found a basis for forming, in some cases for fixing constant artistic gestures and artistic experience, hence on the planes of the fundamental sources of man's spiritual activity, Croce saw a sphere of specific artistic intuition that always seeks and creates its own expression again and again. According to Croce it is impossible to reduce this manifestation of free spiritual activity, where an intuitive act passes over into its expression, to a manifestation based on certain fixed formulas. Indeed, the title of his fundamental work, Estetica come scienza dell'espressione e linguistica generate (1902), already expresses this theory. While he thus rejects the significance of the scientific approach that proceeds by generalization, he also promotes the role of the analysis of the individual work as the way to grasp the uniqueness of artistic expression. For this reason, similar tendencies, especially the literary-stylistic tradition beginning with K. Vossler, found support from him. At that time Croce sharply distinguishes the sphere of art from the extra-aesthetic one; this finds expression in the precise definition of the domain of poetry vis-à-vis literature in all its breadth that he worked out also in his other main work devoted to literature, La Poesia (1936).

3.

F R O M T H E FORTIES T O T H E P R E S E N T

3.1 Without specifying developmental stages for a period which has still not ended and is still too close to us, we are in accord with other observations that the forties mark the culmination of a certain phase and the beginning of the stage which leads up to our immediate present. In the meantime poetics had been divested of the attribute of supratemporality, and its categories had ceased to be ready-made 'forms', of which an author or critic disposed. Emphasis on the individuality of literary creation and on the elaboration of the problems that the concretization of the

898

F. SVEJKOVSKf

work introduces was decisive. Differences in theories depended on the approaches to particular aspects of these new concepts of creation and its functioning. A justification of the effort to grasp not only the continuity, the 'permanence' of the categories of poetics — as theoreticians had been striving to do for a long time — but also to respect the individual, the unrepeatable as a second pole of these categories, will occur in further development. Thus problems of the bipolarity of these categories were raised, and they were accompanied by an effort to reach solutions either by emphasizing one of these poles or by attempting to bridge them. In the first case, the possibility of pushing poetics directly into the center of attention was apparent; in the second, on the contrary, more complex problems, which already exceeded the limits of poetics proper, came to the fore. At the same time problems that had been linked to poetics from the beginning appeared as well: the problems of evaluation and of historical connections, in some cases, of historical poetics. In twentieth-century theories we observe a split with respect to these questions. The traditional supratemporality of the categories of poetics is echoed in the effort to grasp the work in its objective phenomena, which leads to a typology detached not only from historicism but often from questions of the aesthetic value of the work as well. On the other hand, attempts to encompass these problems on a basis which respects the historicity and artistic specificity of the work emerge in an awareness of this situation. Consequently, the field of the poetics of the concrete work was again opened to investigation. 3.11 If we have designated the forties as a definite turning point in the foregoing developments, we have done so primarily because several works signifying the conclusion of previous investigation from various perspectives were published then. E. Staiger's book Grundbegriffe der Poetik (1945) had its source directly in poetics. J. Mukarovsky's Kapitoly z ceske poetiky (1941) summarized the results of literary structuralism in Prague. R. Wellek and A. Warren worked out the most comprehensive synthesis of the contemporary study of theory of literature in their book Theory of literature (first edition, 1942). W. Kayser's work Das sprachliche Kunstwerk (1948) had its main source in interpretational study. At the same time the study that has had the most profound impact on contemporary critical work — New Criticism — was forming in the U.S.A. It signified the culmination of the foregoing evolution of Anglo-American concentration on the literary work (close reading). The most recent phase in the evolution of literary theory and poetics begins with this culmination. In order to understand it in its entirety it is important to take into account not only the work done in this field but also the broader historical situation, which in this instance had an exceptionally immediate impact on many areas of scholarship. This was the Second World War, which created a special situation here: a confrontation of various developmental tendencies, which had originally been restricted to particular countries or cultural spheres, took place. From our viewpoint the results that came from the confrontation of representatives of the new tendencies in

THEORETICAL POETICS IN THE TWENTIETH CENTURY

899

Slavic countries with those of other world trends become especially relevant. The important role of mediator fell primarily to the United States, where R. Jakobson and R. Wellek played the most crucial parts. Examples of this are the advancement of American scholarship and C. Lévi-Strauss's role in post-war France. The activity of L. Spitzer and many other scholars also comes to mind. The role of the U.S.S.R. in postwar political and cultural life is further proof of the penetration of theories and stimuli beyond national boundaries, especially in Europe. This was manifested in a new wave of Marxist oriented studies. The 'discovery' of Russian Formalism or contemporary structuralist concepts in linguistics and literary studies went hand in hand with this (mainly in France and Italy but also in Germany). Similarly, there occurred an increased response to existentialism, psychoanalysis and mythological theories in theory of literature, for the beginnings of these theories actually date from the thirties. However, it was possible to reach a fully developed study on a broad international basis only in the post-war atmosphere of scientific works and artistic expression. It is necessary to say the same thing about the import of older semiotics for the theory of literature. Mention of semiotics leads us to another new aspect in the studies devoted to theory of literature during the last decades, to the effort to incorporate this theory into the broader context of contemporary science. What Roman Jakobson said when he was thinking about the situation of linguistics at the end of the sixties is, in essence, equally characteristic for the specification of this trait: At the beginning of the study of modern linguistics 'it was useful to emphasize autonomy. Today . . . we can ask for the cooperation of interdisciplinary work' (1970:32). This slogan has been frequently applied in contemporary theories; it has also appeared in studies of poetics in response to the methods of statistical and algebraic mathematics, information theory, semiotics, to say nothing of the already traditional relationship to linguistics and its new theories. Finally, an appreciation of connections with tradition has played a noteworthy role in the evolution of theoretical poetics in its relation to theory of literature. We have in mind connections with a long-standing tradition, about which we spoke at the beginning, but also with the tendencies of the first half of the century. This is apparent in the resumption and the elucidation of relationships with older theories from the viewpoint of contemporary conceptions, e.g. of Frenchmen to Valéry and Thibaudet, of Russian theoreticians to the studies of the second and third decades. Another manifestation of this tendency is the intensive translating activity which has aroused interest in Russian Formalism and Czechoslovak Structuralism in nonSlavic countries, interest in French Structuralism in Slavic countries or in South America, the worldwide response to G. Lukàcs' works, etc. 3.12 The development of the theory of poetics that has taken place during these last decades under the influence of new conceptions of science and art, which have helped to reveal new aspects of artistic creation, has crystallized in several major trends. In essence, we detect in them a further developmental modification of the

900

F. SVEJKOVSKF

theories that came to the fore during the preceding decades. There are primarily two basic orientations: a. The first reflects an approach to the work in its general essence, to the work reduced by abstraction to the phenomena determining its specificity as a literary expression or linking it to the very foundations of man's existence and activity. During the last decades the axes of this study tended even further toward levels of possible generalizations in order to project upon them the pivotal categories of poetics. It could be said schematically that these new stages are characterized, on the one hand, by the line from the individual creator, toward whom previous attention was primarily directed, toward man-the-creator in general (here we find the impulses of existentialism, psychoanalysis, archetypal theory and logic; cf. 3.2) and, on the other hand, by the line from the individual text toward the text in general (cf. 3.3), studied against the background of modern logic or semiotics, in some cases of information theory. — In other words, these tendencies orient the theory mainly toward the study of CODES determining the existence of a work. b. The second of the pivotal orienting forces leads, on the contrary, to penetration into the essence of the concrete work (cf. 3.4), in the complexity of its internal and external relations, whereby a onesided concentration on the work in itself, on its autonomy, is counterbalanced. Here the theory emanates from the existence and totality of an INDIVIDUAL ARTISTIC EXPRESSION and it seeks at this level, within the work itself, the universal principles and regularities. If need be, only from here it passes to the studies of broader, interdisciplinary connections. 3.2 Several important works solving the problems of poetics with respect to man-the-creator have appeared since the forties. At the same time they have differentiated their conceptions on the basis of their starting points which range over a truly broad scale: from anthropology to logic. These are above all Staiger's Grundbegriffe der Poetik (1945) and his other works that elaborate a theory of poetics in what he calls 'Musterpoetik'. He reached his Grundbegriffe by the path that phenomenology had prepared for him: das Lyrische, Epische, Dramatische, existing permanently as objective realities. Their interpretation against the background of Heidegger's ideas evokes the author's old conviction that literary study can be 'the contribution of an entirely independent science to general anthropology' (1939:9).25 Here this conviction receives a theoretical formulation. The author has examined the problem of poetics in its most general and hence initial stage. At the same time the problems of genre and style (style in the sense in which we know it from Staiger's studies devoted to interpretation) are combined in his Grundbegriffe. Staiger's other theoretical essays that followed this book were partly answers to some criticisms, partly 25

In connection with the generalizing tendencies in Staiger's theory we are reminded of Heidegger's characterization of the approach to the essence of the literary work, as Ludwig Landgrebe formulated it in the book Philosophie der Gegenwart (Bonn, 1952), p. 135: 'Das Kunstwerk rückt neben das Werk der Denker. Es sind die "Dichter und Denker", in deren Sagen ursprünglich eine Welt aufgestellt wird.'

THEORETICAL POETICS IN THE TWENTIETH CENTURY

901

a further elaboration of a theoretical poetics. Staiger criticizes older conceptions of poetics for trying to squeeze individual phenomena into traditional and fixed formulas. To this he juxtaposes the demand for rendering just the field of certain devices — 'Spielraum' — and for defining fundamental notions in it. Thus this scholar, too, deals with questions which will more than once be prominent in contemporary studies, with questions of the tension between the permanent and the changeable, between the general and the individual. At the same time Staiger revealed a relationship with traditional Musterpoetik, with its well-defined concepts and norms by means of his category 'Spielraum': '. . . mit dem Begriff des Spielraums schiene es mir grundsätzlich erlaubt, den Grundbegriffen eine lange Musterpoetik anzuschliessen und die Fragen aufzuwerden: Was ist im Raum der Ode, der Elegie, des Romans, der Komödie möglich?' (1948:290). Therefore he proceeds from the model of Musterpoetik, but he gives its basic concepts and norms a sense that corresponds to his theory: 'Der fundamentale Begriff des Spielraums vermittelt zwischen dem Kanon einer Musterpoetik im alten Stil und dem historischen Wissen um die mannigfaltigen Weisen der Dichtung, da wir nie verleugnen dürfen. . .' (1961:32). Thus he believes that having removed himself as far as possible from traditional poetics, he has found it again on the basis of this conception. At the time when E. Staiger was formulating his theory, Günther Müller was arriving at the general interpretation contained in his "Morphologische Poetik" (1944a), which shifted the starting points of the principles of creation directly to the level of the basic natural forces which manifest themselves in man. 'Die Kraft, von der die sprachgetragene Wirklichkeit der Dichtung hervorgebracht wird, ist eine Kraft der Natur', he writes, and elsewhere he says that 'Dichtung als eine ursprüngliche Auswirkung menschlicher Naturkräfte den eigensten Gegenstand der Literaturwissenschaft ausmacht . . . ' (1944b:226). Müller subscribes to the tradition of Goethe's Morphologie (1944a) and in its development he sees the basis which makes it possible to encompass historical aspects and at the same time to bridge 'den Spalt zwischen Natur- und Geisteswissenschaft' (1944b:246). From this basis Müller attempts to deal with all the fundamental aspects of poetics, above all with the question of genres and linguistic means, with the work as a structure, with the relations between the poet, the work and the reader. He also includes in his observations questions of the aesthetic value of the work: 'Diese Poetik ist nicht Ästhetik, sondern Gestaltkunde. Sie unterscheidet die Gebilde nach dem Bildungsgesetz, nicht nach dem ästhetischen Reiz, und sie hat mit einer Definition dessen, was auf diesem oder jenem Standpunkt als "schön" anzuerkennen oder als "unschön" zu verwefen ist, nichts zu tun' (1944b:241). At the same time it is apparent from this definition that Müller's work is essentially a re-evaluation of older traditions of twentieth-century German poetics on the basis of 'morphology'. For this reason he also speculates about a specific 'morphologische Schönheit' — hence he finds a place for it in his system and establishes it as a counterpart of general 'Schönheit': 'Das Goethesche Wort "Wie wahr, wie seiend!" beim Anblick von Seemuscheln

902

F . SVEJKOVSKtf

und Krebsen ausgesprochen, deutet die Erscheinungsart des morphologisch Schönen an, und dies morphologisch Schöne ist ein machtvolles Urphänomen . . .' (1944b: 242). This, we believe, sufficiently elucidates both the methodological basis of Müller's theory and its starting point in the 'natural forces'. The theory of archetype and myth, which had long ago been applied to different areas of the study of literary works, provided another meaning and other possibilities for projecting the problems of poetics into connections with the basic expressions of human existence and culture. This theory complied with the search for the unity of the general and the individual as well as with considerations for the specificity of the literary work and its anthropological basis. Starting points are found in the very foundations of culture, in basic expressions of these foundations. There we also encounter the beginnings of the tradition of the principles and means of literary creation. The suggestiveness of these ideas for the study of problems of poetics has been most systematically elaborated in Northrop Frye's book Anatomy of criticism (1957 r= 1965). In his conception poetics coalesces with 'theory of criticism': 'A theory of criticism whose principles apply to the whole of literature and account for every valid type of critical procedure is what I think Aristotle meant by poetics . . . Meanwhile, the opening words of the Poetics, in the Bywater translation, remain as good an introduction to the subject as ever, and describe the kind of approach that I have tried to keep in mind for myself: "Our subject being poetry, I propose to speak not only of the art in general but also of its species and their respective capacities . . . . " ' (1965:14). The relationship between general theoretical aspects and the intrinsic problems of the concretization of the work through literary means and devices is characteristically emphasized by the bipartition of the titles of the chapters in Anatomy of criticism: Historical Criticism: Theory of Modes; Ethical Criticism: Theory of Symbols; Rhetorical Criticism: Theory of Genres; The Mythos of Spring: Comedy; The Mythos of Summer: Romance; etc. At the same time this reveals the connections of the fundamental categories of poetics with primitive forms and utterances that expressed man's relation to the world and to life. Genealogical questions, above all, become the focal point of interest, for it is possible to grasp in them the relations between the problems of poetics and the pivotal expressions of the human view of reality. From the viewpoint of the entire evolution of poetics and its place in contemporary studies of the literary work it is necessary to emphasize not only that poetics constitutes an integral part of the theory of the literary work, as the above quotations have suggested, but also that Frye demonstrates the union of two aspects in his theory: a tendency toward the general foundations of the theory of poetics and a tendency toward the concrete work. This union is realized by the practice of criticism. As it has been shown, those components of the work that are usually understood as semantic have come to the fore in the majority of these theories. We find their counterbalance, namely an emphasis on the problems of expressive means or directly on the signifiant, in the second basic line of contemporary studies which we will

THEORETICAL POETICS IN THE TWENTIETH CENTURY

903

examine later. First, however, we want to call attention to the expression of tendencies which take up the aforementioned orientation of theoretical studies and carry it right to the level of logic; in some cases, to the level of the theory of the symbol. We are referring to Käte Hamburger's book Die Logik der Dichtung (1957).26 The author proceeds from the problems of literary genres, which has been the starting point or the basis of the great majority of the works that we have considered so far; she designates genres as the 'zentrales Problem der Poetik' (1957:1). She approaches them in order to elucidate them in their most general essence, and she has thus come close to the furthest limits of generalization about the act of creation. She starts from the premise 'dass es eine Logik, oder ein logisches System der Dichtung gibt', and in this she sees the possibility of introducing literature into a relationship with a system of thinking and to reach the boundaries where poetry 'in die Prosa des wissenschaftlichen Denkens übergeht' (1957:243). These are the possibilities which literature as a linguistic expression offers. It is precisely a study of the relationship between logic and language that brings the author into contact with the problems of poetics, that bring to the foreground the study of the basic forms of verbal expression, hence of literary genres, and the study of questions concerning the subject in a work which determines the character of the literary utterance. From the viewpoint of her theory she divides the traditional genres into two spheres: die fiktionale oder mimetische Gattung (subdivided into 'die epische Fiktion', 'die dramatische Diktion', 'die filmische Dichtung'), and die lyrische oder existentielle Gattung. Moreover, K. Hamburger finds grounds for speaking separately about 'Sonderformen', where she places her considerations about the ballad and 'Icherzählung'. Right at the beginning of the book, however, she has emphasized that logical points of view entail a necessary reduction in the approach to the literary work: 'Die Logik der Dichtung hat es mit der dichtenden, aber nicht mit der dichterischen Sprache zu tun'. Thus it is necessary to deal with a second basic aspect of the work, which the author calls the aesthetic one. The ties of literature to the whole specific sphere of art are expressed in this aspect. In this connection her thesis about the relationship between Dichtungslogik and Dichtungsästhetik is especially noteworthy for the study of poetics. This relationship is the closer 'je tiefer sich die ästhetische Betrachtung in Fragen der Technik und Strukturen einlässt' (1957:244). Here the problems of literary theory already begin to exceed the object of the author's interest, and a field that K. Hamburger reserves for the study of symbolic forms (cf. the concluding chapter of the book, "Zum Symbolproblem der Dichtung") is opened up to investigation. 3.3 26

This projection of the problems of poetics both onto the level of logic — in

The author further elucidates her basic theoretical starting points in the following articles (as well as in other works devoted mainly to literary genres): 1) "The theory of statement: Prolegomena to a theory of literature", Algemeen Nederlands tijdschrift voor wijsbegeerte en Psychologie 56 (1964), 251-256; 2) "Zur Theorie der Aussage", Zeitschrift für philosophische Forschung 20 (1966), 23-56.

904

F. SVEJKOVSKÎ

some cases, onto that of the theory of the symbol — and onto the level of anthropology indicate the immense scale along which contemporary study is realized. It is symptomatic that even those contemporary theories which did not make the creating subject and the genesis of the work connected with him their starting point but proceeded from the text, the structure of the work itself, encompassed the same span as some aspects of French structuralism will show further on. The diversity of methodological starting points resulted in the effect that the theories representing the two basic lines never coalesced in these aspects even though they were located on the same plane. Let us confront the following statement by R. Barthes with the theory based on the individual and on questions of the genesis of the work: 'On nous demande d'attendre que l'écrivian soit mort pour povoir le traiter avec "l'objectivité!" ' 'La science de la littérature ne peut qu'apparenter l'oeuvre littéraire, bien qu'elle soit signée, au mythe, qui, lui, ne l'est pas' (1966:59). And let us add this summarizing characterization of the 'science of literature' in the conception of French Structuralism: 'Pour que soit possible une science de la littérature, il faut dépasser la bibliographie des auteurs, l'histoire et la sociologie de leur époque, les commentaires philologiques de textes, et d'agencement. Le modèle d'une telle science est linguistique' (J.-P. Fages 1967:105). This leads us to an examination of the second pole of contemporary studies. 3.31 The most directly delineated pointed theories proceeding from the text as a linguistic expression and from the traditions of linguistic poetics have come from the circle of French theoreticians grouped around Roland Barthes. They are based broadly on the ties between literary studies and other scientific disciplines (particularly linguistics, anthropology, psychology, theory of art, philosophy), the common basis of which is structuralism. These studies, however, cannot be regarded — even in the field of literary theory — as an expression of a unified 'school'. Ideas are differentiated, and not all French theories focus attention on the questions which are the subject of our examination (e.g. L. Goldmann's 'le structuralisme génétique?). Since the fifties France has again played a striking — and in many respects a directly polemical — role in the contemporary development of theoretical studies through the works of the theoreticians in R. Barthes'circle. Until that time the former bifurcation of literary theory and critical practice which had been characteristic of the preceding decades of the twentieth century had, in essence, continued to prevail. It was characteristic of the situation that even at the beginning of the sixties Henri Morier summarized the basic problems of poetics in the form of a dictionary and speculated about the critical status of contemporary French study in his introduction (cf. his Dictionnaire de poétique et de rhétorique, 1961).27 Thus the new advance" H. Morier sees the function of his Dictionnaire de poétique et de rhétorique (Paris, 1961) in relation to the needs of the contemporary French reader who lacks a modern work devoted to poetics. The aim of the book is: 'Revenir aux figures; étudier leur raison d'être, leur mode d'action, leur valeur poétique, telle était en partie notre tâche. Une rhétorique moderne ne devrait pas se borner aux indications qui permettent de fabriquer une fogure; elle devrait surtout

THEORETICAL POETICS IN THE TWENTIETH CENTURY

905

ment of theoretical investigation has resumed the line represented in the past by the theory of authors and critics who created a counterbalance to the traditional (or — as it was called—the universitarian) concept of the literary work and continues the modern 'querelle entre les anciennes et les modernes'. Such names as Valéry, Mallarmé, but also Thibaudet and others, recur in contemporary theoreticians' references to tradition. The tremendous potential of the stimuli for contemporary theoretical study had already accumulated in the development of French works with a literary-historical or actually critical orientation, since during the forties and the fifties the stimuli of the most varied trends — above all, of psychoanalysis, existentialism, Marxism — had developed in them. No less relevant for the formation of contemporary theory in France have been linguistic studies focusing interest on the literary text and its problems. In retrospect Barthes (1968:4) mentioned the significance of the works of R. Jakobson and the Russian school, L. Hjelmslev and É. Benveniste. Finally, the theories of structural anthropology represented by C. Lévi-Strauss and the evolution of semiological studies have also been decisive for the methodological approach to the literary work. As can be seen, that which we have emphasized above as a typical feature of the most recent phase of studies — the coalescence of local traditions with developments overreaching the boundaries of countries and traditional cultural spheres — has appeared in a striking way in this creatively elaborated synthesis of stimuli. A total conception of the problems of the literary work, as some French Structuralists — notably R. Barthes himself — have developed it, has been decisive for the approach to questions of poetics. The goal has been to create a literary science, the object of which is 'non des oeuvres déterminées, c'est-à-dire inscrites dans un procès de détermination dont une personne (l'auteur) serait l'origine, mais des oeuvres traversée par la grande écriture mythique où l'humanité essaye ses significations, c'est-à-dire ses désirs' (Barthes 1966:60-61). The work has been viewed as a structure of linguistic character specifically organized with respect to the essence of the literary work, which lies in 'literariness' (cf. the follow-up of Jakobson's conception from the beginnings of Russian Formalism). Thus it is necessary to proceed from linguistics but at the same time to differentiate between it and literary science. Barthes speaks of the need 'limiter la tyrannie (ou la prestige) du modèle linguistique' (1968:6). Linguistics provides a generative model and even provides the motivation for considering a 'faculté de littérature' as a counterpart to man's 'faculté de langage' (Humboldt, Chomsky) (1966:58). These fundamental conceptions have for the most part thrust these authors' work into the field of poetics, as has been the case with other theoreticians working on a linguistically oriented poetics. In his study "Poétique", which will be discussed later, T. Todorov was able to state: 'notre l'étudier du point de vue psychologique, voir ce qui se pase dans l'âme du lecteur au moment où la figure y pénètre, comment elle s'y décompose en développant une energie qui émeut la sensibilité, en un mot comment elle agit pour causer en nous cet émerveillement qui est un effet de l'art' (p. VII).

906

F. SVEJKOVSKY

emploi du terme poétique. . . coïncide d'assez près avec que Roland Barthes a appelé la "science de la littérature" ' (1968:103, note 2). However, if we want to specify the conception of poetics, we must take into account one more pivotal thesis of French Structuralism. In contradistinction to those tendencies which attempted to find a path from a general level to the concrete work, as we have observed, e.g., in Staiger's deliberations about Musterpoetik or in the biopolarity of Frye's Anatomy of criticism, R. Barthes introduced a sharp dividing line between 'la science de la littérature' and 'la critique' (and besides this he distinguished a third approach, 'la lecture')-. La science de littérature 'ne pourra être une science des contenus . . . mais une science des conditions du contenu, c'est-à-dire des formes . . .'; ' . . . son objet ne sera plus les sens pleins de l'oeuvre, mais au contraire le sens vide qui les supporte tous' (1966:57). In contrast to this 'la critique n'est pas la science. Celle-ci traite des sens, celle-là en produit. Elle occupe . . . une place intermédiaire entre la science et la lecture' (1966:63). Poetics, as well as theoretical generalization, in which the traditional tension between the individual and the general is reflected, derives from the following fact: 'L'objectivité requise par cette nouvelle science de la littérature portera, non plus sur l'oeuvre immédiate (qui relève de l'histoire littéraire ou de la philologie), mais sur son intelligibilité' (1966:62). Barthes' ideas became a stimulus for other scholars; as a result works devoted to poetics as a whole or to some of its individual problems came forth.28 In general, it may be said that the basic study of questions of poetics has been determined by the fact that the objects of examination are not works in their concreteness but in their 'conditions' (Barthes), 'realisations possibles' (Todorov), works as 'l'objet de l'activité structurelle', 'des types de discours et non des types d'oeuvres'. Barthes' conclusion concerning literary genres can characterize the whole approach to the conception of poetics and the modes of its elaboration: 'Pour préserver la liberté de l'analyse, il suffira — ce qui est déjà fait par la plupart des chercheurs — de situer la spécificité du genre, non plus dans des règles très générales de composition, dans des macro-structures (à la manière de la poétique aristotélicienne), mais dans des schémas syntaxiques élémentaires: répétition/attente pour le poétique, nom/ verbe pour le narratif . . ( 1 9 6 8 : 7 ) . Similarly, with regard to semiology it is necessary 'remplacer des genres par une typologie des textes', says J. Kristeva (1968:10). The last two quotations characterize a typical feature of the group of scholars in their conception of poetics: the shifting of traditional categories into general positions determined by a specific definition of the literary work, of 'literary science' and by their incorporation into semiological studies. A rapprochement with linguistics, logic, the theory of myth is accompanied (just as in linguistics) by a retrospective glance at traditions of grammar, rhetoric and poetics. Of course, these notions, along with the abovementioned categories of poetics, are interpreted by a shift onto those levels about which we have spoken, i.e. the sphere of language and of semiotical aspects. Let us compare, e.g., Barthes' "Rhétorique de 28

Attention has been devoted above all to the question of literary genres and stylistics.

THEORETICAL POETICS IN THE TWENTIETH CENTURY

907

l'image" or the note concerning poetics in his essay "L'activité structurelle": ' . . . what is new is a mode of thought (or a "poetics") which seeks less to assign completed meanings to the objects it discovers than to know how meaning is possible, at what cost and by what means' (1964 = 1967:87). Todorov's works "La grammaire du récit" and La grammaire de Décamerone follow a similar course in re-evaluating the concept of grammar. Tzvetan Todorov produced the first general work on questions of poetics, and he uses this word directly in the title: "Poétique" (1968). In this study one can see a synthesis of a certain stage of the scholars' works and other conclusions or stimuli from Barthes' circle. The object of a theory of poetics is the 'literariness' of the work, and the modes of its realization are sought at the level of 'l'analyse du discours littéraire', as Barthes has defined it. 'Toute oeuvre n'est alors considérée que comme la manifestation d'une structure abstraite beaucoup plus générale, dont elle n'est qu'une des réalisations possibles.' 'L'oeuvre se trouvera alors projetée sur autre chose qu'elle-même, comme dans le cas de la critique psychologique ou sociologique; cette autre chose ne sera plus cependant une structure hétérogène mais la structure du discours littéraire lui-même. Le texte particulier ne sera qu'un exemple qui permet de décrire les propriétés de la littérarité' (1968:102). These are both the starting points and the objections of the poetics that T. Todorov works out in relation to his subject which is 'discours littéraire'. After formulating the process of creation and the categories of poetics linked to it by means of elementary schemes, for which Barthes called, Todorov switches to "Perspectives" in his conclusion. In it he deals with other aspects of theoretical poetics which were not included in his model: the aspects of historical and aesthetic evaluation. Finally, he returns once again to a deliberation of the object of poetics and comes to a conclusion corresponding to the general conception of literary science and its object, as the French Structuralists present it: 'Que la poétique est son propre objet' (1968:163ff.). In Elements of semiology Barthes distinguished in addition to inner relations outer, external relations, relations to the code and to the context. J. Kristeva primarily directs attention to these external relations in the study of the literary work. Her theory of the 'idéologème', which is supposed to establish ties to the historical and social context, is characteristic of this: 'L'idéologème est cette fonction intertextuelle que l'on peut lire "matérialisée" aux différents niveaux de la structure de chaque texte, et qui s'étend tout au longe de son trajet en lui donnant ses coordonées historiques et sociales' (1968:104). Surely it is no accident that Kristeva often focuses attention on questions of historical poetics and more than once draws support from the stimuli of the last stage of Russian Formalism, represented by 'Baxtin's school'. 3.32 The long-standing traditions of linguistics in Russia and particularly the relations between linguistics and literary theory have to a large extent been revived in the U.S.S.R. Together with stimuli from outside, from the development of analogous study in other countries, they have given rise to new currents in the field of

908

F. SVEJKOVSKf

theoretical poetics. Thus we can ascribe several works from the U.S.S.R. to the movement that has just been characterized by the French theories. Here the most striking statement is Aleksandr K. Zolkovskij's and Jurij K. Sceglov's programmatic essays 'O vozmoznostjach

postroenii

strukturnoj

poetiki'

(1962) and

'Strukturnaja

poetika — porozdajuScaja poetika' (1967b). The latter title already suggests its linguistic orientation. The authors understand their conception of poetics as 'one of the possible directions in the study of it'. For these scholars, as well as for many contemporary theories, 'process', the act of forming the work, is the object of poetics: Very briefly speaking, it seems useful to understand the work of art as a 'concrete object' intended for a maximally effective execution of the theme, its own kind of apparatus for suggesting the theme to the reader; it can be compared to an invention realizing any concrete technical task. The aim of literary-scientific studies, then, must consist partly in describing the organization and the work of such artistic 'machines', in showing how they 'are assembled' by proceeding from the thematic task. It is natural to expect that the translation of a theme into the system incorporating its means is an objective process subordinated to definite — albeit the most general — laws. Many such laws have essentially been known for a long time (Zolkovskij and Sceglov 1967b:82).

The emphasis of the traditions of Russian Formalism and the evaluation of its significance for contemporary scholarship are characteristic. The two authors have devoted another article, "Iz predystorii sovetskych rabot po strukturnoj poetike" (1967a), to this topic. Both the role of Russian Formalism and the name of Roman Jakobson, the leading representative of linguistic poetics, who unites the two approaches and who has preserved and developed this tradition up to the present, have recurred again and again in linguistically based works on poetics. Jakobson's articles "Linguistics and poetics" (1960a) and "Poetry of grammar and grammar of poetry" (1960b) in particular are signposts of the present. The formulations which we find in these articles are a testimony not only to the compactness of the author's work in poetics but also to its development in accord with the development of linguistic scholarship, in which Jakobson has played an essential part. Emphasis on linguistic poetics preserves continuity. 'Poetics deals primarily with the question: What makes a verbal message a work of art? Because the main subject of poetics is the

differentia

specifica of verbal art in relation to other arts and in relation to other kinds of verbal behavior, poetics is entitled to the leading place in literary studies... Since linguistics is the global science of verbal structure, poetics may be regarded as an integral part of linguistics' (1960a:350). In connection with the recent development of linguistics he then states the problem of 'the poetic function of language' in a formulation that has received a wide response (from statements which have surpassed it to critical replies): 'The set (Einstellung) toward the MESSAGE as such, focus on the message for its own sake, is the POETIC function of language. . . The poetic function projects

the principle

of equivalence

from the axis of selection

into the axis of

THEORETICAL POETICS IN THE TWENTIETH

CENTURY

909

combination' (1960a:355). The essence of Jakobson's work on poetics lies in an effort to direct attention to the linguistic problems of the study of literary creation, to focus attention on the text itself. Connected with this are the selection of the problems which he stresses, the selection of criteria and finally a large number of his concrete analyses of literary works from this viewpoint. His theory concentrates on the work in itself, on the work-text. Jakobson sees a means of solving other problems which extend beyond the level of the text by incorporating the verbal work into broader contexts, for which 'the whole theory of signs, that is, general semiotics' paves the way. Among other general questions of poetics the author also deals with historical poetics and its relation to synchronic study: 'A thoroughly comprehensive historical poetics or history of language is a superstructure to be built on a series of successive synchronic descriptions.' If Jakobson's ideas have influenced and are still influencing linguistic poetics, then it is necessary to see this assertion not only in the trend toward a poetics operating on the abstract level, as is the case with the French theoreticians, but also in theories tending toward interpretation, toward a poetics of the concrete work. After all, Jakobson's studies, whether they have had a theoretical character or have been concerned with some work, have examined these two aspects in constant correlation. Characteristically, T. Todorov deals with this fact at the beginning of his "Poétique", and he understands this dual thrust of Jakobson's conception in the following manner: 'Pour permettre de situer plus facilement notre emploi du terme [i.e. la poétique], signalons . . . qu'il a des traits communs avec l'emploi qu'en fait Jakobson, notamment en ce qui concerne sa relation avec la science en général et la linguistique en particulier, mais s'en distingue en ce qu'il n'englobe pas la description des oeuvres concrètes' (1968:103, note 2). Another work devoted to poetics, Manfred Bierwisch's study "Poetik und Linquistik" (1965), has resulted from the stimuli of Jakobson's works and from the modern methods and theories of contemporary linguistics. On a linguistic basis the author explains the object of poetics which, he says, consists of 'die besonderen Regularitäten, die sich in literarischen Texten niederschlagen und deren spezifische Wirkung bestimmen, und damit letzten Endes die menschliche Fähigkeit, solche Strukturen zu produzieren und ihre Wirkung zu verstehen, also etwas, was man poetische Kompetenz nennen könnte' (1965:51). The author pursues the possibilities which structural linguistics offers the study of poetics. Here it is a matter of an empirical science, the aim of which is not 'literarische Qualitätsurteile wissenschaftlich zu rechtfertigen oder Normen für die literarische Produktion aufzustellen, sondern sie hat zu erkären, aufgrund welcher Struktureigenschaften festellbare Wirkungen Zustandekommen' (1965:59). M. Bierwisch agrees with some contemporary views, about which we have spoken above, in stating: 'Grammatische wie poetische Regeln sind ja zu einem grossen Teil Elemente einer nicht reflektierten Fähigkeit' (1965:60). After the author was shown possibilities that linguistics offers structural poetics, he emphasizes that its role in the entire complex of the problems of the

910

F. SVEJKOVSKf

literary work is a partial one. His noting of the need for a typology 'des poetischen Verständnisses' is a relevant point, for in so doing, he directs attention to other questions stemming from the communicative function of the text. The aim which the author has in mind is, however, a general one, 'eine generelle Theorie der Poetik', that will result from the empirical study of various types of 'poetic rules', the structures of poetic works and their mutual relations. The article suggestively demonstrates how it is possible to draw on modern methods of linguistics as well as statistical methods for constructing a poetics. Another of the German works devoted to poetics, Hans-Peter Bayerdörfer's "Poetik als sprachteoretisches Problem" (1967) purports to be a linguistically oriented counterpart to the views of German theoreticians who proceed from methodologically different bases (especially Staiger, Hamburger) and a development of F. Martin's idea about the linguistic basis of the work, which is the starting point for contemporary scholarship in theoretical poetics. Bayerdörfer's approach concentrates, of course, mainly on those problems that are vital in contemporary German scholarship. On the other hand, he centers the study of his 'sprachteoretisches Problem' around Saussure's pair langue-parole, and by means of it he approaches the analysis and theoretical interpretation of the work as a linguistic phenomenon. In his starting point he seeks a way to include both the general aspects of poetics and the problems of the individual work. He understands poetics as 'Sprachlehre der Dichtung' and places it alongside aesthetics to which, in his opinion, it is necessary to transfer many of the problems connected with the literary work as a work of art. It is symptomatic (and not only for modern times) that the stimulations and the contributions to the different problems of poetics were introduced on many occasions by the linguists; along others we mention here still the special contributions of A. J. Greimas, "La linguistique structurale et la poétique" (1967), S. R. Levin's Linguistic structures in poetry (1962), and in close relation to the problems of interpretation of literature, N. Ruwet's "L'analyse structurale de la poésie" (1963) and "Limite de l'analyse linguistique en poétique" (1968). 3.33 Efforts to find the essence of the process of creation, in some cases to reach a formulation of basic devices, have found a response in so-called mathematical poetics. This approach has been applied primarily where poetics has been studied on a linguistic basis, in a close relationship between theory of literature and linguistics, for it is precisely the latter that has already developed possibilities for applying mathematical principles, and more than once it has even directly included the literary work in its studies. Moreover, a definite tradition of the statistical method already exists in studies devoted to certain aspects of poetics (the Russian Symbolists and Formalists and Prague Structuralism had used this method since the twenties). Similarly experimental psychology, which has influenced the study of poetics, has used it. Finally, at the present time a trend in contemporary studies of aesthetics and theory of art, where the stimuli of algebraic methods as well as information theory are beginning to exert an influence, is in compliance with this interest.

THEORETICAL POETICS IN THE TWENTIETH CENTURY

911

The works that have appeared up to now have crystallized primarily in two spheres. The earlier study of verse has been resumed especially in Poland, Czechoslovakia and the U.S.S.R. Let us mention the recent contributions of A. N. Kolmogorov and A. M. Kondratov29 or the anthology of Polish studies, Poetyka i matematika (1965, with M. R. Mayenowa's introductory essay "Mozliwosci i niebezpeczenstwa metod matematycznych v poetyce"). These methods have been considerably developed in the study of stylistics (cf. R. W. Bailey and L. Dolezel, An annotated bibliography of statistical stylistics, 1968). Besides a number of journal articles and anthologies devoted especially to stylistics, the collections Mathematik und Dichtung (ed. H. Kohlhase and R. Gunzenhauser, 1965) and PoeticsPoetyka-Poetika (ed. M. R. Mayenowa, 1961 and 1966) present a survey of contemporary scholarship. For theoretical poetics the significance of contemporary study in this area lies in the fact that in connection with semiotics and information theory these methods make it possible to go beyond a mere statistical description and analysis of a single work. The perspective of mathematical modelling, which can satisfy the present-day striving for generalizing conceptions of theoretical poetics, the effort to reach the objective factors constituting the essence of literary creation has appeared. Several works indicate the directions of these new attempts which are the first manifestation of their kind. At the level of versification there is R. Jakobson's and J. Lotz's attempt to model a verse system on the example of the Mordvinian song in the article "Axioms of a versification system — exemplified by the Mordvinian folk song" (1952).30 In the works "K analizu struktury li ovskoj narodnoj ballady" (1962) and "K rekonstrukcii praslovjanskogo teksta" (1963), V. N. Toporov and V. V. Ivanov (co-author of the second article) have attempted to construct a model of a literary genre. From the perspectives of theory of information the totality of the literary text has been the concern of M. Bense in his book Theorie der Texte (1962). The scholar works out a conception which he calls 'Topologie der Texte' and which he understands as one of the possible approaches — besides the semantic, aesthetic and statistical — to the literary text. He finds support in the development of typology (cf. especially the reference to W. Franz's Topologie, 1960). We shall find another testimony to the search for an algebraic or — as the scholar himself prefers to say — a structural model of the literary work in the Polish scholar Eugeniusz Czaplejewicz's study "Poetyka matematyczna a badanie poezji" (1965). He takes into account especially the need for respecting the difference between linguistics and literary theory in studies elaborated from the viewpoint of mathematical methods. He also makes allowance for the special problems that a consideration of diachrony introduces. Finally, he considers a model, the elaboration of which could fulfil these requirements as well. He finds support for it in studies devoted to game 29

J. L o t m a n gives a survey of this work in the article "Metodi essati nella scienza letteraria sovietica", Strumenti critici 1 (1967), 107-129. 30 Cf., too, John Lotz, "Metric typology", Style in Language, 1960, p. 135ff.

912

F. SVEJKOVSKY

theory. The latter, as he says, provides the possibility of respecting the role of all 'partners' and 'also accounts for psychological moments' — The book by the Rumanian scholar, Solomon Marcus, Poetica matematica (1970) is one of the most comprehensive works devoted to mathematicalpoeticstoday.He considers the trendsand results of the majority of previous studies, especially the linguistic studies of poetic language and the generation of texts. Linguistic aspects are the main basis of his book. He proceeds from the opposition of scientific and poetic language. Upon this opposition he builds a mathematical model of language, and then in brief chapters he attempts to deal with a wide range of problems concerning poetics, particularly rhythm, rhyme, 'poetic figures', questions of literary genres and composition. He also devotes attention to the aspects of probability and information with respect to poetic language. In the very breadth of the questions raised it is possible to observe an effort to deal with a wide range of problems on the specific basis of the language of literature. Finally, the theater (Chapter 8: "Mathematical methods in the study of the theater") and the dramatic text receive special attention. This is noteworthy because scholarly interest in the theater and dramatic creation has come to the fore precisely in contemporary studies based on semiotics and mathematical methods. 3.4 So far the object of our examination has been theories that have had a uniform thrust: to arrive at the categories of poetics specifying the work in its generality, in its potentiality — the work in general. The characterization of these categories as 'empty forms', 'fields', or their projection into the concepts of logic and mathematical operations indicates this fact. This is also suggested by the wide scale from anthropology to logic and mathematics on which levels the approach to these categories is made today. But the opposite possibility has also been realized in contemporary works: to elaborate this theory in relation to a concrete work. This means the inclusion of other aspects of creation in theoretical poetics and proceeding from the complexity that the concretization of a literary work entails. In sum, this approach takes equal account of both components of the opposition that Jakobson expresses by the relationship code-message, Hjemslev by that of scheme-usage. In past years the theoretical aspect of a poetics directed at the work had been developed mainly in Czechoslovak and Polish Structuralism. The premises that these theories included have made it possible to develop earlier domestic theories still further and to establish contact with the evolution of scholarship during these years. This has found expression essentially in two trends: in a reconciliation with Marxist theory and in a reconciliation with the new stimuli and goals of literary structuralism in the other countries. 3.41 An evolutionary dividing line between the older and newer stages of Czechoslovak Structuralism is observed in the second half of the forties. At that time the activity of the traditional Prague Linguistic Circle ended, ideological criticism came to the fore and a confrontation with Marxist theory took place.31 Con31

The basic literature is summarized in Giulio C. Lepschy's book La linguistica

strutturale

THEORETICAL POETICS IN THE TWENTIETH CENTURY

913

sequently, the role of structuralism in literary studies was temporarily suppressed. Since the end of the fifties, however, it has begun to be applied again, and then not only in the works of scholars who had represented it in its previous phase but also in those of younger scholars. Possibilities for its application appeared as soon as attention began to be directed again toward the problems concerning the artistic qualities of the work, since these questions had previously disappeared in the sociological and ideological approaches that had dominated literary theory and practice. At the same time preconditions appeared for resuming contemporary trends of world science on the basis which the prior development of structuralism had attained in its last conclusive works at the end of the forties. These were primarily the notion of the work as a sign and, after the overcoming of certain one-sided approaches or unsolved problems, a regard for both the role of the individual and a suprapersonal 'consciousness'. These considerations made possible the broadening of the view of the literary work, which had originally been focused on the work itself. The structure of the literary work was incorporated into a broader cultural and social context by various ties (of both an artistic and extra-artistic nature); at the same time the possibility of combining the aspects of synchrony and diachrony on a common basis was shown. Thus possibilities of increasing knowledge about the dynamism of the categories of poetics were also presented. It was necessary to seek the basis for the continuity of the tradition of means and devices — the pole of permanence — in the 'collective consciousness', in a complex of supraindividual artistic experiences (traditions and conventions), whereas an individualization of means and devices, their conscious approximation of or, on the contrary, deviation from convention, norm — the pole of changeability — is attained with every concretization of a work (through the mediation of the creator and the receiver). The elaboration of historical aspects in relation to the basic theses of structuralism played an important part in the further development of these points of the theory. F. Vodicka above all played part in this. His works "Literarni historie, jeji problemy a ukoly" (1942) and Pocatky krasne prozy novoceske (1948) combine theoretical problems with the solution of concrete tasks of the historical interpretation of individual works and whole periods. In their theoretical conclusions, therefore, they are also a contribution to the study of general and historical poetics. Consideration for the relationship between a system and its realization provides the possibility of examining the categories of poetics in their complexity and dynamism. Vodicka has studied this in connection with questions of the response to the literary work in various historical and social contexts (Vodicka 1969). In Slovakia the role of initiating structuralist studies fell to J. Mukarovsky's student, M. Bakos, who has (Torino, 1966). Cf. also the literature cited above (2.4 and 2.42) and the following: Le Cercle de Prague, Change, N o . 3 (1969) 156; A Prague School reader on esthetics, literary structure, and style (ed. P. L. Garvin, Washington, D.C., 1964); J. Mukafovsky, "Kam smSfuje dnesni teorie umSni", Slovo a slovesnost 11 (1949) 49-59; J. Mukafovsky, "Ke kritice strukturalismu v naSi literdrni v2d6". Tvorba 20 (1951), 964-966.

914

F. SVEJKOVSKÍ

also combined theoretical and literary-historical considerations (e.g. in the works "Problémy literárnej vedy vcera a dnes" (1965) and "Problémy vyvinovej periodizácie slovenskej literatúry" (1944)). As we have mentioned above, it is possible to trace a confrontation with Marxist theory from the end of the forties. This has manifested itself especially with regard to historical aspects of the study of literary works and in the suppression of theoretical studies (e.g., the collision between the Marxist theory of reflection and the structuralist theory of the literary work as a sign has been symptomatic). Other new tendencies influencing the development of Czechoslovak Structuralism have derived from the stimuli provided by contemporary literary theory or poetics with a linguistic orientation and from semiotics (M. Cervenka and O. Cepan in particular have elaborated these), information theory and mathematical linguistics (e.g. J. Levy, F. Miko); in the second half of the sixties J. Levy began to develop his theory of generative poetics.32 Mukafovsky's theory of the 'semantic gesture' as the unifying principle of the work has been made use of again in studies devoted to the structure of the literary work (especially in M. Jankovic's studies). In these works the problems of poetics have been explained from different viewpoints, in that a regard for the concrete linguistic structuring of the work has remained the main starting point of these theories. Special attention has been devoted to questions of stylistics in relation to literary-theoretical problems, and a theory of the text has been worked out (L. Dolezel, K. Hausenblas, J. Mistrík, V. Marcok, F. Miko). More recent works have continued the broad tradition of studies in versification, while their methods have been amplified by contact with new currents in the study of verse (J. Hrabák, K. Horálek, V. Kochol, M. Cervenka, V. Turcány and other authors who have presented a part of their results in the collections Teorie verSe — Theory of verse).33 Attention has also been devoted to geneological questions (N. Krausová, M. Grygar) and to the relations between folklore and literature (K. Horálek). — In the article "Nová koncepce psychologie literárního tvoíení a 'psychopoetika' ", O. Sus (1966c) has deliberated about the object of 'psychopoetics' on the basis of the theory of structuralism: 'The specific object of psychopoetics thus consists not of accidental moments given by changeable individual psychology during the aesthetic consideration but of structurally motivated correlations between certain components of a literary structure and between "psycho-semantic", in a broader sense psychopoetic, components.' 3.42 Polish studies, especially those which reflected an interest in stylistics and the theory of verse, were close to the development of Czechoslovak Structuralism as early as the thirties. Since the forties Polish scholarship has resembled that in 32

J. Levy presented the concept of 'generative poetics' at the international meeting in Kazimierz nad Wisla (Poland), shortly before he died. 33 The basic surveys of the results in Czech versological studies are: E. Macek, "Ceská versologie v letech 1945-1966", Ceská literatura 14 (1966), Supplement; J. Hrabák, Üvod do teorie verse, 4th ed. (Praha, 1970).

THEORETICAL POETICS IN THE TWENTIETH CENTURY

915

Czechoslovakia in more than one respect.34 However, the theory of verse (especially in the works of M. Dhiska and J. Woronczak, a member of the following generation) and stylistics (M. R. Mayenowa, K. Budzyk) have stood their own ground. During the last decades Polish scholarship has attained a significant place in theoretical studies and has even taken great initiative, as the organization of international conferences and anthologies devoted primarily to questions of poetics testifies: Poetics. Poetyka. Poètika. (vol. I. 1961; vol. II, 1966). A generally conceived series of publications, Poetyka: Zarys encyklopediczny, in which volumes devoted to the problems of verse prevail, has been founded. M. R. Mayenowa, who is also interested in the new methods and tendencies in contemporary literary theory (esp. in stylistics, versification, semiotics), has played a major part in the organization of these works. A collection of studies, Poetyka i matematika (1965), in which the younger generation of Polish scholars made their debut, has also been published under her direction. The separate volumes of the journal Zagadnienia rodzajdw literackich have shown further possibilities for international collaboration in Poland. As its title indicates, this journal is mainly dedicated to the problems of genres; however, its contents have frequently overlapped other fields of literary theory, especially poetics. The initiative in this has come from J. Trynadlowski, W. Ostrowski, and mainly from S. Skwarczynska, the author of a comprehensive three-volume work, Wstçp do nauki o literaturze (1954-1965), which includes the entire contemporary state of theory of literature and its problems. Recently S. Skwarczynska has devoted special attention to genological studies. In his book Nauka o literaturze (1966) J. Krzyzanowski has concentrated on the place of poetics in the theory of literature and on its basic problems. An interest in stylistics pervades the work of K. Budzyk, e.g. his book Stylistyka, poetika, teoria literatury (1966). One of the most important statements devoted to the Marxist conception of poetics has also come from Poland. This is S. 2olkiewski's article "De l'intégration des études littéraires" (1961) in which the author takes into account the entire situation in world scholarship. Polish interest in a Marxist theory of literature has grown considerably since the end of the Second World War (cf. the survey in H. Markiewicz's book cited below). At the same time these studies have absorbed the stimuli of other contemporary theories and have critically dealt with simplifying sociological or ideological conceptions of literature on a global basis in order to construct a modern Marxist theory. Questions of poetics had not been systematically worked out in previous Marxist theories of literature. They had proceeded from the traditions of the older descriptive poetics, and the focal point of both theoretical interest and literary-historical and critical practice had been limited to the realiza34 For the common evolutionary features in these countries cf. O. Sus, "Typologie tzv. slovanského formalismu a problémy prechodu od formâlnich skol ke strukturalismu", Ceskoslovenské pfednâîky pro VI. mezinârodnî sjezd slavistû v Praze (Praha, 1968), p. 293, and also E. Botjâr, "L'école 'intégraliste' polonaise", Acta litteraria Academiae Scientiarum Hungaricae, Tomus 12 (1970), pp. 65-87.

916

F. SVEJKOVSKf

tion of the social function of the literary work and, in connection with this, to the relationship between the work and reality (in addition to the Soviet theoreticians, G. Lukâcs had contributed the most important works; cf. also L. Goldmann's conception of 'genetic structuralism', in which problems of the linguistic realization of the work, problems of poetics are also not prominent), ¿olkiewski, on the contrary, emphasizes the importance of poetics, above all linguistic poetics, which directs attention to the text. At the same time he refutes conceptions which proceed from the theory of the literary work as a sign and which are restricted only to the linguistic and synchronic study of works. His starting point is 'la thèse marxiste de l'unité de la langue et de la pensée, ainsi que la théorie marxiste du reflet'. On this basis he finds a focus for the future development of poetics: 'lier étroitement la poétique linguistique avec l'histoire de la littérature et cette dernière avec la sociologie et les disciplines apparentées'. Since the study of the literary work must also take into account synchrony and diachrony, whereas 'les recherches dans la domaine de la poétique linguistique ont avant tout un caractère synchronique'. Zolkiewski's comments also show how Marxist theory deals with the stimuli from semiotics and information theory. The same trends are apparent in works published in the U.S.S.R. 3.43 In his historical survey "Poètika za pjat'desjat let" V. V. Kozinov (1967) speaks about the middle fifties as a new stage in the development of research on poetics, whereas in the preceding period he finds only 'disintegration'. The signal of the changes in the fifties was, on the one hand, the revival of the earlier tradition which was represented, e.g., by the works of B. V. Tomasevskij and G. O. Vinokur and to which were added a new edition of M. Baxtin's book on Dostoevsky and further studies by V. V. Vinogradov, V. Zirmunskij and D. S. Lixacev, and, on the other hand, the publication of new works which attempted to overcome the insufficiencies of previous scholarship. The interest in an approach to the literacy work as a verbal structure has been characteristic; this interest has been closely related to the advancement of linguistics, whereby not only problems of structuralism but also those of the theory of signs have come to the fore; cf., e.g., the anthology Problemy strukturnoj lingvistiky (1963) or L. A. Abramjan's Gnoseologiceskie problemy teorii znakov (1965). In addition to V. V. Kozinov, V. V. Vinogradov has presented an evaluation of the global situation in studies of poetics in his book Stilistika — Teorija poèticeskoj reci — Poètika (1963), in which he completes his earlier studies that were characterized by an effort to reconcile linguistics and theory of literature on these very levels (cf. also his work Sjuzet i stil (1963)). L. Timofeev's book Teorija literatury (1948) presents a summary of the traditional conception and an elaboration of poetics as a description of fundamental formal means and devices. V. V. Kozinov (1967) and G. N. Pospelov (1967) have compiled a comprehensive list of recent publications. The anthologies Teorija literatury. Osnovnye problemy v istoriceskom osvescenii (1962-1965), the collection Slovo i obraz (1964) and Teorija stixa (1968) are some recent examples of the interest in special problems of poetics.

THEORETICAL POETICS IN THE TWENTIETH CENTURY

917

In the sixties, however, yet another noteworthy trend in literary-theoretical studies, a trend closely connected to theoretical poetics, appeared in the U.S.S.R. In the previously cited article V. V. Kozinov still passed over this trend with a brief, partly negative comment (1967:443-44). It involves an individual attempt to create a special type of literary structuralism on the basis of relations to the worldwide evolution of scholarship and to a considerable local tradition (from Russian Formalism to the contemporary Marxist theory of literature). Symptomatic is the effort to encompass the entire span of problems which in contemporary study have brought to the fore, both of the basic lines in the approach to the literary work, and to art in general. It is precisely here that we find the most striking connection with the new atmosphere pervading linguistic and history of art studies, an atmosphere that began to form in Soviet scholarship during the sixties. We have already had an opportunity to acquaint ourselves with certain elements of these studies above. Most of the attention devoted to poetics has proceeded from the theory of the concrete work and its historical connections. This new trend in research has been guided both by an effort to deal with the impulses of contemporary methods in literary theory and consideration for the need to surmount the previous insufficiencies in the local study of literary works. This situation was evident in the disregard for poetics, and in some cases, in the incorrect understanding of poetics as an expression of 'Formalism'. The conflict with traditional conceptions occurred right on the pages of the journals Voprosy literatury, Voprosy jazykoznanija and elsewhere;35 to some extent it was part of a more general discussion about structuralism, about the applicability of new mathematical methods, information theory and cybernetics. Results of the new theories as well as a wider local and foreign response to them have already appeared since that time. The interest directed to the study of the literary work and especially to questions of poetics has ranged from general perspectives on the theory of the sign, on the one hand, to problems of the concrete, individual work, on the other. Jurij M. Lotman most systematically has elaborated on this basis, and the institution in which he works, Tartu University, has become an important center of these studies. He presented the first synthesis in the book Lekcii po struktural'noj poUtike: Vvedenie, teorija stixa (1964). The essence of his theory is contained in two basic conceptions of art: 1. as a specific model of reality, and 2. as a sign of iconic character. The literary work is considered to be a secondary modelling system, that is, a sign system originating on the basis of a primary sign system—language. The fundamental structural notion: 'the artistic device is not a material element of the text but a relation' (1964:49) is decisive for the conception of the categories of poetics. Thus a functional viewpoint is decisive, and it replaces the previous 'nomenclatural-morphological' view. The main part of the exposition is devoted to 'the problem of the struc35

J. Lotman provides a survey in his article "Metodi esatti nella scienza letteraria sovietica" (cf. footnote 28 above), and H. Giinther does so in "Zur Strukturalismus-Diskussion in der sowjetischen Literaturwissenschaft", Die Welt der Slaven 14 (1969), 1-21.

918

F. SVEJKOVSKY

ture of verse'. The theoretical principles formulated at the beginning of the book are systematically elaborated. The functional viewpoint leads to a constant emphasis of the relation of an individual component of a work to its whole. Thus in the study of verse J. Lotman considers it necessary to proceed from broad contexts, beginning with questions of language in the literary work and the relationship prose-verse and ending with considerations of the relationship between verse and composition or subject (sjuzet). It is precisely this view of the problems of poetics in their relations, not in their separate descriptions, that is typical of Lotman's conception and that incorporates it into contemporary trends in scholarship. Moreover, a regard for linguistics is applied systematically (esp. a regard for the works of Trubetskoy, Jakobson, and contemporary Soviet linguists). Thus in Lotman, too, we can find an inclination toward the stimuli of phonology which are often applied today in structurally oriented literary studies. In his study of verse he especially develops the principle of opposition. The themes which J. Lotman outlined in his book primarily in relation to verse have been further developed into a systematic theory of the work in his recently published second book, Stuktura xudozestvennogo teksta (1970). In both of his works he directs attention to 'extra-textual structures' as well as to the structure of the work. They are, in fact, the background against which a text is perceived: 'reality, literary norms, tradition and ideas'. A consideration for aesthetic qualities and the evaluation of the work are also included in these structures. Here Lotman deals with a problem that Czechoslovak Structuralism solved analogously by referring to the 'collective consciousness'. In both cases the role of the receiver is included in the process of the concretization of the work, and there is a regard for the changeability of a work of art that stems from the confrontation of the text with a complex 'set of existential and ideological concepts'. The consistently applied principle of the relation of a certain phenomenon (hence of the literary work as a whole) to a broader context, a part of which it is, has led Lotman to the study of general aspects of art and culture. In his studies he works out these problems by proceeding from contemporary anthropological and semiotic works (e.g. "Problèmes de la typologie des cultures" (1967b)). He completes his structural conception with the relation of parts to higher wholes, and thus gets into the realm of generalizing tendencies in theory of literature. Boris A. Uspenskij's Poètika kompozicii: Struktura xudozestvennogo teksta i tipologija kompozicionnoj formy (1970), another work originating from the same circle of scholars and published in the same series as Lotman's second book, is in accord with the latter's intentions. The author understands his theme as a part of the semiotic studies which deal with art in its entirety (in 1964 he worked out a theory of art as 'a secondary modelling system'). His book, however, concentrates mainly on the literary work. It is a contribution to the typology of the modes of narration (point of view), which have already been at the center of interest for a long time and which also have had a long tradition in America (Henry James). Uspenskij for the

THEORETICAL POETICS IN THE TWENTIETH CENTURY

919

most part proceeds from the Russian tradition beginning with the Formalists. 3.44 Recently the structural theory has evoked significant responses in Italy. This can be seen not only in the effort to re-evaluate the leading orientation of prior development, the Crocean tradition, and the responses to the stylistic criticism represented by L. Spitzer, but also in the effort to re-examine a line of study which began with G. De Robertis, S. Debenedetti, G. Contini and developed in contact with Czechoslovak, Soviet and French Structuralism. Just as with New Criticism in America, these currents in Italy have been reconciled with intrinsic traditions, primarily in their emphasis on the concrete work and its analysis as the main goal of study. The pages of the journal Strumenti critici and an extensive translating activity reflect this situation in striking fashion. Russian theories from Formalism up to the present and Czechoslovak Structuralism36 have received special attention. C. Segre's I segni e la critica: Fra strutturalismo e semiologia (1969) is an expression

of the concern for contemporary problems on the basis of these modern theories. The book I metodi attuali della critica in Italia (ed. M. Corti and C. Segre, 1970)

presents a survey of the development of this study and other trends in contemporary Italian criticism. Umberto Eco, above all, has devoted special attention to questions of poetics. In the introduction to his book Opera aperta (1963 = 1967) he deals with the conceptions of poetics among the Russian Formalists and Prague Structuralists37 and in Paul Valéry (from 'Première leçon' of Cours de Poétique). Afterwards he presents his own definition: 'Noi intendiamo "poetica" in un senso più legato all'accenzione classica: non come un sistema di regole costrittive (L'Ars Poetica come norma assoluta), ma come il programma operativo che volta a volta l'artista si propone, il progetto di opera a farsi quale l'artista esplicitamente o implicitamente lo intende' (1967:8). Here, too, it is possible to find points of contact with modern views and at the same time to see new bridges between theoretical poetics and the poetics of an author or an individual work. Morover, theoretical poetics is reconciled here with the analysis and interpretation of the individual work, since the structure of the individual work is in the forefront of U. Eco's interest: "... poetica come progetto di formazione o strutturazione dell'opera' (1967:8). From the individual work this scholar envisions a path to the broader context of culture, for the elucidation of which he values the significance of Anglo-American cultural anthropology. Eco's book Opera aperta looks at problems of poetics through the prism of modern works. Upon their modes of composition he builds a model of the poetics of the 'open work', and we again encounter a characterization of the categories of poetics in their internal complexity, as investigation during the twentieth century has revealed them more 36

Cf. e.g. the anthology I sistemi di segni a lo strutturalismo sovietico, ed. R. Faccani and U . E c o (Milano, 1969) or I. Ambrogio's Formalismo e avanguardia in Russia (Roma, 1968). — Cesidi literatura 17 (1969), N o . 3. 37 T h e scholar connects these t w o theoretical traditions in a s o m e w h a t simplified manner w h e n he says that they understood poetics as 'lo studio delle strutture linguistiche di un' opera letteraria' (p. 8).

920

F. SVEJXOVSKi

than once:'. . . the work of art is becoming more and more an opera aperta, an open, ambiguous work which instead of a coherent and orderly world of values tends to suggest a "range" of meanings, a "field" of possibilities, and to this end calls more and more for active intervention, for an operative choice on the part of the reader or viewer' (1963). A developmental trend which not only solves the problems of poetics at the level of the individual work but makes it coalesce directly with its analysis and interpretation thus finds expression in U. Eco's views: '. . . poetics becomes tantamount to a concrete analysis of the works. The point is not to draw a distinction between their ugly and their beautiful aspects in order to establish which parts of them are valid, but to describe structural models. It is in this sense that the examination of poetics can be a kind of critical activity, and its results used as a contribution to the critical understanding of a particular work or writer. . . So the question now arises, . . . whether we are not fast approaching the moment when research into poetics will be the only possible form of criticism'. Although U. Eco values the significance of generalizing theories of the literary work, his own conception lies at the opposite pole, and he himself pushes the theory of the individual work to the extreme when he connects poetics with a concrete artistic expression (cf. also his work La struttura assente (1968)). Here he approaches the theories of Czechoslovak and contemporary Soviet structuralism. At the same time his works show that the most systematic theoretical work on poetics in Italy has developed primarily in studies connected with contemporary structuralism and with the interest in the analysis of the text. 3.5 The reconciliation of theoretical interest with the study of the individual work, however, was due not only to the elaboration of a theory of the concrete text and its realization by the perceiver but also to the elaboration of theoretical approaches of literary criticism, of the study of individual works. The surpassing of the former tradition of positivistic description or, on the other hand, of impressionistic criticism was sought precisely in the progressing evolution of critical activity. And therefore theoretical considerations which focused attention on questions of poetics came to the fore more and more conspicuously and systematically, especially when interest in the linguistic realization of the work was prominent. Although theoretical aspects were not the actual objective of these works and therefore were not systematically worked out, their significance for an examination of the global situation is considerable, since they not only complete the picture of this situation, but they also permit us to take a look at other spheres that have participated in the formation of theoretical poetics. 3.51 In this sense American New Criticism, above all, attracts our attention. Although concerned in their theoretical studies mainly with the 'ontological criticism' (Ransom 1941), the representatives of the New Criticism oriented their attention on that basis to the individual work and its semantic interpretation. They revealed problems which did not relate to the traditional categories of poetics but

THEORETICAL POETICS IN THE TWENTIETH CENTURY

921

to the modern expansion of its conception in relation to the very act of creation (that is to the origin and the reproduction of the explicated work) and to the work itself, to the work-sign. This approach, ahistorical in substance, tended towards the study of substantial principles and phenomena of literature which were of a perpetual nature. We can mention, for example, the concept of the literary work as an organic unity — or the main aspects of the analysis of the work, outlined by C. Brooks: paradox, irony and tension, which are the expression of an effort to discover the basic elements, principles of creation (The well wrought urn, 1947), — or W. K. Wimsatt's orientation to the criticism of the 'intentional and affective fallacy', and to the concept of the work as 'verbal icon' (The verbal icon, 1954). In Ransom's distinction between 'structure' and 'texture' we encounter a reaction to the problems that a modern conception of the work, examining both the atomization and the structural unity of an artifact, makes necessary (1947). A. Tate's notion of 'specific objectivity' of the literary work (1940) and his concern with the problem of denotation and connotation in the literature are similarly symptomatic for trends in contemporary studies. — Together with these trends it is also necessary to emphasize C. Morris's and C. S. Peirce's semantic theory; their significance grows especially from the fifties onwards, and their ideas influenced the approaches to the traditional problems of the whole theory of literature. In the same way it is necessary to weigh the separate stimuli stemming from the response to psychoanalytic theories or from cultural anthropology. These stimuli sometimes became the basis of the methods of a critical practice that was turning away from the conceptions of New Criticism. In connection with theoretical questions of poetics it is also necessary to take into account works in which these questions extended into the field of philosophy. We have in mind chiefly K. Burke's books A grammar of motives (1945), The philosophy of literary form (1951), A rhetoric of motives (1955) and S. K. Langer's Philosophy in a new key (1942) and Feeling and form (1953). In this whole extensive area of studies which deal with the work and its theoretical aspects we find many concepts coinciding with the global trend determining the contemporary approach to poetics from the broader standpoint and context. We even find correspondences in the choice of problems, — but we also find differences in their conception and solution. The variety of directions from which these questions were approached made problematic both the older conceptions of poetics and the stability of its terminology. Accordingly, the creators of new theories were often compelled to elucidate the meaning of the terms that they used; they even resorted to incorporating special glossaries into their presentations. The polemical approach to the standpoints represented by New Critics, and especially to the linguistically-oriented theories, are represented by a group of Chicago scholars. Along with their deliberations on the meaning and method of critical work, which were based on the Aristotelian tradition, these scholars revived interest in individual aspects of poetics and in the study of individual categories with respect to historical connections. On the basis of analytical studies of

922

F. SVEJKOVSKi

Aristotle's Poetics, R. S. Crane (The language of criticism and the structure of poetry (1953)) formulated the basic problems of poetics: 1. 'the primary concern of poetics will not be with the actual process of poetic creation but rather with the poetic reasoning — from the character of the end to be achieved to the necessary or desirable means — that is reflected in this process when it terminates in an artistically successful product'; 2. 'the productive science of poetry . . . will deal with necessities of two kinds — those which can be identified as essential conditions or parts of all poems, once "poems" have been distinguished from other things, and those that are specific to different poetic forms; and that, moreover, since knowledge of the former is incomplete without knowledge of the latter, poetics will aim characteristically at a causal analysis of species, or at least of those that differ most radically from the others in their particular structures and "powers" '; 3. 'since poetics is a practical science, the end of its inquiries must be the discovery of what is the best possible state, consistent with the specific natures, to which different kinds of poems and their parts may be brought; and this means, . . . that it must occupy itself not only with the necessities but also with the possibilities of the forms it treats in so far as these can be known at any given time' (1953:43-47). From the individual problems of poetics the monograph analysis of the literary genres were realized in the first place, for example in W. C. Booth's The rhetoric of fiction (1961), Eleder Olson's Tragedie and the theory of drama (1961) and The theory of comedy (1968). Also the historical dimension is a relevant aspect of the approach to the problems of poetics as was shown by the representative publication Criticism and critics (1952). And — last but not least — symptomatic was an interest in the terminology of contemporary criticism as a constituent part of the tendencies motivated by the efforts for propounding the scientific basis of the theory of literature, as, e.g. R. S. Crane's The languages of criticism and the structure of poetry (1953) or R. McKeon's essay "The philosophic bases of art and criticism" (1943) indicates. 3.52 French criticism was just as stimulating for theoretical studies of the literary work as American New Criticism. It developed chiefly on the bases of existential, phenomenological, or the Marxist approach and was given its final form by the study proceeding from mythological and archetypal theory. Its linguistic orientation received motivation not only from worldwide currents but also from local tradition, in which E. Benveniste and A. Martinet38 represented the close ties between linguistics and literary studies. Not all the development of French literary studies got as far as questions of theoretical poetics in their apparatus and interest, but when they did, they provided stimuli for and made contributions to a number of aspects of its modern makeup. Futhermore, these studies generally comprised the area in which French theoretical 38

In the book La linguistica strutturale (Torino, 1966) G. C. Lepschy devotes special attention to Martinet's work and its relation to Prague Structuralism.

THEORETICAL POETICS IN THE TWENTIETH

CENTURY

923

interest in the literary work made its appearance apart from the well defined group around R. Barthes. It is, however, characteristic, that theoretical interest existed in close relationship with the study of concrete literary creation. (Cf. e.g. S. Doubrovsky's Pourquoi la nouvelle critique (1966)). After all, the origin of the literary theory of French Structuralism had some of its roots in this concrete study, as, e.g. R. Barthes' activity or the problems to which G. Genet addressed himself indicate. In its conception the central object of French criticism — 'theme' — can serve as evidence of a phenomenon that we have already encountered more than once: the réévaluation of traditional notions of poetics against the background of modern conceptions. Here specific interest in one of the important components of the work has derived from the stimuli of Freudian and Jungian theory, and its roots have again been sought outside the boundaries of literature itself and its tradition of subjects: man's most basic experiences and situations. J.-P. Weber's work, Genèse de l'oeuvre poétique (1960), above all, has contributed to the elaboration of this theory. G. Bachelard has also arrived at the problems of certain constant-images. He elevates the role of those which are 'rooted in matter' and emphasizes the relationship between the author and 'the element which naturally corresponds to the author'. The beginnings of Bachelard's theories belong to the end of the thirties, as the chronology of his works indicates: La psychoanalyse du feu (1937), L'eau et les rêves (1942), L'air et les songes (1943), La terre et les rêveries du repos (1948), La terre et les rêveries de la Volonté (1948), La poétique de l'espace (1957) and La poétique de la rêverie (1960). The titles of these works are the best evidence of his choice of basic 'themes'. — If we take into consideration the importance of these approaches and compare them with the role of iconological studies in the visual arts (e.g. in E. Panofsky), with the studies of Russian Formalists (including the studies of folklore), with the work of E. Frenzel, Stoffe der Weltliteratur (1962), of E. R. Curtius's theory of topoi — we discover one of the stimulative importances for the perspective of theoretical poetics. French psychocriticism, a particular modification of the impulses of psychoanalysis, likewise touches upon a number of aspects of poetics as well as focusing attention on the text and analyzing it. This theory, developed by Ch. Mauron in his books devoted to literature (besides works on aesthetics and art) — Des métaphores obsédantes au mythe personel (1963) and L'inconscient dans l'oeuvre et la vie de J. Racine (1957), is essentially a reaction to two sets of problems in contemporary research, the roles of the conscious (le conscient) and the unconscious, subconscious (l'inconscient) in creation. Ch. Mauron indeed sought a way to differentiate these components in the various parts and categories of work. 3.53 While the French study of the literary work was taking place in an atmosphere pervaded by a mutual interest in literary history and the study of contemporary literature, in Germany theoretical questions arose mainly in connection with historical aspects of analysis and interpretation. Developments around the end of the

924

F. SVEJKOVSKÍ

thirties and the beginning of the forties already suggested that the inspiration for theoretical interest had shifted from the realm of elaborate systems to the study of individual works. Books such as J. Pfeiffer's Über das Dichterische und die Dichtung (1957) or H. Enders' Die Werkinterpretation (1967), e.g., besides the abovementioned works, illustrate these new paths. The participation of writers themselves, who in this atmosphere most often directed attention to theoretical problems of poetics in contemporary creation (e.g. T. Mann, B. Brecht, J. R. Becher, G. Benns, F. Dürrenmatt), gave the global situation its final shape. The range of the problems which interpretation reveals in relation to theory of literature and especially to poetics is apparent from Erwin Leibfried's broadly based work Kritische Wissenschaft vom Text (1970). As the title of the book suggests, the author identifies Literaturwissenschaft with the science of the text. His intention is to arrive at 'a system and a theory of poetology — an ontology of the text' on the basis of stimuli that the contemporary state of studies introduced. Ontological orientation opens up an approach to the text and to some aspects of poetics, particularly in the parts devoted to typological questions and the technique of narration. In his book these questions of poetics appear on a general level, for the author pursues 'eine formale Theorie der Region Text' and distinguishes it from the problems of the individual work: 'Wir trennen also streng zwischen formalen und materialen Aussagen. Formales gilt grundsätzlich für alle Erscheinungen am Text; Materiales gilt nur für bestimmte ähnliche Grupen einzelner individueller Texte' (1970:279). Consequently Leibfried's book belongs not only to the evolution of German theoretical works as a new attempt at a systematic elaboration of a wide range of questions pertaining to the literary work, but also to the entire line of contemporary studies tending toward the poetics of 'the work in general'. C. Heselhaus's critical voice has also been heard in the contemporary atmosphere of German Interpretation der Texte: 'There is concern that in some schools what began as the art of interpretation will turn into a renewed rhetorics, stylistics and science of forms' (1957:265). Opposed to these tendencies criticized by C. Heselhaus, the work of the scholars centered around the anthologies published under the symptomatic title Poetic und Hermeneutik has gained special importance in the field of interpretation. On the one hand, they have been working out the hermeneutic approach by striving to counterbalance formalizing tendencies, by emphasizing a philosophical-ideological component in the analysis and interpretation of the work; on the other hand, they have been approaching the solution of specific questions of poetics from these positions. Deserving of special attention in connection with the interest in general problems of poetics is the attitude of some scholars to studies in Slavic countries, above all the attention devoted to Russian Formalism or Czechoslovak Structuralism (besides D. Cizevskij, who in his origins and works is directly connected to these movements, we should mention especially J. Striedter and also H. R. Jauss et al.). Hermeneutics found a large reception especially in Germany owing to the older tradition (F. E. D. Schleiermacher, W. Dilthey) and to the

THEORETICAL POETICS IN THE TWENTIETH CENTURY

925

modern philosophical basis represented by H.-G. Gadamar and his work Wahrheit und Methode (19652).39 3.54 In the meantime the problems of interpretation and the renaissance of hermeneutics attained indeed an extensive reception and at the same time an influential role in the field of theoretical studies, as demonstrated by E. Betti's Teoría generóle della interpretatione (1955) and Die Hermeneutik als allgemeine Methodik der Geisteswissenschaften (1962) or by the P. Ricoeur's studies, e.g. "Le conflict des hermeunétiques: Epistémologie des interpretations" (1962), "Structure et hermeunétique" (1963); P. Ricoeur is giving his attention to the relations between hermeneutics and literary structuralism (cf. his book Le conflict des interprétations (1969)). Similarly, E. D. Hirch's Validity in interpretation (1967) or Jost Hermand's Synthetisches Interpretieren (1968) are broadly based theoretical books influencing contemporary theoretical studies.40 3.6 In order to complete the picture of contemporary theoretical studies concerning poetics we must mention works that deal with some separate aspects of literary creation and that have appeared most often in the fields of stylistics, genology or versification. These works either have originated from the stimuli of theoretical research or have been connected to the study of literary history or criticism. Very often a theoretical interest has been combined with a critical, in some cases, an historical interest. The number of works of this nature is considerable, and thus it is difficult to select from them without running the risk of distorting the picture. For this reason we can refer only to some special sources and general surveys in addition to the works that have already been mentioned. 3.61 The studies devoted to stylistics cover an immense range. Here the interests of the literary work. In America Spitzer published among other works his A method striking fashion so that the definitions of the approach to the text from these two positions, in some cases the points of contact between them, have been a frequent object of deliberation. The tradition of German stylistics, which primarily due to L. Spitzer extended to America as well, has maintained its importance for the study of the literary work. In America Spitzer published among other works his A method of interpreting literature (1949) and thus laid the groundwork for confrontations with New Criticism. Besides Spitzer, H. Hatzfeld has represented these European traditions, and E. Auerbach's works have also had an influence. At the same time, however, close ties between linguistics and literary theory have been established on the basis provided by theories of structuralism and contemporary American trends 39

Cf. R. E. Palmer, Hermeneutics (Evanston, 1969). Let us also compare H. P. H. Teesing's "Der Standort des Interpreten", Or bis litterarum 19 (1964), 31-46. Among other theoretically oriented articles let us mention: Niels Thulstrup, "An observation concerning past and present hermeneutics", Orbis litterarum 22 (1967), 24-44; J. T. Reichert, "Description and interpretation in literary criticism", The Journal of Aesthetics and Art Criticism 27 (1969), N o . 3; Michael Hancher, "The science of interpretation and the art of interpretation", Modern Language Notes 85 (1970), 791-802. 40

926

F. SVEJKOVSKÎ

in linguistics. Thus a wide and significant area of research, in which European developments have converged with the interests of scholars working mainly in the United States, has been created in the last decades. These scholars are represented especially by the anthology Style and language (ed. T. A. Sebeok, 1960), which has gained a wide response in Europe. The most recent publications including the work of the Symposium on literary style in Bellagio (Italy, 1969), Literary style: A symposium (ed. S. Chatman, 1971) testify to the continuity of this research and also to the ongoing elaboration of a number of open problems. European scholars, above all the representatives of French Structuralism, R. Barthes and T. Todorov, have also contributed to this collection. P. Guiraud's La Stylistique (1970) and his Essais de stylistique (1970) illustrate new trends in modern stylistics.41 L. Spitzer welcomed the revival of French interest in stylistic studies in the persons of P. Guiraud, G. Antoine, F. Deloffre and P. Teyssier (L. Spitzer 1961). But, the French studies represent only a part of the broad international field. It is significant that the literary text comes to the fore in the majority of these modern studies, and the theoretical works have become an important link for the study of special questions of poetics. Orientational surveys of these studies can be found particularly in the bibliographies of R. Bailey and his co-workers, Sister D. Burton (1968) and L. Dolezel (1968), the bibliographies of H. A. Hatzfeld (1953, 1961, 1966) and L. T. Milic (1967), T. Todorov's study "Les études du style" (1970), H. Seidler's book Allgemeine Stilistik (19632) and, for the Slavic area, especially in H. Markiewicz's book Glôwne problemy wiedzy o literaturze (19703). Genological studies42 comprise another, equally extensive and many-sided area of investigation. The stimuli for dealing with the most varied sets and aspects of literary genres have come simultaneously from theoretical interest and from works concerned with concrete literary texts. Typological and historical considerations stand out; the problems begin with basic questions of change and stability, of the functional aspects of literary genres, of their division — and end with separate problems of individual genres. Today there are even journals which focus special attention on genological studies — above all the Polish Zagadnienia rodzajow literackich (since 1957) or the recently founded American Genre (since 1968). From the extensive bibliography of separate works or surveys we have chosen primarily those which provide an orientation in theoretical questions: Irene Behrens, Die Lehre von der Einteilung der Dichtung, vornehmlich vom 16. bis 19. Jahrhundert (1940); Jean Suberville, Théorie de l'art et des genres (1948); E. Lämmert, Die Bauformen des Erzählens (1955); M. Fubini, Genesi e storia dei generi letterari (1956); H. Seidler, Die Dichtung (1959); F. Hellens, Les formes poétiques (1961); W. Flemming, Bausteine zur systematischen Literaturwissenschaft (1965); 41

Now we can add M. Riffaterre's Essais de stylistique (Paris, 1971), a collection of his studies from the 1960s; this scholar's work developed in close relations to the French tradition. 42 The term was used by P. van Tieghem in the essay "La question des genres littéraires", Helicon 1 (1939), 95-101.

THEORETICAL POETICS IN THE TWENTIETH CENTURY

927

S. Skwarczynska, Wstçpdonauki o literaturze, vol. Ill (Warszawa, 1965); F. Sengle, Die literarische Formenlehre (1967); W. Ruttkowski, Die literarischen Gattungen (1968); S. Petrovic, "The dictionary of literary terms and the concept of literary terminology" (1969); J. Hermand, "Probleme der heutigen Gattungsgeschichte" (1970). H. Markiewicz's Glôwne problemy wiedzy o literaturze is again the basic survey of special Slavic studies. (Refer also to the abovementioned works by N. Frye, K. Hamburger, R. Wellek and A. Warren, R. Petsch and to the theory of the Chicago School or of the French Structuralists.) 3.62 Studies in versification have expanded mainly under the influence of contemporary interest in statistical and mathematical methods in research on the literary work. They serve primarily textual criticism and the analysis of literary works. Less frequent, however, are attempts to proceed from this knowledge to the study of more general questions, as had been the case in the works of Slavic theoreticians and some other scholars, to which we have previously referred; R. Jakobson and K. Taranovski, above all, have aroused interest in these questions in America. These theories were substantially influenced by the linguistic approach. At this level the development of studies turned attention during the last decades also to semantic interpretation of versological problems (indeed the first steps were made by the formalists and structuralists of the 1920s and 1930s).

4. THE RESPONSE TO THE MAIN THEORIES IN OTHER COUNTRIES

4.1 The revelance of the theories that have come to the fore during the last decades lies not only in their contribution to the ferment and extensive development of research seeking new ways and new possibilities, but also in their significance for the global trend of contemporary European and American theory. Proof of this is the considerable impact which they have had as much — and frequently even more — upon the very practice of literary criticism and literary history as upon theoretical study. The following factors, it seems, have played not a small part in this: responses to structuralism, which has penetrated into various branches of science; the propagation of a contemporary theory of literature and an orientation according to the present state of scholarship primarily through the mediation of R. Wellek's and A. Warren's book; and finally interest in the results obtained in this field in Slavic countries. However, the influence of the main force of the twentieth century has been foremost the effort to understand the specificity of the work of art and to encompass it in a fully elaborated theory in accord with the entire cultural and social situation. 4.11 The present situation in South America, e.g., and the entire state of Hispanic studies in the world, which is closely tied to it, testify to the scope and range of these trends. The older tradition, represented by A. Alonso (cf. his work Materia y forma en poesia (1955) or by D. Alonso, and stemming from the traditions of

928

F. SVEJKOVSKÍ

stylistic interpretation, have continued to play a significant role here. In 1950 D. Alonso published a collection of essays, Poesía española: Ensayo de métodos y limites estilísticos, in which he presented his fundamental theoretical principles, primarily the consideration of analytical work for 'significante y significado' in de Saussure's sense, for 'forma exterior, forma interiorbut mainly his basic thesis: 'La Estilística será la única "ciencia de la literetura"'. In the meantime both this work and another book which he wrote with C. Bousoño, Seis calas en la expresión literaria española (3rd ed., 1963), have appeared in further editions. In the latter work an interest in the generation of the text is elaborated on the abovementioned basis, and certain principles of its modelling are asserted. In South America a response to the stylistic tradition and a relation to Alonso's studies can be found in the works of J. García Morejón (Limites de la Estilística (1961)) and, above all, Carlos Bousoño (Teroría de la expresión poética (5th ed., 1970)). Following the paths provided by D. Alonso's works, C. Bousoño arrives at a goal, which is in his words 'La teoría general sobre la expresividad poética, y más ampliamente sobre la expresividad literaria y lingüística' (1970:13). In essence, he starts from the basis of traditional poetics, or rhetorics, and in confronting it, he seeks to overcome its normativeness and descriptiveness by moving away from 'sensory' (sensibilidad) to intellectual (that is, scientific) experience. This arrival at a scientific approach to the literary work is the realization of scholarly tasks, as D. Alonso saw them in his pivotal work. However, besides this trend, which already has a firm place in the realm of Hispanic studies, we can find manifestations of contact with other contemporary movements. Evidence for this is, e.g., the anthology Anais do segundo congresso brasilie ero de critica e historia literárie (1963) or Tendencias da literatura contemporánea (1969) and Maria Luiza Ramos's Fenomenología de obra Literaria (1969). A relationship with the study of concrete literary works has been developing with regard to contemporary trends in theoretical studies: cf., e.g., Joaquim Ribiero's Teoría e hermeneutica literaria (1969) or A. Carlos Cabral's Texto praxis (1967), where attention is directed to questions of poetics, especially in the section 'Texto e derivado'. To this list may be added works written from other angles, such as F. Martinez Bonati's La estructura de la obra literaria (1960) or the collection Structural studies on Spanish themes (ed. H. R. Kahane and A. Pietrangelli 1959), which introduces other theoretical impulses into contemporary thought. 4.12 A parallel between older conceptions and the penetration of new trends has also been characteristic for northern European countries. Here F. J. Billeskov Jansen's Poetics, in particular, has occupied an important place; two volumes of it have recently been republished (Poetik: vol. I: Systemetik, vol. II: Aestetike Kritik). F. J. Billeskov Jansen again expressed the methodological basis of his theories of poetics as 'the doctrines of poetic motifs, their essence, realization, value and history' in the article "Une école norvégienne d'esthétique littéraire" (1958): ' . . . an aesthetics which makes the claim of understanding and interpreting poetry in general must take into account the biological and psychological sciences, must encompass

THEORETICAL POETICS IN THE TWENTIETH CENTURY

929

man's general essence' (1958:21f.). Evidence of other paths, particularly of the impulses of the linguistic tradition in literary theory, is provided by A. StenderPetersen's studies, e.g., his "Esquisse d'une théorie structurale de la littérature" (1949), published in Travaux du Cercle Linguistique de Copenhague. Among other works we can also mention M. Kristensen's Digningensteori (1958) and H. Hultberg's Semantisk Literaturbetragtning (1966), linked primarily to New Criticism and German studies. 4.13 Frank C. Maatje's Dutch work, Literatuurwetenschap (1970), in which special attention is devoted to problems of poetics, deals with contemporary theories on the broad basis of literary studies. Another Dutch scholar, T. A. van Dijk, approaches the literary text and poetics from a modern linguistic perspective (cf. his article "La metateoria der racconto" which refers to a book in preparation, Aspects d'une théorie générative du texte poétique, 1970:141f.). 4.14 If the evolution of theoretical literary studies in Slavic countries was concentrated mainly in the U.S.S.R., Czechoslovakia and Poland at the beginning, participation in contemporary research has increased in Yugoslavia as well during the last decades. Here we find a creative elaboration of Marxist theory together with other contemporary theoretical trends in world literary studies. The most systematic interest has been concentrated in Zagreb. In the book Udio kritike (1970), B. Donat speaks of 'the so-called Zagreb school of linguistic criticism' and classifies its work as a new stage in the continuing evolution of research on theory of literature in Yugoslavia, a stage which has been developing since the sixties (1970:150f.). An extensive translating activity comprises an element of the interest in contemporary theories; among the recent publications we should at least mention Sovjetska knjizevnost: 1917-1932 (1967), an originally conceived anthology of Russian Formalist writings that incorporates them into a wider framework of the essays of that time devoted to literature, or the collection of studies, Strukturalizam (1970). Among the original works there is the extensive Uvod u knjizevnost (ed. Fran Petré and Zdenko Skreb, 1961) which proceeds to a synthesis of contemporary theories of the literary work from the traditions of the Marxist theories of the fifties; attention is devoted to a basic characterization of the problems of poetics within this framework. The following works illustrate another relation to the literary work that has developed constantly under the influence of new theories and problems: S. Petrovic's Kritika i djelo (1963), Alexander Flaker's and Zdenko Skreb's Stilovi i razdoblja (1964) or the collection published in English, The art of the word. Umjetnost rjeci: Selected studies 1957-1967 (ed. Z. Skreb, 1969). In these works the Yugoslav theory of literature has become closely linked to contemporary problems of stylistic interpretation and to the general theory of the work; however, it has also found a path to the special study of versification. In comparison with this expansion of studies, the Bulgarian P. Zarev's work, Strukturaliz"m, literaturoznanie i esteticeski ideal (1969), testifies to a negligible interest in the more general evolution of contemporary scholarship, for even his

930

F. SVEJKOVSKtf

view of the situation in Slavic countries is very simplified. In contrast to this, we find evidence in Rumania of contact with theoretical currents in literary studies, e.g. in the anthology Metodologia istoriei si criticii literare (1969) or in A. Marino's book Introducere in critica literara (1968), which elaborates a theory of the literary work on the basis of contemporary structuralist theories. We have already mentioned above S. Marcus's book devoted to mathematical poetics.

5. HISTORICAL POETICS

5.1 Finally, we must include one other feature of modern scholarship in our examination: the development of historical poetics. During the twentieth century this area has received special attention. The problem of historical poetics has revived in particular wherever theories have renounced traditional historical approaches to the literary work or wherever attention has been directed to typological study. Thus, it came to the fore among Russian Formalists in the resumption of older local impulses, mainly A. Veselovskij's work, his plans to construct a historical poetics as the history of genres and styles. The Russian Formalists found in this possibilities to realize — in accord with their conceptions — the aspect of diachrony — to study the 'dynamics of literary forms': 'Given our understanding of literary evolution as dialectical change of forms, we did not go back to the study of those materials which had held the central position in the old-fashioned historical-literary w o r k . . . . We were interested in the very process of evolution, in the very dynamics of literary form insofar as it was possible to observe them in the facts of the past. For us, the central problem of the history of literature is the problem of evolution without personality — the study of literature as a self-formed social phenomenon', wrote B. Ejxenbaum (1965:136). And in another place he cited Jakobson's famous formulation: 'The object of the science of literature is not literature, but literariness — that is, that which makes a given work of literature.... The literary historians used everything — anthropology, psychology, politics, philosophy. Instead of a science of literature, they created a conglomeration of homespun disciplines. They seemed to have forgotten that their essays strayed into related disciplines — the history of philosophy, the history of culture, of psychology, etc. — and that these could rightly use literary masterpieces only as defective, secondary documents' (1965: 107). Jakobson wrote this at the beginning of the twenties; in his most recent studies devoted to poetics he deals with historical poetics in the following manner: 'Literary studies, with poetics as their focal point, consists like linguistics of two sets of problems; synchrony and diachrony.... Any contemporary stage is experienced in its temporal dynamics, and, on the other hand, the historical approach both in poetics and in linguistics is concerned not only with changes but also with continuous, enduring, static factors. A thoroughly comprehensive historical poetics or history of language is a superstructure to be built on a series of successive synchronic descrip-

THEORETICAL POETICS IN THE TWENTIETH CENTURY

931

tions' (1960:352). It is apparent here how new functions and aspects of historical poetics arise in connection with actual conceptions of poetics. In the introduction to the French edition of Baxtin's book on Dostoevsky, J. Kristeva, proceeding from the suggestions of this author's method, wrote: ' . . . la poétique historique s'achemine vers une définition de son objet comme type de système signifiant, et vers une conception de l'historicité propre aux modes de signifier, sans les subordonner à un déterminisme sociologique' (1970:12). Here it is evident that the problems of historical poetics are not exhausted by a mere description of the evolution of a certain phenomenon. On the contrary, new possibilities have arisen along with modern conceptions of literature and its theory; the Slovak scholar M. Bakos has formulated the scope of these possibilities the most broadly: 'In the study of literature it is historical poetics which proceeds from the aesthetic essence of artistic literature, which makes it possible to combine a synchronic (systemic) and a diachronic (evolutionary) viewpoint and which thus represents the "internal history" of literature. In a synthesis conceived in this way it is possible to integrate all the aspects of the approach to the literary work' (1969: 165).43 5.2 Along with this has gone an effort which in the field of theory bridges the labile boundary line between so-called older (or old) and modern literature. On the basis of contemporary theories and methods scholars have attempted to encompass the whole plane of literature and its theoretical problems by means of universal principles and models of literature. Evidence of this can be found primarily in P. Zumthor's Langages et techniques poétiques à l'époque romane. Xle-XIHe siècles (1963), in the studies collected in Essai de poétique médiévale (1971), in the works of T. Todorov, e.g. Grammaire du Décaméron (1969) or in J. Kristeva's Le texte du roman: Approche sémiologique d'une structure discursive transformationelle (1970). In this situation it is likewise possible to refer to parallels in the related sphere of the study of concrete works, especially to the works of the representatives of New Criticism, where the interpretation of 'specific formal qualities' have been realized in regard to the unity of both modern and older literature. 5.3 The varied realizations of these modern approaches to the poetics and to the theory of literature could arise only from the basis created by the intensive studies of former concepts and works; and this basis was arranged. It was only a natural consequence of the general orientation of the twentieth century toward the work of art and questions of its specificity that an intensified interest in them spread during this time and motivated an effort to reach both the historical sources of past creations and the appropriate literary-aesthetic principles that were prevalent during their origin in different periods. Therefore, in comparison with the last cenCf. also the theories of the Chicago School (p. 922 above).

932

F. SVEJKOVSKtf

tury v/hich was oriented mainly toward the tradition of the poetics of antiquity under the influence of classical philology, interest has now turned to the specificity of medieval and Baroque poetics in particular, and to the relationship of poetics to rhetoric, grammar, logic (i.e. to the relationship which is symptomatic of some contemporary trends in the field of theory of literature). Considering the whole complex of individual studies or books testifying to this analytic, interpretative and editorial work, it would be unfair to single out only some of them, if we cannot examine the entire situation in more detail. We select only some of them, representing the span and the variety of approaches. The following provide a general evaluation of the history of criticisms: G. Salisbury's History of criticism and literary taste in Europe (1900-1944) and Literary criticism — a short history by W. K. Wimsatt, Jr. and C. Brooks (1957). — The fundamental books representing the studies of ancient poetics are: Karl Borinski, Die Antike in Poetik und Kunsttheorie (1914-1924); C. S. Baldwin, Ancient rhetoric and poetic (1924); W. R. Roberts, Greek rhetoric and literary theory (1928); J. W. Atkins, Literary criticism in antiquity (1934); S. Francesco, Contributo alia poetica degli antichi (1961); G. M. A. Grube, The Greek and Roman critics (1965); let us also compare H. Lausberg's Handbuch der literarischen Rhetoric and R. Barthes's "L'ancienne rhétorique: Aide-mémoire" (1970). — Medieval studies reached an extraordinary broad development, as the following works indicate: C. S. Baldwin, Medieval rhetoric and poetics (1928); H. Brinkman, Zu Wesen und Form mittelalterlicher Dichtung (1928); L. Arbusow, Colores rhetorici (1948); Francesco Tateo "Retoricd' e "Poetica" fra Medioevo e Rinascimento (1960). To these books may be added works written from the broader angle of aesthetics: E. de Bruyne, Études d'esthétique médiévale (1946) and L'Esthétique du Moyen Age (1947); R. Assunto, Die Theorie des Schönen im Mittelalter (1963). The interest in Renaissance poetics is represented by the following works: D. L. Clare, Rhetoric and poetic in the Renaissance (1922); C. S. Baldwin, Renaissance literary theory and practice (1939); A. Buck, Italienische Dichtungslehren vom Mittelalter bis zum Ausgang der Renaissance (1952); G. della Volpe, Poetica del Cinquecento (1954); B. Weinberg, A history of literary criticism in the Italian Renaissance (1961); O. B. Hardinson, The enduring monument: A study in the relationship between Renaissance and literary practice (1962); B. Weinberg, Trattati di poetica del Cinquecento (1970). — The studies of the poetics of Baroque were concerned mostly with the analysis of the individual works (especially in German and Italian literature). M. Raymond's Baroque et Renaissance poétique (1955), L. Anceschi's Le poetiche del barocco (1963), and R. Montano's survey L'estetica del Rinascimento e del Barocco (1962) have a synthetical character with ample bibliographical data. — More than one of these works has even had a direct influence on the general study of theoretical poetics. Precisely for this reason, therefore, we shall be content merely to refer to two, already classic works which represent the basic tendencies and the development of the line that we have traced in the study of theoretical poetics:

THEORETICAL POETICS IN THE TWENTIETH CENTURY

933

E. Faral's Les arts poétiques du XHe-XIIIe siècle (1923) and E. R. Curtius's Europäische Literatur und lateinisches Mittelalter (1948). The first book, along with the publication of medieval texts, initiated a new stage of these studies; the second indicates the intensity that scholarship had reached in a short time. In the person of Curtius we also find a scholar's simultaneous involvement with problems of the past and the present of literature. Of the less developed studies of theoretical poetics dealing with the past of Slavic literatures we shall mention explicitly only D. S. Lixacev's synthetic work Poètika drevnerusskoj literatury (1967) and D. Cizevskij's studies combining theoretical aspects with the analysis and interpretation of works.

6. CONCLUSION

6.1 The study of literary works that began around the turn of the century initiated — as it appears — indeed a new stage in the evolution of modern theories of literary work. At the same time, attention was directed to that area which traditional poetics had occupied, and substantial changes took place. If these often occurred in pointed formulations, directed primarily against previous tradition, and if contradictory results appeared more than once, it is necessary to attribute this to the rapid pulse of the search and the exchange of views. Time itself soon gave weight to the suggestiveness of ideas. The result is a widely spread fan of knowledge and open problems, from which research is proceeding with the naturalness of the rhythm of evolution. If we consider the development of studies in the field of theoretical poetics up to now, several pivotal features draw attention to themselves. 6.11 The importance of the role of poetics in the formation of modern theory of literature has been demonstrated. At the same time, however, a decisive renunciation of normative or regulatory poetics has occurred. In connection with the broader basis of theory of literature the fundamental thread of this study has turned out to be the conclusion that a work of art is not a phenomenon which cannot be grasped by generalizing categories of science (as representatives of some conceptions emphasizing the uniqueness of every artistic act have wished), but neither is it a mere elaboration of 'content' by means of stable 'forms'. The work is approached as the lesult of an act of creation or as it is gradually being stressed more and more — as the result of an act of choice, realized on the basis of a wide field of possibilities. This field of possibilities is determined, on the one hand, by supraindividual 'artistic experience', artistic tradition and, on the other hand, by the creative participation of every subject who contributes to the realization of the work. Therefore, the work itself as a result of this act is not a sum of means and devices but a specifically hierarchized structure. The realization of the work as an aesthetic object depends not only on the activity of the originator, the author, but also on the activity of each receiver — reader, listener, spectator. Thus the various paths of research heretofore have revealed a complex set of factors which determine the final shape, the

934

F. SVEJKOVSKf

final functioning of the work. In light of this the conception of poetics and its categories has necessarily changed. Poetics no longer finds its objects completely fixed or even given a priori, i.e. existing outside the work, but precisely in the work as a resultant of an act of creation. It depends only on which of the components of this act of creation the theoretician takes into consideration. And the way is paved for a differentiation of views. 6.12 If we have just emphasized the specificity of poetics, we must also see, however, the opposite fact, namely that poetics has lost its previous, well-defined character with respect to other aspects of the theory of literature. It no longer stands isolated and definitely formed in the sense of a normative or regulatory poetics, for the new conception stemming from the dynamics of the act of creation binds it to a broader complex of phenomena, allow them to interpenetrate: poetics thus becomes an integrating component of the theory of literature. And only in this framework does it acquire full meaning, as it has been expressed; e.g., in Wellek and Warren's inclination toward the tradition of the opposition 'the extrinsic approach to the study of literature' and 'the intrinsic study of literature' or in anthropological theories, attempts to construct typologies of culture, etc. In this dual aspect of poetics—as a particular discipline at the same time incorporated into a broader context of literarytheoretical problems — we find not only a specific conception of creation but also a contemporary solution of the age-old coexistence of the two approaches accompanying the theory of literary work: the relationship between the epistemological and the technological viewpoints. 6.13 A relation to the aesthetic function of a work and to its incorporation into history has stood out as a special set of problems; this has also been reflected in modern poetics. These aspects were implied in the older poetics of a normative character, in the regulatory type of poetics. The solution of problems of aesthetics and the general theory of art during the twentieth century, on the one hand, and the differentiation between the work (text) and its concretization by the receiver, reader, on the other hand, have opened up new approaches to the aesthetics of literature. While aesthetic considerations have attracted marginal interest, rather they have been consciously detached from the problems of poetics, the necessity for taking them into account has, nevertheless, been apparent, for they participate in the final formation of the individual means and devices of creation. Similarly, consideration for the work in general, for the work in itself has led to the emphasis of a typological approach tending toward the basic principles of creation; the opposite tendencies in the study of theoretical poetics have stressed the role of the phenomenon of historicity in the definition of categories of poetics. 6.14 The focal point of theoretical poetics has shifted from the older conception of notions and categories as fixed and classificatory to an understanding of them as components of the process of creation: they include both the pole of permanence and the pole of changeability. They are encoded in experience, in consciousness, but at the same time they undergo innovations with every new applica-

THEORETICAL POETICS IN THE TWENTIETH CENTURY

935

tion, with every new concretization of a work. True, this oscillation provides various possibilities for theoretical approaches. One may emphasize the changeable and take the standpoint of the Crocean tradition. One may, on the contrary, concentrate on the phenomena of permanence and move away from the individual work or its realization and refer these problems elsewhere, as in the case, e.g., in Barthes's definition of 'la science de la littérature'. But it is also possible to build a theory on the effort to encompass the dynamics of creation and to take account of the factors of permanence and changeability at the level of concrete works, as other theories have demonstrated. And on this basis one may define the task of poetics which in its categories would reflect a concert of the individual components of communication. In this connection it is no accident that emphasis on the role of the reader, of the reader's consciousness, has appeared precisely in recent times, as we can observe in the theory of Czechoslovak Structuralism. Besides F. Vodicka's works let us mention M. Cervenka's article "Literârni dilo jako znak" (1969), or the other studies: H. Weinrich's "Fiir eine Literaturgeschichte des Lesers" (1967), M. Gïowiâski's "Wirtualny odbiorca w strukurze utworu poetckiego" (1967) or W. Iser's Die Appellstruktur der Texte (1970). 6.15 Two basic poles, which still determine the extensive field of present-day studies on poetics, have been prominent in the twentieth-century trends that we have just mentioned: a) One has brought poetics into relation with generalizing and synthesizing tendencies in contemporary scholarship. The act of creation, rather the work itself, has been examined primarily in its essence, which appears at the most basic levels of human activity and which sometimes even gives it a validity existing outside of time and space. Consequently, ways to link these theories to anthropology, philosophy and even to logic have become evident. Possibilities of maximal formalization have become evident and have also been indicated. Opportunities for development in the realm of contemporary semiotics (semiology) have been revealed. A modern formulation of constants, permanent principles and categories, which authors of poetics and poetic theories had been seeking since long ago, has been able to be realized together with this. At the same time consideration for the essence of creation has included aspects of the genesis — or of the generation — of the literary work, and therefore even in this sense it is a new variant of questions examined before on different planes or by means of different methodological devices. b) The other pole has led to development of theories that permit the problems of the work to be viewed from the opposite side: they have oriented poetics toward the individual artifact. Attention has been directed to the study of the system which is the starting point as well as the basis for the realization of the concrete work and which is again, on the contrary, determined by a relation to the level of concrete creation. In opposition to the interest in reaching the permanent, general phenomena constituting the essence of a literary work, consideration for the oscillation between the permanent and the changeable, between a system and an individual act of

936

F. SVEJKOVSKY

creation has primarily been asserted here, for contact with the work in its real shape (in its definite appearance and in its contact with the receiver) has made this necessary. What Mukafovsky had formulated generally is essentially valid here for the conception of the categories of poetics: 'A concept appears... as an energetic means of a constantly renewed apprehension of reality, a means always capable of internal reorganization and assimilation'. The fact that this conception tends primarily toward structural analysis and in its consequences is close to the realms of interpretation or literary history is connected with the approach to the workreality, not to the work in its potential existence and generality. 6.2 Therefore the two basic approaches to the problems of the work, when they are viewed through the prism of poetics, complement one another. The signals of contact with the 'eternal' problems that theory poses with regard to the literary work resound once again in this differentiation and, above all, in the goals that are set. Since antiquity two possibilities of approaching creation have been considered: to direct one's approach either toward methodology, hence abstractly and theoretically, or toward history, exegetically, toward 'poetarum ennarationem', as Quintilian said.

SELECTED REFERENCES

and S. D . N . BURTON. 1 9 6 8 . English stylistics: A bibliography. Cambridge. BAILEY, RICHARD W., and L . DOLEZEL. 1 9 6 8 . An annotated bibliography of statistical stylistics. Ann Arbor, Mich. BAKO§, MIKULÁS. 1 9 6 9 . O problémoch umeleckej a vedeckej avantgardy. Avantgarda 3 8 . 1 5 8 - 6 9 . Bratislava. BARTHES, ROLAND. 1 9 6 4 . L'activité structurelle. Essais critiques. Paris. English trans, by Richard Howard, Partisan Review 3 4 . 8 2 - 8 8 (Winter, 1 9 6 7 ) . . 1966. Critique et vérité. Paris. . 1 9 6 8 . Linguistique et littérature. Langages 1 2 . 3 - 8 . . 1970. L'ancienne rhétorique: Aide-mémoire. Communications 16. BIERWISCH, MANFRED. 1965. Poetik und Linguistik. Mathematik und Dichtung, ed. by H . Kohlhase and R . Grunzenhäuser, pp. 4 9 - 6 5 . München. BILLESKOV JANSEN, FREDERIC J . 1958. Une école norvégienne d'esthétique littéraire. Orbis litterarum 1 3 , Supplementum 1 1 . 2 1 - 2 8 . BOUSOÑO, CARLOS. 1 9 7 0 . Teoría de la expresión poética. 5th ed. Madrid. BOSCH, RUDOLPH. 1928. Die Problemstellung der Poetik. Eine historische Untersuchung über die Methoden und Grenzen wissenschaftlicher Wertbestimmung. Leipzig. BRIK, OSIP. 1 9 2 9 . T.n.formal'nyj metod. Lef 1 . 2 1 3 ff. CERVENKA, MIROSLAV. 1 9 6 9 . Literární dílo jako znak. Orientace 4 . 5 8 . - 6 4 .

BAILEY, RICHARD W . ,

THEORETICAL POETICS IN THE TWENTIETH CENTURY

937

1966. Estetika Jana Mukaïovského. In Studie z estetiky, by Jan Mukarovskf, pp. 341-364. Praha. CZAPLEJEWICZ, EUGENIUSZ. 1 9 6 5 . Matematická poetika a Studium poezie. Ceská literatura 13.419-27. Polish version in Wiersz i pozija, pp. 81 ff. Wroclaw, Warszawa, and Kraków, 1966. DIWALD, HELMUT. 1 9 6 3 . Wilhelm Dilthey. Erkenntnisstheorie und Philosophie der Geschichte. Göttingen. DONAT, BRONIMIR. 1 9 7 0 . Udio kritike. Zagreb. Eco, UMBERTO. 1 9 6 3 . The analysis of structure. Critics Abroad 7 , The Times Literary Supplement, September 27, 1963. . 1967. Opera aperta. 2nd. ed. Milano. . 1968. La struttura assente. Milano. ÊJXENBAUM, BORIS. 1924. Lermontov. Opyt istoriko-literaturnoj ocenki. Leningrad. . 1927. Teorija 'formal'nogo' metoda. Literatura. Teorija. Kritika. Polemika. Leningrad. Trans, in Russian Formalist criticism: Four essays, by L. T. Lemon and M. J. Reis. Lincoln. Neb, 1965. ELIOT, THOMAS STEARNS. 1 9 2 1 . The sacred wood. New York. . 1932. Selected essays. New York. ERLICH, VICTOR. 1 9 6 9 . Russian Formalism. 3rd ed. The Hague and Paris. FAGES, JEAN-PAUL. 1 9 6 7 . Comprendre le structuralisme. Toulouse. FRYE, NORTHROP. 1 9 6 5 . Anatomy of criticism. 7th ed. Princeton, N . J . GLOWINSKI, MICHAL. 1967. Wirtualny odbiorca w strukturze utworu poetyckiego. Studia z teorii i historii poezji, pp. 7-32. Wróclaw, Warszawa, and Kraków. GREIMAS, ALGIRDAS JULIEN. 1 9 6 7 . La linguistique structurale et la poétique. Revue internationale des Sciences Sociales 1 9 / 1 . 8 - 1 6 . HAMBURGER, KATE. 1957. Die Logik der Dichtung. Stuttgart. HATZFELD, HELMUT A. 1953. A critical bibliography of the new stylistics, applied to the Romance literatures: 1900-1952. Chapel Hill, N. C. . 1966. A critical bibliography of the new stylistics, applied to the Romance literatures: 1953-1956. Chapel Hill, N. C. HATZFELD, HELMUT A . , and YVES L E HIR. 1 9 6 1 . Essais de bibliographie critique de stilistique française et romane ( 1 9 5 5 - 1 9 6 0 ) . HERMAND, JOST. 1 9 7 0 . Probleme der heutigen Gattungsgeschichte. Jahrbuch für internationale Germanistik 2 / 1 . 8 5 - 9 4 . HESELHAUS, CLEMENS. 1 9 5 7 . Auslegung und Erkenntniss. Gestaltprobleme der Dichtung. Bonn. JAKOBSON, ROMAN. 1 9 2 1 . Novejsaja russkaja poèzija. Praha. . 1928. Problemy izucenija literatury i jazyka. Novyj lef 12. 36-37. Reprinted in Readings in Russian poetics, pp. 101-102. Ann Arbor, Mich., 1962. . 1952. Axioms of a versification system — exemplified by the Mordvinian CHVATÎK, KVÈTOSLAV.

938

F. SVEJKOVSKY

folk song. Acta Instituti Hungarici Universitatis Holmiensis, Series B, Linguistica 1.5-13 (1941-1945). . 1960a. Linguistics and poetics. Style in language, ed. by T. A. Sebeok, pp. 350-77. New York. . 1960b. Poetry of grammar and grammar of poetry. Poetik. Poetics. Poetyka. Poètika, pp. 397-417. Warszawa. . 1970. Poesie und Sprachkultur, ed. by F. Ph. Ingold. Zürich. JAKOBSON, ROMAN and PETR BOGATYREV. 1929. Die Folklore als eine besondere Form des Schaffens. Donum Notalicium Schrijnen, pp. 9 0 0 - 1 3 . Nijmegen —Utrecht. Reprinted in R. Jakobson, Selected writings IV, pp. 1-15. The Hague and Paris. and . 1931. K probleme razmezavanija fol'kloristiki i literaturovedenija. Lud Slowianski 2. 230-33. Reprinted in R. Jakobson, Selected writings IV, pp. 16-18. The Hague and Paris. JAUSS, HANS ROBERT. 1967. Literaturgeschichte als Provokation der Literaturwissenschaft. Konstanz. KOZINOV, VADIM V . 1967. Poètika za pjatdesjat let. Izvestija Akademii Nauk SSSR, Serija literatury i jazyka, T . 2 6 . 4 3 2 - 5 6 . KRIDL, MANFRED. 1 9 5 1 . The integral method of literary scholarship: Theses for discussion. Comparative Literature 3 . 1 8 - 3 1 . KRISTEVA, JULIA. 1 9 6 8 . Le texte clos. Langages 1 2 . 1 0 3 - 2 1 . . 1970. Une poétique russe. Introduction to Mikhaïl Bakhtine, La poétique de Dostoïevski. Paris. KRZYZANOWSKI, JULIAN. 1966. Nauka o literaturze. Wroclaw, Warszawa and Kraków. LEHMANN, RUDOLF. 1908. Deutsche Poetik. München. LEIBFRIED, ERWIN. 1970. Kritische Wissenschaft vom Text. Stuttgart. LEMPICKI, ZYGMUNT. 1921. W sprawie uzasadnienia poetiky czystej. Przeglqd Filozoficzny 2 3 ( 1 9 2 0 ) . 1 7 1 - 8 8 . Republished in the collection Teoria badan literackich w Polsce I. Krakow, 1960. LEPSCHY, GIULIO C . 1 9 6 6 . La linguistica strutturale. Torino. German trans, with new bibliographical data: Die strukturale Sprachwissenschaft. München, 1969.

M. 1964. Lekcii po struktural'noj poètike. Vvedenie. Teorija sticha. Reprinted as Brown University Slavic Reprint V, with an introduction by Thomas G. Winner. Providence, R.I., 1968. . 1967a. Metodi esatti nella scienza sovietica. Strumenti critici 1.107-29. . 1967b. Problèmes de la typologie des cultures. SocScil 6/2-3. 29 ff. MARKWARD, BRUNO. 1961. Das Verhältnis von formulierter und werkimmanenter Poetik. Poetics. Poetyka. Poètika, pp. 733-44. Warszawa. MEYER, THEODOR A . 1901. Das Stilgesetz der Poesie. Leipzig. MILIC, LOUIS T . 1967. Style and stylistics: An analytical bibliography. New York. LOTMAN, JURIJ

THEORETICAL POETICS IN THE TWENTIETH CENTURY MORPURGO-TAGLIABUE,

GUIDO.

939

1967. Linguistics e stilistica di Aristotele.

Roma. 1944a. Morphologische Poetik. Helicon 5. 1-22. . 1944b. Die Gestaltfrage in der Literaturwissenschaft und Goethes Morphologie. Halle, Saale. MÜLLER-FREINFELS, RICHARD. 1 9 1 4 . Poetik. Leipzig and Berlin. MUKAROVSKÎ, JAN. 1934. K ceskému pfekladu Sklovského Teorie prözy. Cin 4. 123-30. Reprinted in Kapitoly z ceské poetiky I, p. 344-50. Praha, 1948. . 1936a. Strukturalismus v estetice i ve vëdë o literature. Ottüv slovnik naucny nové doby 6/1. Reprinted in Kapitoly z ceské poetiky I, pp. 13-17. Praha, 1948. . 1936b. L'art comme fait sémiologique. Actes du huitième Congrès international de philosophie à Prague 1934, p.p. 1065.-73. Prague. Czech version in Studie z estetiky, pp. 85-88, Praha, 1966. . 1966. O strukuralismu [the lecture from 1946], published in Studie z estetiky, pp. 109-16. Praha, 1966. MURPHY, JAMES. 1965. Rhetoric in fourteenth-century Oxford. Medium Aevum 34. 1 - 2 0 . PETROVIC, SVETOZAR. 1 9 6 9 . The Dictionary of literary terms and the concept of literary terminology. Art of the world (Umjetnost rijeci). Zagreb. PETSCH, ROBERT. 1940. Deutsche Literaturwissenschaft. Berlin. POMORSKA, KRYSTYNA. 1 9 6 8 . Russian Formalist theory and its poetic ambiance. The Hague and Paris. POSPELOV, GENNADIÏ N. 1967. Metodologiceskoe razvitie sovetskogo literaturovedenija. Sovetskoe literaturovedenie za pjat' desjat let. Moskva. 1910. The spirit of romance. London. POUND, EZRA. . 1950. The letters of Ezra Pound, ed. by D. D. Paige. New York; quotation from: W. K. Wimsatt and C. Brooks, Literary criticism, p. 662-63. New York, 1957. ROETTEKEN, HUBERT, 1 9 0 2 . Poetik, I . An additional part was published in Euphorion 2 5 ( 1 9 2 4 ) . SKWARCZYNSKA, STEFANIA. 1 9 3 8 . Przedmiot, metoda i zadania teorii literatury. Pamigtnik Literacki, 1938; it was reprinted in the collection Teoria badan literackich w Polsce, II, pp. 270 ff; Krakow 1960; the citation is to the latter. SPITZER, LEO. 1 9 6 1 . Les études de style et les différents pays. Langue et Littérature. Actes du Ville Congrès de la Fédération Internationale des Langues et Littératures Modernes, pp. 2 3 - 3 8 . Paris. SPOERRI, THEOPHIL. 1942. Uber Literaturwissenschaft und Stilistik. Trivium 1. 1-3. STAIGER, EMIL. 1939. Die Zeit als Einbildungskraft des Dichters, Zürich and Leipzig. . 1 9 4 5 . Versuch über den Begriff des Schönen. Trivium 3 . 1 8 5 - 9 7 . MÜLLER, GÜNTHER.

940

F. SVEJKOVSKÍ

. 1948. Zum Problem der Poetik. Trivium 6. 274-96. . 1961. Andeutung einer Musterpoetik. Unterscheidung und Bewahrung f. Hermann Kunisch, pp. 353-62. Berlin. Sus, OLEG. 1964. Anfänge der semantischen Analyse in der tschechischen Poetik: Josef Durdik und seine Theorie der dichterischen Sprache. Zagadnienia rodzajów literackich 7. 42-56. . 1966a. Geneze sémantiky umení v ceské tvarové estetice. Literaria 9. 196-224. Bratislava. . 1966b. Ke vzniku sémantické typologie v estetice Otakara Zicha. K prehistorii ceského strukturalismu. Ceská literatura 14. 393-415. . 1966c. Nová koncepce psychologie literárního tvorení a 'psychopoetika'. Slovenská literatura, 1913. 32-53. . 1967. Predystorija cesskogo strukturalizma i russkaja formal' naja ¡skola. Ceskoslovenská rusistika, 1 2 . 2 2 9 - 3 5 . TATE, ALLEN. 1 9 4 0 . Miss Emily and the bibliographer. Collected essays. Denver, 1959. TODOROV, TZVETAN. 1965. Théorie de la littérature: Textes de formalistes russes. Paris. . 1968. Poétique. Qu'est-ce que le structuralisme? Paris. . 1970. Les études du style. Poétique 1. 224-32. TOMASEVSKIJ, BORIS V . 1927. Teorija literatury. Poètika. 2nd ed. Moskva and Leningrad. TOPOROV, VLADIMIR N. 1962. K analizu struktury litovskoj narodnoj ballady. Simpozium po strukturnomu izuceniju znakovych sistem. Tezisy dokladov. Moskva. TOPOROV,V. N . and V . V . IVANOV. 1963. K rekonstrukcii praslavjanskogo teksta. Slavjanskoe jazykoznanie 5. 88 ff. TROCZYNSKI, KONSTANTY. 1 9 2 9 . Przedmiot i podzial nauki o literaturze. Quotations from the edition in Teoría badañ literackich w Polsce, II. Kraków, 1960. TYNJANOV, JURIJ. 1929. Literaturnyj fakt. Arxaisty i novatory. Leningrad. Reprinted in Readings in Russian Poetics, rev. ed. Ann Arbor, 1971. VACHEK, JOSEF. 1 9 6 4 . A Prague School reader in linguistics: Studies in the history and theory of linguistics. Bloomington, Ind. . 1966. The Linguistic School of Prague: An introduction to its theory and practice. Bloomington, Ind. VAN DIJK, TEUN A . 1970. La metateoria del racconto. Strumenti critici 4 . 1 4 1 63. VODICKA, FELIX. 1966. Celistvost literárního procesu. Struktura a smysl literárního dfla, pp. 8 7 - 1 0 7 . Praha. . 1969. Struktura vyvoje. Praha.

THEORETICAL POETICS IN THE TWENTIETH CENTURY WALZEL, OSKAR.

1925.

WEINRICH, HARALD.

Gehalt und Gestalt im Kunstwerk der Dichtung. Für eine Literaturgeschichte des Lesers. Merkur

1967.

941

21.

1026-1138. 1957. Wilhelm Dilthey's poetics and literary theory. Wächter und Hüter, Festschrift für Hermann J. Weigand, pp. 121-32. New Haven,

WELLEK, R E N É . 1957.

. 1969. Literary theory, criticism and history. Concepts of criticism. 5th ed. New Haven. . 1971. The literary theory and aesthetics of the Prague School. Discriminations. 2nd ed. New Haven. WIMSATT, W . K. J R . and C . BROOKS. 1957. Literary criticism: A short history. New York. ZIRMUNSKIJ, VIKTOR. 1923. K voprosu o formal' nom metode. Problema formy v poèzii. Peterburg. . 1925. Formprobleme in der russischen Literaturwissenschaft. Zeitschrift für slavische Philologie 1.117-52. . 1928. Zadaci poètiki. Voprosy teorii literatury. Stat'i 1916-1926. Leningrad. ZÖLKIEWSKI, STEFAN. 1961. De l'intégration des études littéraires. Poetics. Poetyka. Poètika, p. 7 6 3 - 9 4 . Warszawa. ZOLKOVSKIJ, ALEKSANDR K., and Ju. K. SCEGLOV. 1962. O voznoznostjach postroenii strukturnoj poètiki. Simpozium po strukturnomu izuceniju znakovych sistem. Tezisy dokladov, pp. 138-41. Moskva. . 1967a. Iz predystorii sovetskix rabot po strukturnoj poètike. Trudy po znakovym sistemam 3 ( = Ucenye zapisky Tartuskogo Gosudartsvennogo Universiteta, vol. 198). Tartu. . 1967b. Strukturnaja poètika — porozdajuscaja poètika. Voprosy literatury 11.74-89.

RHETORIC A N D STYLISTICS

P. GUIRAUD

Literature is an art of language insofar as it uses language with regard to specific ends, means, and situations. The conditions which generate the structure of ordinary social communication are changed, and literary form is submitted to influences of its own. This special quality of literary language has been acknowledged since the most ancient times, and its study is, in the West, the object of Rhetoric which we inherited from the Greeks through Rome. RHETORIC is the technique of literary creation, as well as the description and codification of its precepts. It is based upon the theory of genres which is a classification of various literary works according to their end: aesthetic, polemic, didactic, etc.; and their means and media: writing, singing, reciting, etc. The GENRE determines both the form of the expression and the content. The theory of expression — which is the one most elaborated — has resulted in inventories of figures: phonetic, lexical, syntactic. The theory of content deals with COMPOSITION or the ordering of the various parts and arguments, and with INVENTION which comprises an inventory of topoi or common themes with rules for expanding them and setting them, according to well-tried canons, into NARRATIONS, DESCRIPTIONS, PROSOPOPEA, etc. There is no need, here, to repeat the reasons why the system broke down during the nineteenth century, leaving the author his inspiration and the critic his subjective evaluation, evaluation more or less intuitive and most of the time tautological. We simply wish to show how, in the last fifty years, a new theory of literary creation and expression has been built in the vacuum left by rhetoric. First, under the name of STYLISTICS various methods of studying expression have been devised. During this time — after a period of disaffection — a new interest in genres has developed. On the one hand, the notions of function and communication have shown their importance in new genres like the mystery novel, the musical comedy, press editorials, advertising, etc. On the other hand, the difficulty exists of understanding older works without a knowledge of the conditions under which they were created. Through this new approach, more recently, stress has been put upon the form of content which appears as a system of signs, at an upper level, invested with a semiotic function of its own; this system is studied according to models borrowed from current linguistic analysis.

944

P. GUIRAUD

So, a new rhetoric is being built within two main disciplines: stylistics and literary semiotics. The first is a study of linguistic form which is sometimes called LINGUISTIC STYLISTICS; the second, a study of content called LITERARY STYLISTICS, THEORY OF LITERATURE, NEW CRITICISM, etc. Both borrow their concepts, their models, and their methods from modern linguistics. For one who knows the history of recent linguistics — the mutations, the contradictions, and the polemics of the various schools — it is easy to understand the proliferation and often the confusion of our discipline. Every reader is a stylistician and every one has his own approach. So, facing the impossible task of choosing between thousands of works, we will content ourselves here with a broad survey of the subject. I. STYLISTICS A.

The Structure of the Code

Stylistics is the study of the specific form of the message. This definition covers extensive ground, which includes various types of studies, so we must agree on a model and a terminology. We will choose the one that Roman Jakobson borrowed from the theory of communication in order to distinguish between the different linguistic functions. Here it is slightly simplified: code

author

message

— reader

content The form of a message depends, of course, on its content; it depends on the language selected by the author, on the author himself, and on its destination which implies both a medium and a reader. Ultimately — and this is one of the contributions of modern stylistics — the message itself determines and generates its own form. Setting aside, for the present (cf. below p. 953) the form of the content, we have three main approaches: a stylistics of the code (descriptive stylistics), a stylistics of the author (genetic), and a stylistics of the reader (functional). At the same time these three types of studies have two aspects, depending on whether one looks at the data in relation to the text itself or in relation to the linguistic norm. There are two distinct problems: either you study the general value of a metaphor, a regionalism, etc., or you study the effects of these forms in a specific context. So we must distinguish between the form, function, origin, etc., of the language which generates the message and the structure of the message itself.

RHETORIC AND STYLISTICS

945

General bibliography The following are publications of general interest covering the various aspects of the subject of stylistics. These include bibliographies, general theoretical works, manuals and readers, as well as collections of papers and the proceedings of congresses or symposia. Works pertaining to specialized fields and approaches within stylistics will be listed at the end of the appropriate subsections. Bibliographies BAILLY, R. W., and D. M. BURTON. 1968. English stylistics. A bibliography. Cambridge, Mass., The M.I.T. Press. HATZFELD, H . 1 9 6 0 ( 1 9 5 3 - 5 5 ) . A critical bibliography of the new stylistics applied to the romance literatures ( 1 9 0 0 - 1 9 5 2 ) . Chapel Hill, N.C. HATZFELD, H . , and Y . L E H I R . 1 9 6 1 . Essai de bibliographie critique de stylistique française et romane. Paris. MILIC, L. T. 1967. Style and stylistics. An analytical bibliography. New York. SHAPIRO, K. 1948. A bibliography of modem prosody. Baltimore. SHTOKMAR, J. 1933. Bibliografija rabot pro stixoslozeniju. Moscow. Dictionaries and encyclopedias LE HIR, Y. 1956. Esthétique et structure du vers français d'après les théoriciens du XVIème siècle à nos jours. Paris. . 1960. Rhétorique et stylistique de la Pleiade au Parnasse. Paris. MORIER, H. 1961. Dictionnaire de poétique et de rhétorique. Paris, P.U.F. PREMINGER, A . , ed. 1 9 6 5 . Encyclopedia of poetry and poetics. Princeton. VACHEK, J. 1 9 6 0 . Dictionnaire de linguistique de l'Ecole de Prague. Paris. Theories of style Linguistique générale et linguistique française. Berne. Lingua e poesia. Firenze. BUDZYK, K. 1946. Stylistyka teoretyczna w Polsce. Warsaw. DRESSOIR, M. 1948. Die Rede als Kunst. München. GRANGER, G . C. 1968. Essai d'une philosophie du style. Paris. GUIRAUD, P. 1 9 7 0 . La stylistique. 6th ed. Paris. NOWOTTNY, W. 1962. The language poets use. London. TOMASEVSKIJ, B . 1 9 2 9 . O stixe. Moscow. . 1958. Stix i jazyk. Moscow. TROJAN, F. 1952. Der Ausdruck der Sprachstimme. Vienne. ULLMANN, S. 1964. Language and style. Oxford. BALLY, C H .

BERTONI, G .

1944.

1937.

Collections of papers, proceedings of congresses and symposia and S . R . LEVIN. 1 9 6 7 . Essays on the language of literature. Boston.

CHATMAN, S . ,

946

P. GUIRAUD

J., and J. SUMPF. 1969. L'analyse du discours. Langages 13 (March). N. E . , J. SPENCER, and M. J. GREGORY. 1964. Linguistics and style. London. FOWLER, R., ed. 1966. Essays on style and language. London. GUIRAUD, P., and P. KUENTZ. 1 9 7 0 . Lectures stylistiques. Paris. JACOB, A. 1969. Points de vue sur le langage. Paris. Langue et littérature. Actes du VHIème Congrès de la FILLM. 1961. Liège. Linguistique et littérature. 1968. Langages 12 (December). Littérature et stylistique. 1964. Cahiers de l'AIEF 16 (March). Mathematik und Dichtung: Versuche zur Frage einer exakten Literaturwissenschaft, ed. by R. Gunzenhâuser and H. Kreuzer. 1965. Miinchen. Poetics, Pœtyka, Poetika. 1961. Warsaw and The Hague. Proceedings of the IXth International Congress of Linguists. 1964. The Hague. Recherches de stylistiques. 1967. Cahier du CRAL 2. Nancy. SEBEOK, THOMAS A., ed. 1960. Style in language. Boston and New York. STRELKA, P. JOSEPH, ed. 1970. Patterns in literary style. Philadelphia. Style et littérature. 1962. The Hague. La stylistique. 1969. Langue française 3 (Septembre). Théories et problèmes. 1958. Orbis, Litterarum Supplement 2. VACHEK, J. 1964. A Prague School reader in linguistics. Bloomington. DUBOIS,

ENKVIST,

1. Descriptive

stylistics

The distinction between form of language and form of message is essential in judging the works of Ch. Bally, the father of our discipline. Bally, a Geneva linguist and follower of Saussure, was the first to conceive, under the name of stylistique what he defines as: 'the study of expressive linguistic facts from the point of view of their affective content: that is to say, the expression of sensibility by means of language and the action of language or sensibility (Bally 1904:1. 16). He takes care to specify that he studies the affective values (which he calls stylistiques) of signs in language and not the effects the author derives from them in the text. Bally does not study style (i.e. the form of the message), but, of course, he describes the linguistic values which make style possible. He distinguishes between NATURAL EFFECTS and EFFECTS BY EVOCATION, which distinction corresponds to what one might call similarity and contiguity values (one might consider the affective, social, aesthetic, etc., connotations here). His work, which studies sounds, words, and constructions, corresponds to the analysis of 'figures' made by traditional rhetoric. But whereas the latter is based on the authority of the great authors, Bally's stylistics proceeds from a consideration of linguistic form and function as it was conceived at the beginning of the century. Bally has had a very large following and his work is still remarkably relevant and inspiring. Yet, Bally's school could be accused of too often confusing code and

RHETORIC AND STYLISTICS

947

message levels and of finding in the latter stylistic effects which are not necessarily there. And of course its linguistic models are usually obsolete. In this regard, Roman Jakobson renews the subject by introducing modern criteria such as the concept of shifter or of metaphor as opposed to metonymy (cf. below p. 948). It is evident that modern linguistics—functional, distributional, transformational, generative — will furnish descriptive stylistics with new concepts and new methodologies. In fact new models are already being investigated. See one of the numbers of Langages (13 mars 1969) which is devoted to DISCOURSE ANALYSIS formulated in terms of transformational and generative grammar. One will also see the contribution of stylistics to the last International Congress of Linguists. But young stylisticians, until now, are more interested in the message than in the code. In fact, many of them deny the validity of a descriptive stylistics, at the code level, as conceived by Bally. They argue that style is a message property and stylistics the study of the message in itself. While agreeing with this postulate, one might still think that message description is impossible without the previous study of the linguistic values at code level, which generate stylistic effects in the message. Bibliography — Descriptive

linguistics

1948. Colores rhetorici. Gottingen. BALLY, CH. 1 9 0 4 . Traité de stylistique française. Heidelberg. BLACK, M. 1 9 6 2 . Models and metaphor. Ithaca. BODKIN, M . 1 9 6 3 . Archetypal patterns in poetry. London. BROOKE-ROSE, C . 1 9 5 8 . A grammar of metaphor. London. BURKE, K . 1 9 6 2 . A grammar of motives and a rhetoric of motives. Cleveland. COHEN, J. 1966. Structure du langage poétique. Paris. CRESSOT, M . 1 9 6 9 . Le style et ses techniques. Rev. ed. Paris. DELBOUILLE, P. 1 9 6 1 . Poésie et sonorités. Paris. DEVOTO, G. 1962. Nuovi studi di stilistica. Firenze. GRAMMONT, M . 1 9 1 3 . Le vers français, ses moyens d'expression, son harmonie. Paris. GUIRAUD, P. 1 9 5 3 . Langage et versification d'après l'œuvre de Paul Valéry. Paris. KOCH, W. A. 1966. Recurrence and a three-modal approach to poetry. The Hague. MAROUZEAU, J. 1950. Précis de stylistique française. 3rd ed. Paris. MORIER, H . 1943-44. Le rythme du vers libre symboliste étudié chez Verhaeren, Henri de Régnier, Vielé-Griffin et ses relations avec le sense. Genève. SPIRE, A. 1949. Plaisir poétique et plaisir musculaire. Essai sur l'évolution des techniques poétiques. Paris. ARBUSOW, L .

2. Functional stylistics

Every message has a destination. It is oriented towards a reader (or a hearer) to

948

P. GUIRAUD

whom the author wants to transmit some specific information, in some specific condition. The theory of genres, as conceived by Rhetoric, is a classification of literary works according to their function and the ends and media adapted to these functions. The notion of GENRES, of course, still survives in various works devoted to STYLE: poetic, dramatic, historical, narrative, and so on. At the same time, these studies, often a bit out of date, are today renewed by the new criteria borrowed from modern linguistics. The notion of function applied to descriptive stylistics (cf. above p. 946) shows that each sign has potential values which make it more or less adapted to the expression of a specific situation and genre. For instance, there is a specific use of personal pronouns or verbal tenses in the novel. This is a long recognized fact, but functional linguistics explains the source and mechanism of usage, until now studied only intuitively. Take, for instance, the notion of SHIFTER as analyzed by Jakobson. One knows that linguists call SHIFTERS or INDICATORS such signs as personal pronouns, demonstratives, verbal tenses, time and space adverbs, etc., which derive their signification from the person of the locutor and from the time and space of the communication. In every communication there are two subjects, the one who is speaking and the one spoken of; and in the same way time and space are double. The opposition and relation between time of speaking and time spoken of plays a major function in any literary work, especially in the novel. The manipulation of shifters allows the displacement of these two times whether they are separated or combined. Not least interesting is Jakobson's definition of metaphor, a figure of similarity as opposed to metonymy, a figure of contiguity. Here again well known facts are put in a new light as Jakobson shows their stylistic function in literature, based on similarity (Romanticism, Symbolism) or on contiguity (Realism, Surrealism). Jakobson, in the same way, shows the critical power of such notions as syntagmatic and paradigmatic axes, defining poetry as 'the projection of the principle of equivalence from the axis of the selection on the axis of the combination'. The idea — foreign to rhetoric — of an individual style expressing the nature of the author, even his subconscious, leads to the notion of WRITING (écriture) as defined by Roland Barthes in Le degré zéro de l'écriture (1953). As opposed to style, écriture consists of the various means of expression an author obtains from his time, his class, a group, a school, the means by which he imitates and through which he signals his adhesion to an aesthetic, a moral, a philosophy, etc. There has been considerable interest recently in the various forms of écriture: the novel, the political essay or, even, advertisement, and the folk song. And this neo-rhetoric was bound to focus attention again once on traditional rhetoric. The domain of medieval literature — among others — has been thoroughly investigated from a stylistic and rhetorical approach. Works that were considered 'without style' and the expression of a naïve literature are now recognized as the heritage of a very ancient and sophisticated rhetorical tradition. E. Curtius's monu-

RHETORIC AND STYLISTICS

949

mental work, Europäische Literatur und lateinisches Mittelalter (1948), opened the way for these studies among which are J. Rychner's essay on epic form (1955), R. Dragonetti (1960) and P. Zumthor (1963) on poetic technique, etc. By this stylistic and technical analysis medieval literary criticism is at last escaping an approach which was mainly external and historical. Bibliography — Functional stylistics BARTHES, R. 1953. Le degré zéro de l'écriture. Paris. BOAS, F. 1948. Race, language and culture. New York. CURTIUS, E. 1 9 4 8 . Europäische Literatur und lateinisches Mittelalter. Berne. DRAGONETTI, R . 1 9 6 0 . La technique poétique des trouvères dans la chanson courtoise. Bruges. GUIRAUD, P. 1 9 7 0 . Essais de stylistique. Paris. JAKOBSON, R . 1 9 6 1 . Poetry of grammar and grammar of poetry. Poetics. Warsaw and The Hague. . 1963. Essais de linguistique générale, tr. by N. Ruwet. Paris. . 1968. Selected writings, Vol. IV. The Hague. KROEBER, A . L . 1 9 5 7 . Style and civilizations. Ithaca. RYCHNER, J. 1 9 5 5 . La chanson de geste: Essai sur l'art épique des jongleurs. Genève. ZUMTHOR, P. 1 9 6 3 . Langues et techniques poétique à l'époque romane. Paris. 3. Genetic stylistics The postulate that 'style is man' leads modern stylistics to the attempt to define the relation between the form of the message and its author. This is a difficult and controversial problem which, until now, does not seem to have been solved satisfactorily. Even if it is clear that a text depends on the experience, culture, and nature of its author, one might, nevertheless, question the validity of this postulate at the level of style. The form of the message seems to be linked much more to the culture of the writer than to his individual nature and depends on the rhetorical habits of a time, a group, or a school. Barthes' definition of écriture (see above p. 948) opposes this concept to the one of style as the expression of the inherent, unsophisticated nature of the author. But it seems that style, according to Barthes' critical work, would be a characteristic of content more than of linguistic expression. The same conclusion can be drawn from new literary criticism, especially as reflected by its psychoanalytical trends (cf. below p. 954). Indeed, there are many studies devoted to the description or origin of an author's style. But they are usually based on compilations of words, figures, and constructions; and if there is such a thing as an individual language, it must be a system of relations and not a mere inventory of inorganic forms. What is specific to a language — and consequently stylistic — is not its isolated elements but their internal organization.

950

P. GUIRAUD

This is the problem which Leo Spitzer, the master of genetic stylistics, tries to solve. According to him, a literary work is a structured whole, organized around a vision of the universe specific to the author. This 'spiritual etymon' is manifested by specific linguistic features, which are individual stylistic deviations. The task of criticism is to grasp — intuitively — this generative principle of the text, which explains its deep meaning. Spitzer's Stilstudien will remain a monument of modern STYLISTICS. Nevertheless his method, mainly intuitive and subjective, has not attracted any follower of consequence, at least up to now. Pierre Guiraud has tried another structural approach with his theory of STYLISTIC FIELD. He treats the text as a code, proceeding to a distributional analysis of the various forms. Each word's connotation and denotation is inferred from a study of its contexts. The text is analyzed as a closed system, in which each term is defined according to its position in relation to others. The specificity of the distribution as compared with standard usage defines its stylistic value. Thus we hope to reconstruct the work's code, showing the specific net of relations on which it is built and, eventually, establishing its origin and biographical or cultural sources. The problem has been approached in a more objective way with the attempt to define a psychology of style, that is, to link the use of definite linguistic forms to corresponding psychological categories. An example of this method is La psychologie des styles (1959) by Henri Morier; it is a very interesting and often convincing attempt to give an objective linguistic analysis of such intuitive criteria as 'affected', 'mannered', 'biting', and so forth. But it does not seem that stylistic psychology has yet found pertinent criteria in traditional psychology. But Jakobson's analysis of metaphor and metonymy (cf. above p. 948) and the study he made of these two functions in both language pathology and poetic creation show that the attempt to establish a genetic typology of style is not unjustified. Notions like 'time and aspect', 'actual and virtual', 'definite and indefinite', 'intrinsic and extrinsic determination', etc., correspond to a psychology of style that a genetic stylistics might some day be able to use. Guillaume's psychosystematics, for instance, or Hjelmslev's homology between expression and content forms, open the way to such an approach; and generative linguistics is already tending in this direction. Bibliography — Genetic stylistics 1928. Styles et physiologie: Petite histoire naturelle des écrivains. Paris. CROCE, B. 1923. Rettoria, grammatica e filosofia del languaggio. Problemi di estetica. Bari. MORIER, H. 1959. La psychologie des styles. Genève. SPITZER, L. 1931. Romanische Stil- und Literaturstudien. Marburg. . 1957. Language of poetry: Language on enquiry into its meaning and function. New York. . 1961. Stilstudien. Rev. ed. Miinchen. CHASSE, CH.

RHETORIC AND STYLISTICS

951

. 1962. Linguistics and literary history. Rev. ed. New York. . 1966. Eine Methode Literatur zu interpretieren. Rev. ed. München. 4. Quantitative stylistics Considered as an individual deviation from standard usage, style seems the natural domain for statistical analysis. However, the validity of a quantitative approach has been questioned especially by the structuralist schools. The main problem is the definition of a pertinent norm without which the study of deviations is useless. Too often, also, statisticians confound the quantitative and qualitative levels, unable up to now to define their functional relation. Most statistical studies, moreover, consist of inorganic inventories of numerical deviations and coefficients which are rejected by the stylisticians interested in the evaluation of literary and aesthetic values, or those who — from a structural perspective — define style as formal opposition. And, of course, to postulate a stylistic function inherent in the message excludes any reference to a quantitative analysis insofar as stylistic effect is unique (cf. below p. 952). Nevertheless, and notwithstanding the fact that statistical stylistics has not fully confirmed the hopes of some fifteen years ago, a large amount of research, both methodological and pragmatic, has been done in the field, and one cannot see any reason for stylistics to refuse the help of a quantitative study carefully managed. Bibliography — Quantitative stylistics FUCHS, W . 1 9 5 2 . On mathematical analysis of style. Broomerike. GUIRAUD, P. 1 9 5 4 . Les caractères statistiques du vocabulaire. Paris. . 1960. Problèmes et méthodes de la statistique linguistique. Paris. GUIRAUD, P . , and J. WHATMOUGH. 1 9 5 4 . Bibliographie de la statistique linguistique. Utrecht. HERDAN, G. 1 9 5 6 . Language as choice and chance. Groningen. . 1966. The advanced theory of language as choice and chance. Berlin and New York. KREUZER, H., and R. GUNZENHXUSER, eds. 1 9 6 2 . Mathematik und Dichtung. Versuche zur Frage einer exakten Literaturwissenschaft. München. Kvantitavni linguistika. 1963-66. Nos. 2 and 10. Prague. MULLER, CH. 1968. Initiation à la statistique linguistique. Paris. Statistical methods in linguistics. 1962. No. 1. Stockholm.

B. The Structure of the Message The object of stylistics is the study of the form of the message; and this study must be preceded by the analysis of expressive resources at the linguistic level as we have already defined them. Some stylisticians would not however agree. M. Riffaterre,

952

P. GUIRAUD

especially, denies any stylistic value to the study of language at code level as conceived by Bally. He rejects the notion of stylistic deviation and, consequently, any kind of quantitative approach. Finally, he turns down the very notion of individual language on the ground that stylistic impression belongs, not to the author, but to the reader decoding the message. We won't go into this discussion here. We agree, that Bally's epigones too often confound stylistic values with their effects in the discourse. Moreover, we believe that their criteria, most of the time, ignore the accomplishments of modern linguistics. As for genetic stylistics, we have pointed out its limitations which result, probably, from the lack of adequate psychological and sociological criteria. Therefore, it is not without reason — no matter what might or might not be the validity of a stylistics at code level — that the emphasis must be placed upon the structure of the message, which is, at any rate, the proper end of our discipline. Its task is to study the form of the message in itself. This methodological postulate was stated by the Russian formalists as far back as the early twenties and upheld by their heirs, the Prague structuralists, whose Roman Jakobson remains the most eminent representative. The basis of Jakobson's method of description lies in his definition of the poetic function (in its widest meaning): 'the projection of the principle of equivalence from the paradigmatic axis in the syntagmatic axis'. That means, as is well known, that there is a correspondence between sound or syntactic structure and meaning, between meter and rhythm, and between other kind of parallelisms. But Jakobson's formulation, once again, opens the way to more precise and objective criteria. Most of Jakobson's views on the subject will be found in "Poetry of grammar and grammar of poetry", to be reprinted in the third volume of his Selected writings. To this school belongs S. R. Levin's Linguistic structures in poetry (1962) and his theory of COUPLING, which is an inventory of the various structural patterns 'in which semantically and/or phonically equivalent forms occur in equivalent syntagmatic positions, the forms so occurring thus constituting special types of paradigms'. M. Riffaterre stresses the inherence of the stylistic effect as a property of the message in itself. The stylistic effect depends on two main structures, CONVERGENCE and CONTRAST. The convergence on the same sign of different linguistic values is akin to Jakobson's principle of equivalence and Levin's coupling. As for the contrastive effect, it is linked with the situation of the sign in the message — the effect, for instance, of an archaism in an archaic context is different from that of an archaism in a modern context. As far as we are concerned, we see no contradiction between a stylistics of the message and a stylistics of the code (cf. above p. 944); on the contrary, they are complementary. But the structural analysis of the message stresses the necessity of maintaining a clear distinction between the two levels and the awareness of the very common confusion between stylistic value and stylistic effect. It shows, at the same time, the usefulness of new criteria borrowed from modern linguistics. This new

RHETORIC AND STYLISTICS

953

structural stylistics, until now, has been mainly distributionalist; but new approaches — transformational and generative — are already in the making. Bibliography

— The structure of the message

R. 1961. Poetry of grammar and grammar of poetry. Poetics. Warsaw and The Hague. LEVIN, S. R. 1962. Linguistic structures in poetry. The Hague. RIFFATERRE, M . 1 9 5 9 . Criteria for style analysis. SLAMA-CAZACU, T. 1 9 6 1 . Langage et contexte. The Hague. The subject, comparatively new, has not been the object of a comprehensive work. But one finds the beginnings in the numerous examples in the recent collective works in the general bibliography (p. 945). JAKOBSON,

II. LITERARY SEMIOTICS

The specific nature of the literary message is manifested both at the level of expression and content, a fact which was acknowledged by rhetoric when it described for every genre categories of argumentation, rules for composition, and procedures of invention. The decay of rhetoric and the separation of philology and the history of literature into two distinct disciplines lead to the disruption of this critical system towards the end of the nineteenth century. Literary criticism, then, leaves to grammarians the study of the linguistic expression, retaining the commentary on the content. Commentary, which may be brilliant when penned by some critics, remains for the most part empty paraphrase: verses are turned into prose, and one explains laboriously to the reader what the author wanted to say or ought to have said. At the same time, philologists, for their part, draw up inventories of what are supposed to be the most conspicuous forms. We have seen how, from this philological analysis, a STYLISTICS derived, which, little by little, was built upon the new postulates and the new observations of modern grammar. Literary analysis was slower in finding its way, especially due to the fact that it remained for so long under the influence of the historical method, devoted to the external problem of sources, influences, and biographies. The polemic between the internal critic and the external one was mainly begun by stylisticians, especially Leo Spitzer (cf. above p. 950) who conceived of an internal criticism in terms of linguistic forms. The Russian formalists, who as early as the twenties studied the structure of content, were unknown outside their country where the movement miscarried anyway. However, it was revived through the Prague School with the study of content structure in relation to the various genres. At the same time there developed a study of literary themes and their symbolic meaning conceived as signs and therefore structured into a system. Both disciplines borrow their methods and epistemological frame from linguistics,

954

P. GUIRAUD

following in this instance the example of social sciences like anthropology and sociology. Until now, these sciences have mainly used structuralist models; but like stylistics, literary semiotics certainly has a few things to learn from transformational and generative grammar. Literary semiotics is now flourishing in such movements as Anglo-Saxon New Criticism, German Literaturwissenschaft, French Nouvelle Critique, etc.

1. The Semantics of Literature Literary themes have always been studied: love, death, nature, and so on. Modern psychology has recognized the deep roots as well as the symbolic function and universality of such themes as the Sea, the Tree, the Light, the Circle, and so on. Definitive in this regard is the work of Gaston Bachelard on poetic imagination, which shows the relation of literary images to the four alchemical elements: fire, earth, air, water. Proceeding from psychoanalysis — especially Jung's philosophy of archetypes — Bachelard shows that these elementary metaphors are the expression of a poetic experience: water, for instance, is, at a deep and unconscious level, an image of maternity, fire is the symbol of sexuality, etc. From a linguistic approach it appears that the reality thus suggested works to signify another reality. And this is the general case in literary communication; for instance, Charles Bovary's cap, in the first chapter of the novel, is the sign of a complete psychological and sociological situation; so is the evening dress of Emma, or the wedding cake. In a literary work there are two levels of meaning: the things, the characters, their relations and behavior, linguistically significant at one level, have a meaning of their own. Immediate literary content works as a means of expression. This approach is the basis of many works on literary symbolism — especially in France. See, for instance G. Durand's L'imagination symbolique, J. Rousset's Forme et signification, Ch. Mauron's Des métaphores obsédantes au mythe personnel, and so on. More recently this type of study has been formalized by adapting analytical models to modern linguistics. Such is Lévi-Strauss's approach to the study of myths where the literary content is reduced to a set of semic elements organized in a system of oppositions on the model of phonology. Such systems have been discussed and elaborated by structural semantics as well, especially by A. J. Greimas and Barthes.

2. The Morphology of Literature Literary semantics is the study of the metaphoric and symbolic values of the various themes, such as mountains or rivers, wine or blood, and so on. But a literary work

RHETORIC AND STYLISTICS

955

also contains characters, situations, actions, and events. It has always been one of the main objects of criticism to reduce these elements to categories, such as: the lover, the traitor, the ingénue, etc. And, of course, we have always known that the various genres are variations of a small number of patterns. But more recently such patterns have been formalized. For instance, Etienne Souriau's Les deux cent mille situations dramatiques (1950) shows that dramatic situations are built upon the combination of six functions corresponding to six main characters: the Arbiter, the Opponent, etc. The problem had been already posed by Russian formalism; the classic work in this regard is V. J. Propp's Morphology of the folktale, first published, in Russian, in 1928, and first translated into English in 1958. The author defines FUNCTION as 'the action of a character, defined from the point of view of its meaning for the development of the tale as a whole'. The problem is to define and classify the elementary functions of which each tale is a specific combination. The book, which escaped notice at the time, attracted attention with the publication of such authors as Lévi-Strauss, Dumézil, Barthes, Greimas, etc. Novels or myths are thus reduced to their immediate constituents called mythèmes, narrèmes, etc., and organized into a system from which each novel or myth takes its structure and meaning. Models have been elaborated and the review Communications has devoted a number of studies to narrative structure. Bibliography — Literary semiotics DURAND, G . 1 9 0 0 . L'imagination symbolique. Paris. GREIMAS, A . J . 1 9 6 6 . Sémantique structurale: Recherche de méthode. Paris. LÉVI-STRAUSS, C . 1 9 5 8 . Anthropologie structurale. Paris. MAURON, C H . 1 9 0 0 . Des métaphores obsédantes au mythe personnel. Paris. PROPP, V. J . 1 9 5 8 . Morphology of the folktale. 2nd. ed. Austin. ROUSSET, J . 1 9 0 0 . Forme et signification. Paris. SOURIAU, E. 1950. Les deux cent mille situations dramatique. Paris.

LITERARY GENRES

TZVETAN TODOROV

The problem of genres has been discussed so often, and is today so complex, that when dealing with them one must choose between writing a book and presenting a general theoretical overview. The present paper represents the latter solution. First of all, we must eliminate one problem which is raised by the superficial identification of genres with the names of genres. As we all know, some genres have names, sometimes very popular ones, such as tragedy, comedy, sonnet, etc. But it is obvious that if we want to use the concept of genre in a general literary theory, we cannot always trust a historical tradition full of contingencies: some genres never received a name; others, although different, were confused under the same name. The study of genres must take place on the level of structural characteristics, not according to prior naming. But even this apparently very simple problem is not simple at all, since it raises the question which is central to all study of genres: their relation to history. Indeed, two entirely different approaches to the phenomenon of genre can be observed in the existing studies. The first is inductive: it records the existence of genres, starting with the observation of a period of literary history. The second is deductive: it postulates the existence of genres starting with a theory of literary discourse. Although some aspects of one are always related to the other, each approach has its own concepts, methods, and techniques, and the difference between them is so great that one may wonder if it is at all useful to allow that these two kinds of studies have a common object, if it would not be better to speak of GENRES, in the first case, and of TYPES, in the second. Here is an example. Suppose we discover that in the classical period in France, one could characterize the contemporary tragedy by its 'seriousness of action' and the 'dignity of characters'. Starting with this single observation we can follow two entirely different paths and produce two altogether distinct studies. In the first case we will proceed as follows: 1) we will have to justify the use of the terms 'characters' and 'action'; we will observe that they are necessarily present — or not — in any literary text; 2) then we will try to establish a list of properties which will possibly specify this general term: for instance, we will formulate two opposite cases, the characters are either 'worthy' or 'base'; 3) next we will have to study every kind of character (or action, etc.) and these very categories will become the subject of our study; the

958

TZVETAN TODOROV

different combinations of these properties are called literary types. These do not necessarily have a precise historical existence: either they correspond to historical genres or to models of style which were present in very different historical periods; or perhaps they do not correspond to anything, and this is probably the most interesting case: they resemble an empty slot in Mendeleyev's periodical system, and some literary work of the future may fill it up. But if we take a closer look at that sketch of a typological study, we see that there is no significant difference between it and studies in general poetics; 'typological' is here synonymous with structural; the initial observation on genres was just a convenient starting point for an exploration of literary discourse in general. But we can also follow a very different course, starting with the same initial observation on classical tragedy. We will now collect all the works in which we can find the described properties; and these works will form the corpus of 'French classical tragedies'. The concept of DOMINANT FEATURE used by the Russian Formalists, however hard to define, can be applied here: indeed, in order to claim that this work is a tragedy we must be able to show that the relevant features are not only present but also dominant, in a ruling position. From that point on, we will no longer be interested in the categories of literary discourse, but in a certain reaction to literature, proper to this period, which we can observe both in the writer and in the reader. In the first case, it is a certain model of style to which the author refers, even if he wishes to violate it; in the second, it is a certain horizon of expectation, i.e. a set of pre-existing rules which orient the reader's understanding and allow him to receive and to appreciate the text. These are genres: they form a system within any period and they can only be defined by their mutual relations. In this sense there is no genre of tragedy: tragedy must be redefined in each period, in accordance with the other contemporary literary genres. Here we leave general poetics and enter literary history. We can observe the same opposition between genre and type if we compare the two with the individual work. We must distinguish here three cases. In the first, the individual work obeys totally the rules of the genre; this is the case of popular or mass literature. A good detective novel, for instance, does not claim to be original (and if it does it is less 'detective' and more 'novel'), but simply to apply well the prescription. In the second case genre rules are transgressed. We must first note that a work does not always belong to a genre: it is true that each period is dominated by a system of genres but this system does not necessarily cover all the works published during that time. On the other hand, a certain partial transgression of the genre is almost obligatory, otherwise the work cannot be considered sufficiently 'original' (this requirement has varied very strongly throughout history). A transgression of genre rules does not affect the literary system very profoundly. If for instance tragedy implies that the hero dies at the end, and if in one play we find instead a happy ending, this will be a transgression of genre rules. Since the genres still cover a large portion of the literary potentialities of a time, such a transgression

LITERARY GENRES

959

will usually be interpreted as a mixture of genres, here of tragedy and comedy. The idea of mixed genres is the product of a confrontation between two genre systems: the mixture exists only in the terms of the anterior system; any evolution, seen from the past, is mere degradation. But as soon as the 'mixture' becomes a part of a new system, it also becomes an independent genre, as was the case of tragicomedy in France. Finally we can also observe — but much more rarely — type transgressions. The system of literary discourse is not fixed, given once and for all; the whole set of literary potentialities can be transformed, and that is why type transgressions are also possible. If we take the same example again, we axe dealing with a type transgression if the writer has found a new category, neither comic nor tragic (in other words, the genre trangression corresponds to the relation 'both A and non-A', while the type transgression to 'neither A nor non-A'). To transgress a genre rule means to choose a way which was already potentially — but not actually — present in the synchronic literary system; the typological transgression aims at this very system. Ulysses of James Joyce is not only a transgression of the contemporary rules of the novel, but also a discovery of new possibilities for fiction writing in general. Thus the opposition between type and genre may be useful; but at the same time it should not be considered absolute. There is no profound separation between system and history, between the abstract and the concrete, but rather different degrees of inscription in time. This inscription is weakest in the case of types; but as we just saw types are not set apart from history either. It is stronger in the case of a genre: basically a genre exists within the limits of a period; nevertheless, some genre features are conserved after the period during which they were fixed, as were the rules of the tragedy in eighteenth century France. Finally, at the other end of this continuum we find the PERIODS themselves. Indeed, when we speak of Romanticism, or of Symbolism, or of Surrealism, we suppose just as in the case of genres that a set of features is predominant in a group of literary works. The difference is that a period can (and usually does) include many genres, and that it cannot be extracted in any way from history. The period is not a purely literary notion but participates in the history of ideas, of culture, and even of society. We possess innumerable classifications of genres; but very rarely are they based on a clear and coherent idea of the status of genres. Two fallacies are particularly frequent: 1) the confusion of genres and types or, more precisely, the description of genres as if they were types; and 2) the reduction to simple opposition, between a feature and its negation, of what is in fact the conjunction of several distinct categories. More often than not the level of abstraction has not been specified: it is obvious that a genre can be qualified by a few or by many properties; thus some genres include others, that is, have a greater extension. We will look briefly at some of the best known genre classifications. 1. PROSE-POETRY. This very broad opposition is rarely made explicit; there is

960

TZVETAN TODOROV

even a certain ambiguity concerning the meaning of the word 'prose': sometimes it means 'literary prose', other times 'non-literary discourse'. If we accept the first meaning, we shall still observe that this opposition cannot be reduced to a single category: one alternative opposes VERSE to PROSE (depending on the rhythmical or graphic organization of the text; the existence of free verse and of prose poems creates a serious problem here); and another alternative opposes POETRY to FICTION, i.e. a type of discourse which must be read on the level of its meaning, as a pure semantic configuration, is opposed to a representational (mimetic) discourse which we interpret on the level of its designata, as a description of a human universe. We may also add some stylistic prescriptions: emotive, figurative, personal styles are predominant in 'poetry' while 'fiction' is characterized mainly by a referential style, as well as some thematic features: in the first case a relation between an ego and the universe is usually described, while in the second an interhuman relation is generally the topic. Let us add that contemporary literature often tries to neutralize this opposition: the contemporary 'novel' requires a 'poetic' reading, as a semantic construction and not as the representation of an extratextual world. 2. LYRIC-EPIC-DRAMATIC. From Plato to Emil Staiger, many attempts have been made to describe these three categories as the fundamental and even the 'natural' forms of all literature. I wonder if this is not in reality a system of genres characterizing Ancient Greek literature, presented now unjustifiably as a system of types. In this case an effort has been made to discover the underlying categories of the division. Diomedes, in the fourth century, systematizing Plato, suggests the following definitions: lyric = where only the author speaks; dramatic = where only the characters speak; epic = where both are allowed to speak. This is certainly a very clear and formal definition, however one wonders if the selected structural property carries the necessary importance. Goethe distinguishes between poetic 'modes' (which correspond more or less to our 'genres': ode, ballade, etc.) and the 'natural forms of poetry' (analogous to our 'types'), and he asserts: 'Es gibt nur drei echte Naturformen der Poesie: die klar erzählende, die enthusiastisch aufgeregte und die persönlich handelnde: Epos, Lyrik und Drama . . . ' One may interpret this formula, based on the three protagonists of the act of uttering: he (epic), I (lyric), you (dramatic). Roman Jakobson has followed this line of thought: ' . . . können wir sagen dass für die Lyrik der Ausgangspunkt und das führende Thema stets die erste Person der Gegenwartszeit und für das Epos die dritte Person der Vergangenheit i s t . . I n his book dealing with the 'fundamental concepts' of poetics, Emil Staiger explains the three-fold division primarily by time reference: lyric = present; epic = past; dramatic = future (this correspondence was first established by the German Romantic poet Jean Paul). At the same time he uses other categories as well: he speaks of 'lyrische Ergriffenheit', 'epische Uberschau', and 'dramatische Spannung'. He has also insisted on the necessity of distinguishing the categories (or types) epiclyric-dramatic from their concrete realizations (genres) epos-lyrical poetry-drama. Thus in the case of Jakobson and Staiger the division is based on linguistic cate-

961

LITERARY GENRES

gories but these are not immediately observable, as was the case with Diomedes. But if even this division could be established on reasonable evidence, the question still remains — does it really have the importance that literary history has given it? 3. TRAGEDY-COMEDY. This is also a very old division but not as universal as the preceding one. The confusion between genres and types has been very frequent here. Aristotle introduced the opposition without defining it; the Italian and French classics characterized tragedy by seriousness of action, dignity of characters, and the unhappy ending — and comedy by the opposites of these features. Such a definition is clearly generic; Northrop Frye has tried to formulate a typological one: 'There are thus four main types of mythical movement: within romance, within experience, down, and up. The downward movement is the tragic movement.... The upward movement is the comic movement...' (the two other categories are the romantic and the ironic). Let us note here that 1) other esthetic categories have sometimes been put on the same level, such as: the sublime, the grotesque, the marvelous, etc., and 2) each one of these divisions can be further subdivided and specified, e.g. for comedy, into farce, vaudeville, burlesque, etc. 4. The rhetoric of the Middle Ages usually maintained a division into three styles: high, medium, and low, often illustrated by the different works of Virgil; but the word 'style' here covers a meaning closer to our 'genres' or 'types'. It is based both on some purely linguistic features, and on the social level of the characters represented. 5. Another effort to justify genre divisions by 'nature', i.e. by language, was made by André Jolles who deals with the SIMPLE FORMS of literature. According to Jolles, the contemporary literary forms (= genres) are derived from linguistic forms; but this derivation (more logical than historical) is not direct, it passed through an intermediary stage of 'simple forms' which one finds mainly in folklore. Thus, we have a three-level derivation. Jolles' inquiry may be summarized by the following diagram: Frage

Aussage

Schweigen

Befehl

Wunsch

Realistisch

Kasus

Sage

Rätsel

Spruch

Fabel

Idealistisch

Mythe

Memorabile

Witz

Legende

Märchen

His description is certainly very sketchy but it has positive characteristics: the plurality of initial categories (and thus, the greater complexity of the classification); and the attention to genres such as riddles, proverbs, etc., which do have a significant impact on other more elaborated linguistic forms. There has been (and there will be) many other genre classifications. One should not try either to keep them all or to renounce the idea of genres, simply because of this

962

TZVETAN TODOROV

very multiplicity: these classifications are a natural product of any reflection on the internal properties of literary texts. Genre criticism is one of the most fruitful kinds of criticism if the scholar only remembers that, as Northrop Frye has put it, its purpose is to clarify, not to classify.

BIBLIOGRAPHY

1. References to this Article FRYE, NORTHROP.

1957.

Press.

Anatomy of criticism. Princeton, Princeton University

1 9 3 5 . Randbemerkungen zur Prosa des Dichters Pasternak. Slavische Rundschau 7.357-74. JOLLES, ANDRE. 1 9 3 0 . Einfache Formen. Halle, Max Niemeyer Verlag. STAIGER, EMIL. 1 9 4 6 . Grundbegriffe der Poetik. Bern, Francke. JAKOBSON, ROMAN.

2. Bibliographies of Genre Studies BEHRENS, IRENE.

heft 92.

1940.

Die Lehre von der Einteilung der Dichtkunst. ZRPh Bei-

RUTTKOWSKI, WOLFGANG VICTOR. 1 9 6 8 .

Die literarischen Gattungen. Bern, Francke.

3. General 1943,1949. The theory of literary kinds. I: Ancient classifications of literature. II: The ancient classes of poetry. Dubuque, Iowa. HEMPFER, KLAUS W . 1 9 7 3 . Gattungstheorie. München, W . Fink Verlag. JAUSS, HANS-ROBERT. 1970. Littérature médiévale et théorie des genres. Poétique 1.79-101. LÄMMERT, EBERHARD. 1955. Bauformen des Erzählens, 9-18. Stuttgart, Metzler. TIEGHEIM, PHILIPPE VAN. 1 9 3 8 . L a question des genres littéraires. Helicon 1 . 9 5 101. TODOROV, TZVETAN, ed. 1965. Théorie de la littérature: Textes des Formalistes russes, 126-8, 302-7. Paris, Seuil. . 1970. Introduction à la littérature fantastique, 7-27. Paris, Seuil. DONOHUE, J . J .

METRICS

JOHN LOTZ

METRICS

Metrics ( < *(T6XVT|) ^ E X P I K T ) ) is the study of the numerical and prosodic characteristics of the phonetic material in certain language texts called verse. Counteracting a predominantly literary approach of recent times, the study of meter has again become a specific linguistic concern, following in this respect an earlier grammatical tradition which usually included a special appendix or section on metrics in grammars.

VERSE A N D PROSE

Versification, the production of verse ( < Latin versum facere 'to make verse'), implies a specific, converted, or, if you wish, perverted use of language.1 The term verse (Latin versus, '(di)verted'), though referring to the graphic representation of verse in lines, expresses this insight, setting verse in opposition to prose ( < Latin (oratio) pro(or)sa, 'straight forward talk'). The problem of the boundaries between prose and verse is a delicate one; no sharp defining characteristic feature separates the two modes. We have rather the opposition of a small set of language texts characterized by numerical regularity of speech material within certain syntactic frames as contrasted with texts which lack such a characteristic. The outward boundary of verse is so-called free verse, regarded by many as not a bona fide metric phenomenon.2 Verse and prose are opposed to each other as two types of which one, verse, has definable properties, and the other, prose, is characterized by lack of any such features. The members of an unbalanced opposition, such as that between verse and prose, are called 'marked' vs. 'unmarked'. 3 1

This article is essentially an adaptation of a paper prepared for the Modern Language Association of America's publication, Versification-. Major language types. Sixteen essays, edited by W.K. Wimsatt, which was based on three general theoretical papers: Lotz 1943, 1959, and 1960. 2 Cf. Chesterton's view of free verse: 'Free verse is like free love, a contradiction in terms'. a The universe of discourse is divided into two parts unequally. Although the notion of markedness is not universally applicable, there is no doubt that the paired terms 'marked' and 'unmarked' give significant information in many cases. With verse and prose, calling one 'marked' and the other 'unmarked' indicates an imbalance in the division of the universe of discourse. The concept of markedness has been recently reintroduced in modern American linguistics following Trubetzkoy's phonological theory. Difficulties arise, however, in using the terms 'marked' and

964

JOHN LÖTZ

The terminology and the concepts in the field of versification are heterogeneous and often confusing, especially as used in different literary traditions. The basic terms most often used in versification are: prosody, meter, verse, and poetry. 4 Prosody ( < Greek rcpoacoSia 'that which is added to song') originally referred to features of sound (i.e. pitch, duration, and stress) and Greek scholars of antiquity treated prosody in conjunction with sequential arts other than poetry, e.g. music and dancing. They considered it a discipline separate from grammar. In modern phonetic-linguistic usage prosodic features refer to pitch, stress, length, and, often, syllabicity. In the usage of metricians, prosody has an even wider coverage. It refers to versification in general, including rhyme, assonance, alliteration, and other characteristics of verse. Meter ( < Greek n^tpov 'measure') in antiquity was a general philosophical concept which later in the study of versification became restricted to the study of numerical regularity in verse, hence the term metrics. The corresponding Latin term numerus 'number' emphasizes this concept of numerical regularity. 'unmarked' as unclear cover terms without taking into account and systematizing the various types of asymmetrical relationships which can exist between a pair of phenomena. For example, in Russian or German voiceless-tense stops occur in all positions, whereas the voiced-lax stops do not occur finally. This imbalance results from a limited vs. a general privilege of occurrence. Another type occurs in frequency: e.g. in Hungarian /c/ and /£/ are very frequent sounds whereas the corresponding voiced Izl and Izl are rare. A further kind of imbalance is represented by the category of number in Hungarian where the plural category is characterized by reference to a set larger than the set containing one element, but it also contains the constraint that no numerical qualification, definite or indefinite, can precede the plural forms. The other number category, the 'singular', contains a set with a single element or a set where either numerical qualification is given or a generality or lack of countability exists, e.g. hajd 'ship', ot hajd 'five ships', vs. hajdk 'ships'. This opposition represents an imbalance in semantic referential coverage as to specification. The division of language texts into verse and prose represents an extreme imbalance where one of the terms, prose, is characterized only by the absence of any of the positive defining features required for verse. It should also be noted that there are some oppositions that would be very difficult to differentiate as to markedness, e.g. between labial and apical fricatives or, to use another field of reference, between man and woman. Exaggerating the concept of markedness to the extent of requiring each pair of phenomena in a universe of discourse to be characterized by it is a dangerously strong requirement just as demanding pervasive binarity in phonology or deriving all language components from an initial symbol'S' (sentence) is in linguistic analysis. Markedness has to do with opposition, and it is not always possible to determine what is 'marked' and 'unmarked' in an opposition. For instance, in the case of woman and man, there is, in some sense, the same kind of logical relationship as between the singular and plural categories in Hungarian. But whereas there is obviously an imbalance in singular and plural in Hungarian, this cannot be said in the opposition between woman and man, except if one would regard the number of chromosomes as the differentiating characteristic as to markedness. In like manner a linguistic pair of phenomena, such as i and «, can only be characterized as to markedness by the acceptance of arbitrary and ad hoc criteria, lacking any naturalness or compelling logical exigency. 4 Often another term, rhythm, occurs in discussing versification. Rhythm, however, is not a specific property of verse; it occurs in prose as well. Moreover, rhythm is a general phenomenon present in any periodically recurring phenomena, for instance in music, in mathematics, tidal waves, and the recurring seasons. Relevant here is the remark of St. Augustine: 'Omne metrum etiam rhythmus non omnis rhythmus metrum est'.

965

METRICS

In this article we use the term versification, not prosody, to refer to the general study of verse. Prosodie is used in conformance with modern linguistic terminology to refer to pitch, stress, and duration (syllabicity is not included). Verse is defined as a language text viewed in all its linguistic functions and characterized by meter, which is the numerical regulation of certain properties of the linguistic form alone. The literary use of verse is called poetry.

APPROACHES TO THE ANALYSIS OF METER

There are various ways in which verse can be approached. It could be approached from the point of view of art in general and especially from the point of view of music.5 The most elaborate example of this is in Paul Maas's treatment of classical Greek verse, where the constitutive 'metron', for instance, in the iambic trimeter is represented in the following way: anceps j-

biceps rr

breve j

biceps rr

The view that syllabic duration can be matched with the length of notes is based on a partial misconception of the temporal nature of the two phenomena, verse and music. Metric structure does have analogies in its sequential character with music; but in music the notes are completely regulated temporal phenomena, whereas in speech the sounds are evaluated in a looser way. In Greek, for instance, not only the bases consisting of a short vowel, but also those consisting of a short vowel and a single consonant (or a short vowel and a muta cum liquida), are classed as short bases; a long base includes an even larger variety of possibilities, ranging from a single long vowel or a short vowel followed by two consonants to a long vowel followed by four con6

The relationship of verse to the spatial (non-sequential arts) such as painting and drawing, is less immediate. The relationships which occur here are of several kinds. A poem might be represented in a certain shape as Simias's Egg in Antiquity, or Apollinaire's calligrammes or Christian Morgenstern's poem about the funnels: DIE TRICHTER Zwei Trichter wandeln durch die Nacht. Durch ihres Rumpfs verengten Schacht fliesst weisses Mondlicht still und heiter auf ihren Waldweg u. s. w. (Note the use of the letter w to represent the opening of the funnel.) A deeper relationship between graphic art and the semantic implications of poetry exists with artists who are both poets and painters such as William Blake, Henri Rousseau, and Paul KJee; see Roman Jakobson, "On the verbal art of William Blake and other poet-painters", Linguistic Inquiry 1.3-23 (January 1970). It is also possible to establish a mathematical relationship between poem and drawing, where the isomorphic relations can be mathematically described. Cf. Lotz 1965.

966

JOHN LÖTZ

sonants. Thus, the temporal variation is much larger than, let us say, between (J) and (J*). The following phenomena in Greek verse have to be distinguished: (a) the metrical scheme, which requires either a short or a long (-) base in some welldefined positions and which allows variation of short and long bases in other welldefined positions; (b) a concrete verse (a line), where the choices have been made, and every base is in fact either short or long; and (c) the actual performance, where both short and long bases vary in length. Dionysius of Halicarnassus reports that in the performance of the Homeric epics by the bards the long base was longer than the two short bases in obligatory positions. Clearly, verse and music share some common characteristic features, which are most obvious in songs, but the two are not identical, either in score or performance. Verse has its own independent organization, which is not the same as the organization of sound in music. Both temporal and dynamic relations in music are much stricter and organized in a different way from that of poetry. Music cannot be the basis for metric analysis.6 Another approach to verse is the 'objective approach', which takes its point of departure either from the physical recording of the event itself, as Edward Wheeler Scripture did in his treatment of English verse (1929; see too Verrier 1909-10), or from a phonemic transcription such as that employed by modern American structuralists.7 The idea that it is possible to reduce these phonetic data to significant units without making any special assumptions about them is an untenable oversimplification of phenomena employed by pure phoneticians and behaviorists. Only by pre-established rules is it possible to produce satisfactory results in metric analysis. The rules should not be confused with measurements of actual performance. The same Shakespearean line can be performed in a prosaic manner or in a manner that emphasizes the metric structure. The phonemic units cannot be deduced from the sound wave alone. There is a correlation between metric rules and metric performance, but no deductive relationships.8 Thus, within the field of metrics we have to distinguish the following: metric performance, which is characterized by the physical concepts of wave shape, frequency, intensity, time, and their psychological correlates, timbre (especially syllabicity), pitch, loudness, and duration; metric score, in which non-measurable relations are established on the basis of linguistic function; metric line-type, into which scores • 'The measurement of verse is determined by some recurrent linguistic feature, peg, obstacle, jutting stress, or whatever. If we read this recurrence so as to give it equal times, this is something we do to it. Maybe we actually do, and maybe this is a part of our aesthetic satisfaction; still it is not a part of the linguistic fact which the poet has to recognize and on which he has to rely in order to write verses' (Wimsatt and Beardsley 1959:590). ' See Chatman 1965 for references to American phonemic approaches to verse. " We can, of course, subject both metric numbers and metric performance to statistical analysis, and this kind of analysis is valuable especially for stylistic and historical purposes. But statistical analysis, even of metric numbers, cannot define and determine meter.

METRICS

967

are abstracted (in this case we may have variables in certain positions, besides the fixed values); and finally, metric system, in which all the various aspects of meter are subsumed. The most common approach to verse is to regard it as part of literature under the sub-heading poetry. This is the usual Western tradition but it is especially ethnocentric.9 In Sinology, the Chinese language is not exclusively associated with Chinese literature but with a number of other disciplines such as history, art, religion, and philosophy. The function of verse is by no means solely a literary-aesthetic one. Meter, the constitutive property of verse, does not serve a single function or purpose but can correspond with a variety of functions. It is a purely formal phenomenon which refers to the language signal alone without reference to the semantic content of that signal.10 Even in our culture, where the role of meter is predominantly literary-aesthetic, verse is used in advertising jingles. Old Germanic legal texts are couched in verse to prevent alteration. Poetic function is only one, though the most common, function of verse. Poetic function consists in the fact that versification with its paraphernalia of rhyme, assonance, alliteration, and the like is used for literary-aesthetic effect. The classification of verse according to its poetic function is not satisfactory because it does not establish a comprehensive definition of the term. It would, of course, be possible to subsume verse in all its varied uses under a single psychological label, such as attention-getting, but a labeling of the extension of a term does not mean an intensive definition of that term.11 An added difficulty in approaching verse from the literary point of view is the delimitation of a well-defined universe of discourse and the specification of verse within it. The correct approach to verse, in my opinion, is the linguistic one, which regards verse as a subset — to be sure a very small subset — of all language phenomena. Since all verse is a product of the use of language, it is entirely within the competence of linguistics. If there were a uniform function behind verse, it would be possible to approach it from the point of view of this specific function and to regard the language aspect as differentiating. Since there is no uniform function, it is impossible to define verse from any of the above points of view. The only feasible basis and workable method for the analysis of verse and meter is linguistics. (In the following we are concerned only with meter, the specific mark of verse.)

8 In American language and literature departments, the specifically linguistic component has often been neglected and linguistics has had to evolve on its own as a powerful, separate discipline. 10 Though meter is a formal notion, it may convey symbolic effect, as onomatopoeia does in speech. Contrast, e.g., the following two examples from Virgil, where meter expresses first equine speed, then slow force: 1) 'Quadrupedante putrem sonitu quatit ungula campum' and 2) 'Illi inter se magna vi bracchia tollunt'. Likewise the increase of bisyllabic valleys between the stressed peaks expresses the growing tension and fright of the child in Goethe's Erlkönig. 11 'Metrics is less an inherent feature of language than a conventional use to which language is put. By the same token, meter, in poetry, is a convention rather than an inherent feature' (Uitti 1969:224).

968

JOHN LÖTZ

THE LINGUISTIC BASE OF METRICS

In most languages there are texts in which the phonetic material within certain syntactic frames, such as sentence, phrase, and word, is numerically regulated.12 Such a text is called verse, and its distinctive characteristic is meter. Metrics is the study of meter. A non-metric text is called prose. Numerical regulation may refer to a variety of phenomena; therefore, verse and prose are distinguished not as two sharply differentiated classes, but rather as two types of texts. (This, however, should not obscure the fact that verse and prose are polar opposites, as mentioned above.) Meter can mean a strict determination of the occurrence of the syllabics in the poem and, in some cases, also of certain prosodic features. The regulation can be very strict. For instance the haiku in Japanese has exactly seventeen syllables (this is metrically simple, but culturally intricate): Shiki (1867-1902) Tsurigane ni tornante hikaru ho tara ka na

On the temple bell has settled, and is glittering, a firefly

In comparison to the simple haiku, the sonetti a corona, invented by the Academia degl'Intronati of Siena in the second half of the cinquecento, imposes exceedingly strict technical requirements on the poet. This highly stylized and complicated genre is an arrangement of fifteen sonnets in cycles. The rules require that the first fourteen sonnets shall form a coherent, connected, and closed series of variations on the subject, the transitions being effected by the repetition of the last line of each sonnet as the first line of the next, and that the whole chain of thought shall return to its starting point at the end of the fourteenth sonnet; and also that the fifteenth or master sonnet shall constitute a thematic summing-up of the preceding ones, so that both the formal expression and the arrangement are determined by the initial lines of the base sonnets. In the sonetti a corona the number of closely regulated syllables is about 2,240 as compared to seventeen in the haiku}3 But the regulation in verse can also be relaxed in various ways. In some types of Hungarian poetry a metrically relevant phrase can contain either six or five syllabics; and such a variation can be even wider, as in West Siberian folk poetry or in Hebrew Psalms. Or the numerical regulation can refer only to some parts of the text, as when prose and verse are mixed in the same literary work, e.g. in the Chinese fit genre. A text can be even more intricately interwoven, as in proto-Indo-European verse, where the beginning of the line was fairly f ree, but a substantial final portion was regulated. 12 It is very likely that not all cultures have verse, though I cannot think of any cultures, except dying cultures with only a few survivors, which do not have ritualized texts where some limitation of the language material is required. 13 The exact number of syllables in the sonetti a corona is between 2,100 (15 x 14 x 10) if all the lines are decasyllabic and 2,320 (15 x 14 x 11) if all the lines are hendecasyllabic.

METRICS

969

In certain types of classical rhetorical texts, only a small final portion (clausula) was determined. (The difference between these last two modes is a question of degree.) Furthermore, such a numerical regulation may involve a strict determination of syntactic units, with a wide variation in phonetic material. Free verse often tends to move in this direction. In Ghe, an African language, there are texts which are repetitions of units of five or three syntactic phrases bound together by rhyme. Other 'rhythmic' texts appear in the enumeration of sport scores on the radio (team-scoreteam-score). All these types of verse deviate from straight prose; the deviation can be put in terms of numerical regularity, or meter, and this regularity is the differentia specifica of verse. We could argue about where verse ends and prose begins, or whether or not we should introduce new intermediary types. In the following, however, we shall concentrate only on texts in which the entire phonetic material in its syllabic, and sometimes also in its prosodic, aspect is regulated. Such strictly formalized verse gives a contrasting background to other less regulated phenomena where they do occur. The use of the term 'versification' without the notion of a strictly regulated verse is a contradiction in terms. As a first example let us take the text of a Hungarian folksong: Hey peacock, hey peacock, Peacock of the Empress! If I were a peacock, I would rise in early morning, I would go to running water, I would drink the running water, I would rustle my wings, I would let my feathers fall. A pretty girl would gather them would fix them to the hat of her beloved sweetheart like a pendant plume.

Hej pava, hej pava, csaszarne pavaja Ha en pava volnek, jo reggel felkelnek, folyovizre mennek, folydvizet innam, szarnyam csattogtatnam, tollaim hullatnam. Szep leany felszedne, az o edesenek a kalapja melle bokretdba tiizne.

In this text the regularity is very simple: after every six syllables there is a phrase break, and a certain undefined number of such stretches, usually a small number, ends a sentence; the total utterance, the song, consists of an undetermined number of such intervals. As a second example let us take a scheme from another type of poetry which I will not reproduce in words, the classical T'ang Dynasty poetry: . . . .

X O O X

. . . .

O X X O

. (R) . (R) . . (R)

[. X O R

= = = =

any syllable syllable with distinctive tone, even or non-even; the other distinctive tone; rhyme]

970

JOHN LÖTZ

The total of twenty syllabics, repeated a number of times in the poem, includes four pentasyllabic phrase groups. In the first such stretch the second position must be filled by a syllable with either an even or a non-even tone, the fourth position must be filled by the opposite distinctive tone; the rest of the syllabics are undetermined. The second and third pentasyllabic stretches mirror this line, so that the second and fourth positions are the opposites of what they are in the first line; the other syllabics are undetermined. The fourth stretch repeats the scheme of the first. Here tonal features are regulated in addition to the number of syllabics. It is of interest to note that in Classical Chinese poetry both the syllabic line types and the strophic arrangement are highly restricted and the metric structure is strongly underscored by rhyming, alliteration, and parallelism. As a third example let us take the late Sapphic verse (as normalized by Horace) where the versification scheme is as follows: 1) word limit occurs after the eleventh, twenty-second, and thirty-third syllabic pulses; 2) sentence limit after the thirty-eighth syllabic pulse (though occasionally strophic enjambment does occur); 3) syllabics 1, 3, 5, 8, 10, 12, 14, 16, 19, 21, 23, 25, 27, 30, 32, 34, and 37 are either long or diphthongal, or, if short, are followed by at least two consonants; 4) syllabics 2, 6, 7, 13, 17, 18, 24, 28,29, 35, and 36 are short and not followed by more than one consonant; 5) the poem ends after a certain number of the stretches described in this versification scheme.14 It may be represented graphically: — w — x~ww_w_x _ w — X

x (anceps) = w or —

ww — w — X

— w — X — ww — w — X

— WW — X

14 The rules could be given in a mathematically more elegant form. E.g., Rule 4) would be: Positions (2, 6, 7) modulo (11 n), where n = 0 or 1 or 2, and 35, 36 are brevia (w). The hexameter requires a more complicated description in that first the stretch of bases constituting the hexameter line (of which there are a varying number ranging from 13 to 17 — in the rare versus spondiacus 12 bases might occur) has to be converted into a stretch of time units (morae) by assigning 2 morae to a long base and 1 to a short base as follows; the last syllable, an anceps, is counted as 2 morae:

-

=

2

W =

1

x = 2 In the following mapping P symbolizes bases, |i morae, indices ordinal numbers indicating position in the line, a the final base position: Pi Pz Pa • •• P = 24 morae This sequence of morae forms a numerically constant, positionally well-ordered set of exactly 24 morae. The rule can thus be formulated more elegantly in mathematical terms as: The 24 morae are represented on the base level by a long base if the position divided by 4 gives the remainder (symbolized by R) 1 or 2; i.e. R = 1 or 2 (modulo 4); other positions are optionally filled by either a longum or 2 brevia. Positions 19 and 20 are generally filled by 2 brevia (except in the rare versus spondiacus), positions 23 "and 24 by a single base (anceps). This approach can be generalized for an entire poem, e.g. the 8,745th morae in the Iliad must be part of a long base (8,745:4 = 2,186 remainder 1).

METRICS

971

Thus Horace. Ode, I.xxii: Integer vitae scelerisque purus non eget Mauris jaculis neque arcu nee venenatis gravida sagittis, Fusee, pharetra ...16 As a last example let us take English blank verse, or iambic pentameter, where there is: 1) word limit after every tenth or eleventh syllabic pulse: 2) sentence limit after a certain number of such decasyllabic or hendecasyllabic stretches (number undetermined); 3) syllabics in even numbered positions relatively heavier than syllabics in odd numbered positions as a thoroughgoing tendency (deviation at the beginning of the verse is frequent); 4) text ending after an indeterminate number of the three stretches just described. For instance, Shakespeare: 'I come to bury Caesar, not to praise him'. The above examples from several, diverse systems of versification demonstrate the wide variety of possible schemes of phonetic material in syntactic frames. They all represent verse types where the number of syllabics in the verse line are determined (isosyllabic). There are, however, other regular metric types where the syllables may vary, but some feature, other than syllabicity, regulates the verse line. For instance, in the Greek or Latin hexameter the number of syllabics varies from thirteen (twelve in the rarely occurring versus spondiacus) to seventeen, but the durational aspect is regulated in that each line contains twenty-four morae. A complete analysis of any system of versification would, of course, include a detailed linguistic description of all the means employed to differentiate language material from regular use of language. The study of the phenomena which differentiate verse from prose constitutes a special case of grammar which might be called poetic grammar. Poetic grammar is a selective and specific aspect of the total grammatical description of a language, just as phonology is a selective evaluation and classification of physical-physiological acoustic phenomena for linguistic purposes. It includes the linguistic analysis of all verse phenomena, such as phonology, morphology, syntax, and vocabulary. The constitutive aspect of poetic grammar, however, is the specific organization of the phonetic material in certain grammatical frames, namely meter. Other aspects of the poetic grammar may be utilized in a system of versification but do not by themselves create verse. We are concerned here only with those aspects of language that are relevant for metrical purposes. Yet even among those that are relevant we have to distinguish the basic constitutive factors of meter and the additive-variative ones. Linguistic phenomena are selected according to the principle of metric relevancy, in analogy with the principle of relevancy in phonological and grammatical analysis; for example, syllabic stress is not metrically relevant in Classical Greek, whereas it is metrically relevant in English. Hungarian offers a clearcut example of how phonetic 16

Early Sapphic verse was somewhat simpler; the third and fourth lines represented a unit not necessarily divided by a word boundary; the result was a three-line structure.

972

JOHN LÖTZ

material of the same language can be used in two different metric systems; base length is disregarded in the 'national' versification, yet it is the constitutive element in the 'quantitative' versification. Thus, metric relevancy implies a selection among the various phonetic phenomena of a language for metric purposes. The linguistic study of meter has two parts: a) study of the linguistic constituents and b) study of the metric superstructure. 16

STUDY OF THE LINGUISTIC CONSTITUENTS The linguistic components of meter are divided into two parts in accordance with the structure of natural language: o n the one hand, a small number of physiologically produceable, acoustically perceivable, and linguistically relevant sound features (phonological constituents) and, on the other hand, a large number of meaningful signs (syntactic, grammatical, or semiotic constituents). Therefore, our study of the linguistic constituents of meter is subdivided into two sections: 1) study of the phonological constituents of meter, and 2) study of the syntactic (also called grammatical or semiotic) constituents of meter.

Phonological

Constituents

All strictly regulated metric systems are founded on syllabification, which occurs on the physiological, physical, and psychological levels of speech transmission. It is, however, the occurrence of a syllabic pulse characterized by a syllabic peak rather than the syllable 17 as a sound stretch that is metrically relevant. Syllable implies both 16

All of the above discussion deals with speech sound. Terminology is often determined, however* by the graphic representation of verse: text, line, stichic and astichic verse, eye-rhyme, etc. In discussion of language, there is often confusion between speech and script, sound and letter. 17 The notion of syllable is one of the most obscure issues in linguistics. There is no doubt that there is an underlying pulse phonomenon in all speech more readily perceived than phonemes. The consistent matching of speech pulses with musical notes and the rhythmic chanting of children are evidence for the reality of syllabicity. (This pulse phenomenon, while universal, is in part socially regulated.) The difficulties stem from the fact that on the one hand a great number of correlated phenomena are subsumed under the term syllable and on the other hand, two disparate sets of phenomena, one of culmination and the other of delimitation, are thrown together into one unit. Some structural linguists, such as the Danes Brondal and Hjelmslev, and, later, American linguists, such as Hockett, Haugen, and Pike have used the term 'syllable' either for the place of the phonologically relevant accent, or as a minimal distributional model. In the first usage, languages like French, which have no distinctive accent, would have no syllables, and, of course, such a theory would be useless for metric purposes. The second interpretation, syllable as a basis for a distributional model, cannot be used as a constitutive feature for metrics either. I think 'syllable* needs an intensive phonetic definition, just as p is defined by bilabial closure, plosion, and open glottis. Such a theory was offered by Stetson, who ascribed 'syllable' to a single action of the intercostal muscles. His theory is original and of great importance. Yet it does not seem quite convincing to me. Menzerath thought that the rhythm of jaw movements would provide a cue for the syllabic pulse. I do not think that this will prove correct either, since one can articulate

METRICS

973

culmination and delimitation; this latter is not relevant metrically. Even in the socalled 'quantitative' meter, as in Classical Greek, the metrically relevant stretch is not identical with the syllable. It is possible, however, to use 'syllable' to mean a portion of sound determined by one syllabic crest, where the boundaries are not necessarily definite. A more or less arbitrary convention for syllable boundaries ('syllabification') comes from hyphenation and the breaking of words at line-ends in script or print. The simplest metric type is one which utilizes only the number of syllables within the syntagmatic frames. Here the metrically relevant notion is syllabicity, and the metric unit is the syllable. The regulation involves the number of syllables. This type is represented in many types of folk poetry: as, for instance, in the Hungarian folk song quoted above. A more complex metric system is one in which, besides syllabicity, a prosodic feature is relevant.18 Here we have to differentiate between various types of syllabic materials, which we may call bases. Thus, in the tonal meter of the T'ang Dynasty, the even tone was opposed to the other tonal classes for metric purposes. In durational, or 'quantitative', poetry—as in Classical Greek or in Arabic—the distance from one syllabic onset to the next one was organized into long and short bases. In dynamic types—like English or German—the syllabics are classified into heavy and light bases, with a great degree of freedom in the interpretation for metric purposes.

Syntactic

Constituents

The syntactic constituents of meter provide the frame within which the numerical regulation of phonological material can be stated. The meaningful semiotic frames which always function in verse are sentence and word, consisting of various kinds of morpheme groupings marked differently in each language. The smallest syntactic speech with constant mandibular angle, for instance when a pencil is stuck between the teeth. I have a different idea about the nature of syllabic pulse and its phonetic structuring and linguistic role. In my opinion, syllabification depends on the optimal filtering pulses in the buccal cavity. It represents a speech sound which is a filtered, sustained relative maximum in the speech chain. The pulse begins with the onset of the syllabic peak and continues until the onset of the next syllabic peak and does not depend on syllabic cuts as required in conventional usage. This approach solves the controversy about a case like split, which obviously has two energy and loudness peaks in s and i, but s has nofilteredformant structure, and the word is monosyllabic. This would also explain why aipdxa in Greek is regarded as a phyrrichius (y whereas fiaxpa, containing the same amount of sound material, is a trochee (" The term prosodic is replaced in modern American structural linguistic usage by the term suprasegmental. The term implies that speech has, in addition to segmental features, properties which extend beyond a single segment. Use of this term is misleading, in my opinion, because it does not accurately characterize the group of phenomena aimed at, while including other phenomena which one does not want to consider among suprasegmentals. For instance, voicing in English, which is not regarded as a suprasegmental feature, extends inherently beyond one speech sound in all cohesive obstruent clusters. On the other hand, tone, which is regarded as a suprasegmental feature, is not longer than the underlying timbre, e.g. Cantonese d (rising tone) vs. a (falling tone).

974

JOHN LÖTZ

unit, the morpheme, does not figure constitutively in meter, although it may be used to indicate verse structure, as in parallelism. The syntactic units, phrase or clause, cannot be utilized as such for metric purposes, because they may be discontinuous stretches; a metrically relevant unit obviously must be a cohesive stretch of the language sequence. The relationship of word and sentence can be expressed in a diagram indicating all the syntactic relations in a sentence (which for linguistic purposes would have to be specified and labelled). An example where the syntactic relations are fairly simple and lend themselves to a linear, sequential representation can be seen in a Middle High German line from Walter von der Vogelweide's farewell elegy:

Ó wê war sind verswunden alliu miniu jâr A syntactically more complex line, where the syntactic phrases are discontinuous and the relations are intersecting, is represented by the following line from Blake's Tyger :

Did He who made the lamb make thee? In addition to sentence and word, there is another syntactic sequence which often serves as a basic semiotic frame for meter. This unit is the colon ( < KO>X,OV 'member'), a notion which comes from Greek metricians but which is often unrecognized in modern metric theory. Although cola sometimes correspond with the concept of phrase or clause, they are not syntactic units in the same sense. The colon is a cohesive, sequential stretch of the verse line characterized by syntactic affinity or connectedness utilized for metric purposes. The lower limit of the colon is one word, but this is usually not the case; if it were, the notion of colon would be unnecessary. Besides being characterized by sequential cohesiveness and syntactic affinity, a colon is relative within the sentence-morpheme span. For instance, 'big house', 'very big', and 'very big house' are all cola, though of different degrees. In the Hungarian folk song quoted above, each line constitutes a colon. The verse line of Walter von der Vogelweide is divided into two cola for metric purposes: I) O we war sind verswunden and 2) alliu mitiiu jar. In the Blake example, on the other hand, the division of the line into cola is metrically irrelevant. In the study of metrics, the colon is a syntactic unit of material utilized for metric purposes and is related to the notion of caesura. In an overwhelming number of

METRICS

975

Greek hexameters there was a cut in the third metric unit, either following the obligatory longum (-) (masculine caesura); or, if the variable position in the third unit consisted of two short bases (yj), after the first breve (feminine caesura). In the French alexandrine there is a syntactic cut in the middle, dividing the line into two comparable syntactic frames; later a tripartite division was made by the Romantics. Colon is an especially dominating metric concept in folk poetry (see the first Hungarian example above). Metric usage of cola is also dependent on 'syllabic weight'. For instance, a twoword attribute noun phrase in Hungarian is evaluated one way if the constituents are monosyllabic and another way if they are polysyllabic: for example, szép no 'beautiful woman' consists of a single colon, whereas, gyonyoruséges boszorkány 'ravishing enchantress' consists of two. It should be noted that the concept of colon is not restricted to metrics alone. Though unrecognized as a functional linguistic unit, it is the basis for intonation and stress patterning in general language usage.19 Besides the constitutive linguistic components necessary for meter, there may be other phonological and grammatical features which underline and emphasize the metric structure. These, however, do not by themselves create meter, and they may also function independently in prose. Such features include assonance and alliteration, which can function as indicators of cohesion in the line; rhyme, which is usually correlated with inner response in the metric structure; 20 refrain, which often indicates a higher order of construction; and parallelism, which shows a correspondence in the grammatical structure of responsive lines. These phenomena can even be required constituents in a given metric system, e.g. alliteration in Finnish, rhyme in Classical Chinese, parallelism in Ob-Ugric folk poetry. METRIC SUPERSTRUCTURE

Metric units are those syntactic frames for which the numerical regulation of the phonological material can be stated. A metric utterance, the poem, can be constructed repetitively of shorter units, called lines, or it can be constructed more freely, astichic " The earlier preoccupation of American linguists with phrase-structure grammar was based on a desired identification of linearity and coherency of cola with syntactic structure in general. 20 The idea advanced by some American structuralists that 'pure' rhyme should be incorporated as a cornerstone into phonological theory for determining phonemic units is based on an ethnocentric Western view of rhyme. Rhyme is a culture-bound phenomenon. In Germanic languages the ideal rhyme is completely identical phonetic material beginning with the vowel of the last stressed syllable and continuing to the end of the line. (There are, of course, deviations for emotive effect; e.g. Poe's Ulalume contains the rhyme: sister, kissed her, vista.) In Arabic, however, only the final consonants are identical; in Spanish assonance only vowel identity is relevant; whereas, in Hungarian the identity of the final vowels is required, and only a similarity in the manner of articulation of consonants, not their total identity, is regarded as desirable (e.g. hat and kap but not hat and hal). It should also be noted that the use of rhyme is not restricted to verse alone. For instance, a popular genre in Arabic, the Makama, uses rhyme systematically in prose.

976

JOHN LÖTZ

verse. Lines may be organized into higher-order constructs such as strophe and cycle, and may have an internal organization into segments characterized by certain types of word boundaries called caesurae. 21 The relation among comparable elements in the metric structure is called response. Where there is only one line in a poem, such as a single hexameter, this response may refer to other examples of metric structure given in the culture. Thus the proverbial hexameter 'Natur(am) expellas furca, tamen usque recurret' echoes similar lines in classic Latin poetry. The components of the linguistic base and the metric superstructure may be summarized: Linguistic Base Metric Superstructure a. Astichic a. Phonological: Vowel Consonant Syllable ß. Stichic: b. Semiotic (grammatical): Segment (Morpheme) Line Word Strophe Colon Cycle Sentence Utterance = Poem

METRIC TYPOLOGY

In the metric typology that follows we have considered only those aspects of language that are relevant for metrical purposes. The aim is to include in a unified framework the diverse systems of versification known from various cultures. The presentation concentrates on the central core of phenomena and does not follow up the tangents which can lead off at every single aspect of this most deliberately formulated and experimentally varied use of language. We consider a variety of metric systems, each of which presents a complete regulation of a set of texts. Typologies can be set up in different ways, and they vary accordingly as to their use. The primary characteristic of the typology presented here is that it is deduced 21 Another unit which often figures in the literature on meter but which is disregarded in this presentation is the foot. Elaborate lists of terms are given to the various types of feet, especially in Classical Greek and Latin and Arabic poetics and in the various West European metric traditions which follow the Classical tradition, e.g. Choriamb -w-, ionicus a minori ^ — , besides the customary iamb, trochee, dactyl, and anapest, or in Arabic al-Wafir al-Mujtath It seems, however, that such stretches are not needed logically for the description of a metric scheme in addition to the semiotic frames on the one hand and syllabics, prosodically defined bases, and their paradigmatic replacement possibilities on the other as postulated above. The notion of a foot is transferred from the concept of measure (Takt) in music. The same holds for the notion of nfexpov in Greek tradition, a stretch which is usually longer than the foot. In principle, foot — or metron — could be used as a descriptive device, but it is derivable from the framework established here and is neither a basic nor a constitutive unit.

METRICS

977

from observation of diverse versification systems, but at the same time it takes into account the general nature of speech itself. It has, therefore, a claim for more theoretical validity than a typology which is only a taxonomic classification of observed data. Speech occurs in two parameters: timbres and prosodic features. Timbres, or sound qualities, are utilized for meter only as to syllabification, which is the obligatory base for all metric systems. Timbres occur in the physical world and can be measured. (They also occur in the psychological world, but there they cannot yet be satisfactorily described.) Prosodic features are three in number: duration, pitch (or tone), and stress. Prosodic features are also anchored in physical reality. Psychological perception of these physical parameters is very complex, but a certain correlation exists between duration and time, between pitch (tone) and frequency, and, to a much less clear degree, between stress and intensity. These four features (syllabification and the three prosodic features) give the total framework for a metric typology, and there are high culture examples for meters based on each of the four. It should also be pointed out that, as in all typologies, transitory metric systems and phenomena do occur. The typology includes Pure-syllabic meter: only the number of syllables within the syntactic frames — word, colon, sentence — is regulated (example: Hungarian folk poetry); and Syllabic-prosodic meter: in addition to the number of syllables certain types of prosodic features must occur. According to the kinds of prosodic features, three subclasses can be distinguished: 1) Durational meter (commonly called 'quantitative'): in some positions the length of the syllabics and, in some cases, the complexity of the following consonant cluster are regulated. Example: Classical Greek and Latin verse. 2) Dynamic meter: besides a certain number of syllabics, heavier and lighter syllabic pulses are required in certain positions. Example: English and German verse. Durational and dynamic meter show an affinity in that both utilize 'weightier' vs. less 'weighty' bases. The distinctive mark of these two types of meter may be subsumed under the term 'accent'. In Old Germanic verse the accented positions were filled by either a long or a stressed syllable, whereas there was considerable freedom in filling out the rest of the verse line. 3) Tonal meter: in certain positions well-defined tonal classes, representing classes of distinctive pitch phonemes of the language, are required. Example: even and non-even tone in Classical Chinese. Recently a theory was advanced that the regular T'ang dynasty verse has been essentially durational (quantitative), not tonal, in nature, the even tone corresponding to a longum (-), the deflected tone to a breve (y). This view seems implausible, however, for two reasons. The measurement of the duration of various tonal segments does not set the even tone apart, and, in addition, what is known about the development of Chinese dialects does not support such a theory. It should be noted that the above metric types can coexist in the same language: for example, in Hungarian both the pure-syllabic meter and the durational meter occur. Besides these 'pure' types, intermediate or mixed (as in most typologies) can be found. For example, French meter is basically 'pure' syllabic, but the final syllabic

978

JOHN LÖTZ

Metric Types

B. Complex Metric relevancy: Syllabicity and Prosodies Metric unit: Base Numerical regularity: Number and Position A. Simple Metric relevancy: Syllabicity Metric unit: Syllable Numerical regularity: Number (Example: Hungarian folk poetry)

a. Tonal: frequency-defined Base classes: 1. Even 2. Changing (Example: Chinese)

b. Dynamic: energy-defined Base classes: 1. Heavy 2. Light (Examples: English and German)

c. Durational: time-defined Base classes: 1. Long 2. Short (Examples: Classical Greek and Arabic)

Fig. 1. Chart of Metric Types

in a segment, not counting a final e-muet, must be heavy; in early Byzantine hexameter, when quantity was disappearing in the spoken language, the prefinal syllabic had to be oxytone; or, in some Hungarian verse, a choriamb is superposed on a pure syllabic structure. If ictus was used in some Classical Greek or Latin 'quantitative' verse — a hotly debated issue — it would provide another example of such a mixture. This typology (summarized in Figure 1) seems to be the most adequate for metric purposes. It would be possible, of course, to use other aspects of verse for erecting different typologies. For instance, we could set up a typology of syntactic frames utilized in verse. (Greek hexameter was delineated more clearly by words and word groups than Latin hexameter; free verse — when it is still verse — often has a firm syntactic composition; folk poetry often uses cola as units.) Or, we could make a typology of the numerical regulations imposed on the meter (strict, loose, or permitting variations; for examples, see the beginning of this article). These typologies, however, would probably turn out to be less informative and, some of them, trivial. In syllabic-prosodic meters the notion of the syllable alone will not suffice. A second concept has to be introduced: the organization of the syllabic material according to certain prosodic features, called long and short base for durational meter, heavy and light base for dynamic meter, and even and changing base for tonal meter. It is interesting to note that the phonological elements are grouped into only two base classes, never into more, although in principle much finer gradations would be possible. For example, in English more than two stress levels could be utilized, but, apart from the tendency in the so-called dipodic meters, there are never more than two classes utilized systematically; in Classical Greek the length of the syllabic and the following consonant clusters would have allowed a large number of classes, but only two types of bases were utilized for metric purposes, the short and the long; in Classical Chinese there were nine (or six) phonemic tones, but for metric purposes

METRICS

979

there was only one opposition, that between even and non-even tones. Dual opposition may prove to be a metric universal or at least a universal tendency. In syllabic-prosodic meters the numerical regulation refers not only to the syllabics as such but to the base classes as well; that is, in certain positions only one base class is allowed. Position must, therefore, be introduced as a second numerical principle, in addition to the number of syllabics, the sole characteristic of pure-syllabic meter. Positions which have to be filled out by a definite base class are called fixed. Other positions, those which allow variations, are called free. No syllabic-prosodic systems exist in which all positions are fixed. Free positions may be filled by either of the two base classes (anceps).22 Or there may be more complex substitutions. In a hexameter certain positions can be filled either by one long base or by two short bases. In Corinna's anaclasis the scheme allows four choices: ( - -), ( -), or ^ In this case, no position is fixed; groups of syllabics (feet) have internal compensations. PRESENTATION OF A METRIC SYSTEM

In the metric analysis above and in the establishment of the metric typology a linguistic base was assumed as the necessary point of departure. Within the linguistic frame, however, there exist various possibilities for presentation of the metric materials. Customarily a metric analysis is given discursively, but it can also be presented more rigorously by placing it in the framework of the basic mathematical requirements of symbolic logic or of generative-transformational grammar. Such an approach differs from the usual presentation of metrics in that it requires a stricter, more formalized analysis and format as are common in mathematical and physical sciences. In generative-transformational approaches to verse, metrics is conceived in purely linguistic grammatical terms as the dissolution of preceding symbols by means of rewrite rules into components (cf. definitions and primitives of an axiomatic system). The initial symbol might be regarded as 'verse' or 'meter' resulting ultimately in an abstract metric scheme. The abstract metric scheme is then converted through 'mapping' rules into concrete metric lines, with the possibility of distinguishing between types of lines and concrete lines of performance. The most significant application of this approach was applied to English verse materials by Halle and Keyser (1966). The axiomatic approach to metrics was first applied in a joint paper by Jakobson and Lotz on Mordvinian verse in 1941 where the metric analysis was made, following the assertion that: M

The term anceps, historically much older, can correspond to the notion of neutralization in modern linguistics. Neutralization means that distinctions normally employed are not made in certain well-defined positions. For instance, in English there are two sets of stops as to their force feature, fortis vs. lenis (e.g. pit vs. bit), but the distinction is neutralized after /s/, and only one set of stops occurs (e.g. spit but not sbit).

980

JOHN LÖTZ

The analysis of a metrical system requires the exact determination of all the constituents and their mutual relations underlying any meter of this system; the analysis must fully and unequivocally bring out which meters can and do exist and which cannot occur in the given system. Thus, the whole stock of the actual metrical forms really existing must be completely deducible from the rules established. In a recent revision of this paper, however, the present author rejected as a feasible basis for metric analysis the strong claim of this earlier attempt to set up an 'axiomatic' (i.e. rule-governed) device which generated all and only the existing metrical schemes of a system (Lotz 1972). The realization of a full corpus of metrical schemes and the exclusion of non-occurring and non-permitted metrical schemes is not possible due to the nature of verse itself, which, as the most idiosyncratic use of language, is subject to extreme individualistic manipulation. Therefore, instead of regarding the above account of Mordvinian verse as a closed explanatory system, it is better viewed as a descriptive statement which coherently defines the structural skeleton of a metric system but which is subject to modification with the introduction of additional data, or, possibly, with the expurgation of wrong data. 23 In addition, it should be pointed out that the above approaches do not increase the factual coverage of the metric material; rather, formalization of this kind forces a more careful treatment of the phenomena themselves and establishes an explicit basis for comparison among various metric systems and between metrics and other linguistic fields. OUTLOOK

The above presentation of metrics is based on linguistic assumptions within a linguistic frame and it is descriptive in nature. A comprehensive treatment of metrics would entail the placement of descriptive metrics in a broader framework by relating it to other linguistic subfields and investigating its relationship to phenomena of other social disciplines. For a complete linguistic treatment one would also have to take into account the historical development of various metric systems and their comparative analyses. Comparative work has recently yielded significant results not only in Indo-European studies but also in other well-known language families, such as Uralic. Historically, one would have to take into account changes which take place due to taste, i.e. subjective attitudes toward metric systems as expressions of poetic experience and concommitant changes in the metric form.24 2S

For a full discussion and application of the axiomatic method as applied to verse, see the revised article (Lotz 1972). 24 For example, such changes are the acceptance of trisegmental divisions instead of the traditional hemistichs in the Classical Greek hexameter by lyricists or in the French alexandrine by Romantic poets; or the replacement of the earlier Hungarian 12-syllabic alexandrine, which allowed the syllabic schemes 7 + 5, 6 + 6, and 5 + 7 for the hemistichs, by the more rigid 6 + 6 syllabic scheme for the two segments.

METRICS

981

Besides its linguistic aspects the study of metrics is also connected with other fields ; e.g. literary studies (meter employed primarily as a poetic device), anthropology and ethnology (the existence of folk and 'high' poetry, the cultural use of metric forms), psychology and aesthetics (the special effect of meter compared with the use of language in literary prose and colloquial style), sociology (the stratification of authors and their audience), and cultural history in general (the role of metrics and poetry in the entire context of the totality of culture and the changes it undergoes in time).

SELECTED BIBLIOGRAPHY

AUSTERLITZ, ROBERT. 1958. Ob-Ugric metrics. F F C 174.

CHATMAN, SEYMOUR. 1965. A theory of meter. The Hague, Mouton. DE GROOT, A. WILLEM. 1946. Algemene versleer. The Hague, Servire. GRAMMONT, MAURICE. 1957. Le vers français, ses moyens d'expression, son harmonie. 4th ed. Paris, Delagrave. HALLE, MORRIS. 1970. On meter and prosody. Progress in linguistics, ed. by Manfred Bierwisch and Karl Erich Heidolph, 64-80. The Hague, Mouton. , and S.J. KEYSER. 1966. Chaucer and the study of prosody. CE 28.187-219. HEUSLER, ANDREAS. 1925-29. Deutsche Versgeschichte. Vols. I—III. Grundriss der

deutschen Philologie, ed. by Hermann Paul. Berlin, Walter de Gruyter. JAKOBSON, ROMAN. 1952. Studies in comparative Slavic metrics. Oxford Slavonic Papers 3.21-66.

, and JOHN LÖTZ. 1941. Axiomatik eines Verssystem am mordwinischen Volkslied dargelegt. Thesis I, Hungarian Institute of the University of Stockholm. (Tr. 1951 : Axioms of a versification system : Exemplification by the Mordvinian folksong. Acta Instituti Hungarici Universitatis Holmiensis Series B.I. 5-13. Stockholm). Kenyon Review, Vol. 18, No. 3 (Summer 1956). LEHMANN, WINIFRED P. 1956. The development of the Germanic verse form. Austin, University of Texas Press and the Linguistic Society of America. LEVIN, SAMUEL R. 1962. Linguistic structures in poetry. The Hague, Mouton. LIU, JAMES J.Y. 1962. The art of Chinese poetry. Chicago, University of Chicago Press. Reprinted, in paperback, by Phoenix Books, 1966. LÖTZ, JOHN. 1943. Notes on structural analysis in metrics. Helicon 4.119-46. . 1959. Metrics and linguistics. Report on the tenth annual round table meeting on linguistics and language studies, ed. by Richard S. Harrell, 129—37. MSLL 12. . 1960. Metric typology. Style in language, ed. by Thomas A. Sebeok, 135-48. New York and London, M.I.T. Press and John Wiley. . 1965. The structure of the sonetti a corona of Attila Jozsef. Acta Universitatis Stockholmiensia, Studia Hungarica Stockholmiensia I. Stockholm.

982

JOHN LÖTZ

. 1972. Uralic meter. Versification: Major language types. Sixteen essays, ed. by W. K. Wimsatt, Jr. New York, MLA, New York University Press. MAAS, PAUL. 1929. Griechische Metrik. Gercke-Norden: Einleitung in die Altertumswissenschaft, Vol. I, Fasc. 7 (photocopy reproduction with additions). Leipzig and Berlin, Teubner. NORBERG, DAG. 1958. Introduction à l'étude de la versification latine médiévale. Acta Universitatis Stockholmiensia, Studia Latina Stockholmiensia. Stockholm, Almqvist & Wiksell. POE, EDGAR ALLAN. 1951. The philosophy of composition. The complete poems and stories of Edgar Allan Poe with selections from his critical writing, ed. by A. H. Quinn, 978-87. New York. First published April 1846, in Graham's Magazine. PREMINGER, ALEX, FRANK J . WARNKE, and O . B . BARDISON, JR., eds. 1965. Princeton encyclopaedia of poetry and poetics. Princeton, Princeton University Press. See esp. "Prosody" by Craig La Drière. SCRIPTURE, EDWARD WHEELER. 1929. Grundzüge der englischen Verswissenschaft. Marburg, N.G. Elwertsche Buchhandlung (G. Braun). SIEVERS, EDUARD. 1893. Altgermanische Metrik. Halle a . S . , Max Niemeyer. SNELL, BRUNO. 1957. Griechische Metrik. Gottingen, Vandenhoeck & Ruprecht. TOMASEVSKIJ, B.V. 1923. Russkoe stixoslozenie. Moscow. UITTI, KARL D. 1969. Linguistics and literary theory. The Princeton Studies of Humanistic Scholarship in America. Englewood Cliffs, N.J., Prentice-Hall. VERRIER, P. 1 9 0 9 - 1 0 . Essais sur les principes de la métrique anglaise. 3 vols. Paris. WILAMOWITZ-MILLENDORFF, ULRICH VON. 1921. Griechische Verskunst. Berlin, Weidmannsche Buchhandlung. WIMSATT, W.K., Jr., ed. 1972. Versification : Major language types. Sixteen essays. New York, MLA, New York University Press. , and MONROE C . BEARDSLEY. 1959. The concept of meter: An exercise in abstraction. PMLA 74.585-98. WRIGHT, WILLIAM. 1955. Prosody. A grammar of the Arabic language, 2.350-90. 3rd ed. Cambridge, Cambridge University Press. ËIRMUNSKIJ, V. 1966. Introduction to metrics. (English edition, with introduction by E. Stankiewicz and W.N. Vickery.) The Hague, Mouton.

PART FOUR

SPECIAL LANGUAGES

NEW FORMAL DEVICES FOR LINGUISTICS

MAURICE GROSS

Algebraic linguistics is by now a well-developed field that constitutes an indispensable background for theoretical linguistics. The important study by Chomsky (1959-62) inaugurated this field of study, and since then it has grown considerably in many directions and under various impulses. The object of algebraic linguistics is the classification of families of sequences of letters called formal languages. If A is an alphabet (i.e. a set of letters), then we call A* the set of all sequences of letters that can be formed with A, and a formal language L a subset of some A*. There are several ways of looking at sequences of letters. Strings of letters are defined in terms of products or concatenations of strings (possibly of one letter); representing a string is very much like writing a word in ordinary language. Sequences of letters can also be considered as written on tapes divided into squares. In each square, we can have either a letter or a blank (noted # ) . Example: String: aaabb, also written a3b2. The exponents (3 and 2) indicate the numbers of successive letters. Tape:

#

a

a

a

b

#

a #

a

a #

b

b #

#

a

a

b

b #

a #

#

b #

#

#

These three tapes are all different, but they correspond to the single string a3b2. Strings are described by formal systems or grammars composed of rules of the type x-*y, where x and y are strings. When a string has the form uxv, the rule is applicable to it, and the result of the application is uyv. Tapes are described by automata which are devices composed of several mechanisms: a central unit that can be assigned different states, and one or several 'reading' and/or 'writing' head(s), under

986

MAURICE GROSS

each of which a tape is moved square by square, and modified in content according to the states of the central unit. Linguists are concerned with the formal investigation of the descriptive tools that they use. They attempt to formulate abstract models for some of the phenomena they observe, then to derive from precise formal hypotheses properties that are significant for natural languages. Computer specialists became interested in algebraic linguistics for several reasons. One of their problems is the strict specification, by means of grammatical rules, of the programming languages used to state computing algorithms. These rules intervene in the translation process into machine language, a process which involves syntactic analysis very much as in the mechanical analysis of natural languages. There are other concerns in Communication Sciences. Computation theory is an attempt to formalize the notion of (mechanical) computation, automata theory deals with certain of these aspects. In linguistics, automata have been considered as models for various syntactic properties (push-down automata for self-embedded constructions, for example, Chomsky 1963 and Schiitzenberger 1969). In computation theory, automata can characterize various operations (certain additions, multiplications, etc.) in terms of 'hardware' (movement of the tape(s), type of the head(s), etc.). Computer technology deals with automata in a different way. Certain circuits for the transmission of data can be modeled by automata, the corresponding methods constituting an important part of Switching Theory (Harrison 1965). Pure mathematicians have developed new approaches to the abstract study of formal languages. The notion of string has been shown to be a natural extension of the notion of number. Thus, finite-state languages (also called regular, or recognizable, or rational languages) correspond to the rational numbers, while context-free languages correspond to algebraic numbers (Schiitzenberger 1969). A number of devices relevant to linguistics have been reviewed by Chomsky (1963) and by Kiefer (1968). Since then, new formal devices, and new properties of languages have been found in relation to the various interests that we mentioned. We intend to discuss informally some of them, stressing particularly their relevance to language description, and indicating where empirical findings might support their integration into a general theory of language.

1. DESCRIPTIVE POWER OF GRAMMARS

Grammars, and automata, have been classified according to their ability to characterize more or less finely sets of strings, according to their descriptive power. We will, as an example, consider different abstract languages defined on the alphabet of three letters^ = {a, b, c}. A* is the set of all strings made of a's, Vs, and c's, and we will deal with the three denumerably infinite languages:

NEW FORMAL DEVICES FOR LINGUISTICS

987

L j = {am b" c":m,n,p> 0} (the numbers m, n, p, are independent of each other); L2 = {am b" 0} (we have the same numbers of a's and c's in each word, but n, the number of b's, does not depend on m); L3 = {am bm c™ : m > 0} (each string has the same number m of a's, b's, and C'J). We have the relation: L3czL2a ¿ i (with strict inclusions). L3 is 'smaller' than L2, and L2 is 'smaller' than Lx. The smaller languages are, the more complicated (i.e. powerful) are the grammars needed for their characterization. Thus, L, can be characterized by a finite-state grammar ( = f.-s. grammar) which is one of the weakest kinds of grammars. L2 and L3 cannot be described by any grammar of this type; L2 needs a context-free grammar ( = c.-f. grammar), while L3 is beyond the power of context-free description but can be described by a context-sensitive (— c.-s.) grammar. The proofs given in Chomsky 1959-62 and 1963 that correspond to this situation show in an interesting way these differences of power. Let us suppose that we want to describe L2 by means of a finite-state grammar of the most general kind. 1 The shape of L2 imposes a certain form to any (finite-state) grammar that describes it. In particular, it can be proved that the sequences of applications of rules, namely the derivations, have to yield patterns of symbols (terminal and non-terminal) that will match the strings of L2. It can also be shown that when L2 is obtained from these patterns, at the same time, because of the constraints imposed on the grammar, there are other strings, not in L2, which are described as part of the language of the grammar. In other terms, when one tries to describe all strings of L2 by means of a finite-state grammar, one has to describe a 'larger' language. For empirical reasons, we may want to include in the description of Lz certain finite-state features. Then, in order to compensate for the weakness of the finite-state process, we would have to add to the grammar some extra apparatus that would prevent the language from growing beyond the required limits. We will now review various mechanisms that have been defined in order to obtain this effect.

2. GRAMMARS WITH CONTROLLED DERIVATIONS

Linear context-free grammars have rules of the form: A-*xBy or A - y x 1

The circumstances that we describe would be exactly the same if we tried to describe L3 by means of a context-free grammar.

988

MAURICE GROSS

where A and B (e VN ) are non-terminal symbols, and x, y, (e VT ) are terminal strings. Example: VN = {5, T}, VT = { / , g, h) S-yfSh (G): S-*fTh T-+g T T^g

The language L(G) described by this grammar is L2 = {f m g" h m : m, n> 0} Derivations are of the form: S-yfS

h

•>/m Th m *f m

gTh m

(D): *f mg n~ 1 Th m m m *f g" h

They will be called linear derivations, since only one non-terminal symbol occurs in each step (except for the last step which is terminal). Trees associated with such derivations have the shape:

/

/

...

m times/

/

g

8

n times g

m times h

Fig. l

We will consider the sequence of non-terminals that occurs in a linear derivation like (D) as a word on VN , and we will call it the backbone of the derivation, or of the corresponding tree (Gross 1966). Given a linear context-free grammar, the set of its backbones constitutes a finitestate language; in the case of (G) above, the set of the backbones is {S™ T n : m, n > 0}

NEW FORMAL DEVICES FOR LINGUISTICS

989

and L2 is associated with it. Conversely, when linear context-free rules apply to a finite-state set of backbones, the associated language is linear context-free. Instead of leaving the set of derivations determined by all possible patterns of applications of rules, we could restrict it, by giving first a set of backbones that will be submitted to certain linear context-free rules. Thus, the backbones can constrain the possible sequences of applicable rules, restricting at the same time the language that corresponds to the rules. We will restrict the sets of backbones to linear context-free languages, which are obtainable in a simple way. The languages that will be associated with these sets of backbones will be 'smaller' than the context-free ones. The following example provides a language which is not context-free. Let us consider L2 = { f m gnfm : m, n > 0} as a set of backbones, it is obtainable by the grammar (G) above. The following linear context-free rules will generate L2 : f->afc f-*ag g~*g

c

g-*h h-*bh h-yb

where the non-terminal alphabet is {/, g, h}, and the terminal one is {a, b, c}. The associated language is L3 = { o " i V : « > 0 }

a3 b3 c 3 E L 3 . Fig. 2 3.

PROGRAMMEDjGRAMMARS

Rosenkranz (1969) has defined an interesting device that can be superimposed on any

990

MAURICE GROSS

given grammar, and that has a restrictive effect with respect to the type of language that the grammar can describe by itself. Given a grammar, each of his rules receives a name (r), and to each rule one associates two sets of such names: S{V) and F(W) that will indicate which rule is to apply after rule (r). A programmed grammar is thus a finite set of expressions (r) f-+g

S(V), F{W)

where f-*g is a rewriting rule defined as usual by means of two alphabets (VN, VT). Derivations always start with a rule S-*f, where S is the axiom. In the course of a derivation, if a rule (r) applies (i.e./occurs in a step and is replaced by g), the next rule to apply is one of the rules of S( V). If the rule (r) fails to apply, then the next rule to apply is one of the rules of F(JV). Example: VN = {5, A, B}, VT = {a, b} S is the axiom, and e is the null word. The rules are: 1. S-*A B S(2, 3,4) 2. A-yaA S ( 5) 3. A-*b A S{ 6) 4. A-*e S( 7) 5. B-yaB S{2, 3,4) 6. B-*b B S(2, 3,4) 7. B-*e Without the 'programming' apparatus, the rules would describe a finite state language (VT). Here, the device characterizes the language {jcx : x e V^} which is not a contextfree language. S

Fig. 3

Each word of the language consists of an arbitrary string x (on a and b) followed by its copy x. Both sequences x are described by finite-state structures (i.e. right-

NEW FORMAL DEVICES FOR LINGUISTICS

991

branching trees), and the programming device operates an identity check between the two members of the product S — A B. 4.

INDEXED GRAMMARS

Aho (1968) has extended the notion of context-free grammar by allowing non-terminal symbols to be indexed by sets of context-free rules, and by displaying these indices in the same way as the other symbols. An indexed grammar defined on the two alphabets VN (non-terminal), and VT (terminal), has rules of the form: A-+f

AeVN

The right member / is defined on VN u VT in a complex manner, since each nonterminal symbol may carry some indices which are themselves defined on VN u VT. An index is a finite set of context-free rules: B-*g

beVN

ge(VN

uVT)*

(i.e., here, the non-terminal symbols of g do not have any index). There are thus two types of rules, the first ones apply freely and generally involved indices, the others do not introduce any index, and their application is directed by the indices. These two types of rules are used differently. In the first case, a symbol A is replaced by the corresponding right member/. Let r be the index or sequence of indices attached to A, and let the step of derivation to which the rule A-*f applies be vArw, then r will distribute over the (indexed) nonterminal symbols of / . If the rule is: A

XqX^X-^X^X^ ... XfoXfc

where the Xj's are (indexed) non-terminal symbols (i.e. of the form Cs ... t, Ce VN and s ... t is a sequence of indices), and where the X/'s are terminal strings (e VT), then the resulting step will be vx0X1rx1X2xr2

... Xkrxkw

and r does not distribute over the terminal jt,'s. In the second case of rule application, the step of derivation is vArsw (A e VN), r is an index, and s is a string of indices). The rule: A y XQX^X-^JC2X2 • • • XfcXfc belongs to the index r (thus the X,'s do not carry any index). When it applies, the index r is absorbed (i.e. it disappears), and the index s distributes over the Xt's as in the first case, the resulting step being vx0Xisx1X2sx2

... Xksxkw

992

MAURICE GROSS

Example: VN = {S, T}, VT = {a, b} The rules that introduce the indices r and s are: S-*aSr S-*bSs S-> T The indices are: r = {T-> Ta, T-+a} s = {T^Tb, T-+b} When we derive a word, we have: S -> aSr -* aaSrr

aabSsrr

aabaSrsrr -*• aabaTrsrr

by means of the rules of the first type. Then the index rules apply: -> aabaTsrra -* aabaTrrba -> aabaTraba -> aabaaaba. The tree associated with the sequence aabaaaba is

The whole language is L = {xx : x 6 {a, A}*}. It is not context-free, although it has been obtained through a context-free-like process. This possibility is essentially due to the distributivity of the indices. Very much as in §2 the words of such languages are described with a double structure. One structure, given by the first type of rule,

NEW FORMAL DEVICES FOR LINGUISTICS

993

describes context-free constraints between symbols and indices both considered as being the same sort of element. The second kind of rule then makes a distinction between symbols and indices and only describes constraints between symbols.

5.

EMPIRICAL DATA A N D NEW DEVICES

We now indicate various linguistic phenomena which quite naturally suggest that a mechanism like the ones just presented could be used adequately. There is an area of structures discovered by Z. Harris (1964) and called by him parallel structures, which has proved hard to analyze by means of classical devices,2 whereas they appear to play a very central role in language. Conjunctions are examples of parallel structures. It has, for example, often been noticed that sentences like: John knows Paul, and the earth is an ellipsoid. are not grammatically acceptable, the reason being that the two sentences on each side of the conjunction and are totally unrelated in meaning. Other coordinating conjunctions exhibit related features. The sentence: John knows Paul, but John knows Jack. is odd, if acceptable at all, because the sentences on each side of but are too similar. The sentence: John knows Paul, but John does not know Jack. does not have the preceding ungrammaticality, the negation makes the second sentence distant enough from the first. The conjunction and ... too imposes similar restrictions on its two members; we have the following type of paradigm: 1. 2. 3. 4.

John knows Paul, and John knows Jack too. *John knows Paul, and John does not know Jack too. *John knows Paul, and John believes Jack too. *John knows Paul, and Paul believes Jack too.

Example (1) shows that and ... too and but present different patterns of constraints. Examples (2), (3), and (4) show that both members should not be too 'different'. Further study, also with pairs of sentences sharing identical words, suggests that there should only be one difference3 between the two sides of and, as in (1) (i.e. Paul vs. Jack), or in: 2 The term classical device refers to the various types of generative grammars that have been described by Chomsky (1957, 1965, 1972). 3 The sentence: John knows Paul and Paul knows John too. is quite correct, although there are two differences: John vs. Paul, and Paul vs. John. Perhaps the reversal of order between John and Paul should be counted as the only difference.

994

MAURICE GROSS

John knows Paul, and Jack knows Paul too. and: John knows Paul, and John appreciates Paul too. This remark would still be true, if we added to the main verbs (i.e. knows, believes, etc.) other complements ('place', 'time', etc.). Another important case of parallel structure is given by relative clause adjunction. This rule has the effect: A man entered the room WH the man talked to me. -»• The man who talked to me entered the room. where WH is a conjunction,4 and the sentences on each side of WH must have one noun in common. The similarities and differences should not be strictly defined in terms of morphemes; in fact meanings of various types are involved. Sentence (3) above is not acceptable because the meanings of knows and believes are too different, but the sentence: John knows Paul, and John met Jack too. which seems to be like (3) in terms of superficial parallelism, is much more acceptable, presumably because of the common meaning and presuppositions implied by the two parallel verbs. Harris (1968) considers parallel structures as a fundamental device for the analysis of meaning. Thus, sentences like: 5. He wrote poetry because it was Tuesday. will be interpreted with various presuppositions according to the similarities and differences that are present on each side of the subordinating conjunction. Harris analyzes the presuppositions further by reconstructing intermediary sentences that chain the two sentences of (5) and that present constraints of parallelism. One presupposition is thus expressed by: He wrote poetry because it was Tuesday, and on Wednesday, he has the poetry class. The newly introduced sentence has only one word in common with the initial sentence (5). In order to introduce more parallelism between the apparent and underlying elementary sentences of (5), and also at the same time to make explicit meaning, 'classificatory' sentences like: Wednesday is the day after Tuesday. * Actually it could be the case that WH and and are two forms of one abstract conjunction, and that relative clause adjunction is only a special case of conjunction with reduction of the common part, as in The man talked to me and entered the room.

995

NEW FORMAL DEVICES FOR LINGUISTICS

are also reconstructed; another such sentence could be: A poetry class is a class where one learns poetry. We will then have sequences of elementary sentences like: Su

'S3

where St and S2 on the one hand, and S2 and S3 on the other, are parallel. Moreover, S2, for example, is the reconstructed sentence, and it is such that its words (excluding purely grammatical elements) are all found either in Sx or in S3. These conditions allow the elements of S2 to be zeroed, yielding Su S3. In this series of examples, the situation is the following. Each sentence, whether part of a conjoined construction or not, has its own structure. On the other hand, when two sentences enter a conjoined construction, a new structure independent of the preceding ones is imposed on the pair of sentences. Such a double structure cannot be described by means of a phrase structure grammar, and is hard to account for in terms of generative transformational grammars. From the empirical point of view, the formal devices that we described above are much more adequate, since they incorporate by definition a notion of double structure. There are sentences with crossing constraints that contain the conjunctions respectively or according to, and that cannot be described at all by context-free grammars.

(Resp): John, Paul,...,

Alfred are respectively

tall,

fat,...,

stupid.

Fig. 5

The way sentences like (Resp) should be handled by a transformational grammar is by no means obvious. In particular, it is hard to choose and justify a deep structure for them. Except for a morphism,® sentences like (Resp) constitute the language L = {xx \xe{a, b}*}, which can be described by programmed and indexed grammars, and also by grammars with controlled derivations. The following system will characterize L. Backbone rules: VN = {S}, VT = {Juh, gu gi}, e is the null word, S^ASf2, 6

S->glSg2,

S-+e.

This is clear for a language like French which requires gender and number agreement. If we suppose that gender is marked by the morpheme a, and number by the morpheme b, and if we ignore the other morphemes, the sentences take the form xx.

996

MAURICE GROSS

Controlled rules : VN = { f i , f 2 , gi, g2}, A afi, h -> agi, A gi bgu gi bfu gi f2 ~*f2a, f2 g2a, g2 f2~*a, g2~*b.

VT = {a, b} af2, f-> ag2, bf2, gi bg2, bg2, g2 ^ bf2,

In the case of indexed grammars (Fig. 4), as well as with the preceding grammar, the structure of the sentences (Resp) is imposed by the model (Fig. 5).

The vertical edge corresponds to a backbone or to the string of symbols and indices described by the non-directed rules of an indexed grammar. In either case, this edge would be considered as a deep structure where the elementary sentences are embedded one into the other. Fig. 6

The description of parallel structures with differences occurring on each side of a conjunction would be made in a very similar way; a central edge would indicate the identities and differences in a context-free mode, and rules directed by this edge would display the components of the sentences.

NEW FORMAL DEVICES FOR LINGUISTICS

997

There are other phenomena described in the framework of generative transformational grammar for which programmed grammars might provide a very natural formalization. The notions 'obligatory rule' and 'optional rule' are constantly being used, although their significance is far from being clear. There is a very general way of making rules apply obligatorily. While they are usually defined on two alphabets6 (terminal and non-terminal), the language of the grammar is defined as the set of all TERMINAL strings that can be obtained from the axiom(s). This definition thus implies that all non-terminal symbols occurring in the steps of the derivations have to be eliminated. In other words, certain rules that turn non-terminal symbols into terminal ones apply obligatorily. This global method for obligatorily applying certain rules has been made more specific in a number of uses. Thus, Katz and Postal (1964) proposed that all transformations be applied obligatorily to a basic structure, the result being an actual sentence. Their method consists in placing markers (i.e. non-terminal symbols) in the basic structures; transformations involving these markers apply and eliminate them when the intended result is reached. For example, the symbol WH (for 'question') is adjoined to a declarative sentence structure, and the 'question' transformations will eliminate this WH: basic structure: WH John will say something tense-attraction rule: WH NP will will NP. This rule eliminates WH and applies obligatorily; otherwise the basic structure could not become terminal. It also applies to any other auxiliary verb {can, could, have, had, is, may, etc.). The result is the question: Will John say something? Such an obligatory effect could easily be obtained by means of a 'programming' device in the sense of 3) above. Instead of inserting and erasing WH as we did, one could have attached to the last rules that describe the declarative basic structure a set of rules S(F) that would include the rule of tense attraction. This rule would then apply obligatorily too, and would be formulated in a more natural way, namely as: NP will -> will NP The programming device is such that when a rule applies, the next rule is (at least in part) determined. This property of the rule system turns out to have an empirical basis. Harris (1968) has verified the following markovian character of the rules of a transformational grammar: if one considers the sequences of rules that apply between sentence structures as a language, this language is finite. " The formal grammars described in Joshi (1969) do not quite have this character. They manipulate only terminal strings that can be adjoined and/or distributed over other terminal strings. However, there is a limited number of markers that are necessary to indicate structure, and that have a function close to non-terminal vocabulary.

998

MAURICE GROSS

From a formal point of view, generative transformational grammars have a rather important defect: they are much too powerful. They can characterize languages that have nothing to do with human language (i.e. recursively enumerable languages). The devices we have presented are much more restricted, and to a certain extent they are more adequate with regards to the data; we think that they deserve further study. However, most of the empirical problems we raised in order to justify their use (parallel structures, obligatory rules, etc.), have barely been studied, and there is so far no serious basis for choosing one or the other type of grammar we have described.

BIBLIOGRAPHY

AHO, A.V. 1968. Indexed grammars — an extension of context-free grammars. JACM 15/4 (October). 1 9 5 7 . Syntactic structures. The Hague. . 1959-62. On certain formal properties of grammars. I & C 1959-62.137-67. . 1963. Formal properties of grammars. Handbook of mathematical psychology, ed. by D. Luce, E. Bush, and E. Galanter. John Wiley & Sons, Inc. . 1965. Aspects of the theory of syntax. Cambridge, Mass., M.I.T. Press. . 1972. Deep structure, surface structure, and semantic interpretation. Semantics, ed. by D. D. Steinberg and L. A. Jacobovits. Cambridge, University Press. GROSS, M . 1 9 6 6 . Sur certains procédés de définition de langages formels. Automata theory, ed. by E. R. Caïanello. New York, Academic Press. HARRIS, Z . 1 9 6 4 . Elementary transformations. Transformation and Discourse Analysis Project 54. Univ. of Pennsylvania. Philadelphia. . 1968. Mathematical structures of language. John Wiley & Sons, Inc. HARRISON, M. A. 1965. Introduction to switching and automata theory. New York, McGraw Hill. JOSHI, A . K . 1 9 6 9 . String adjunct grammars and transformational grammars. Transformation and Discourse Analysis Project 75/2. Univ. of Pennsylvania. Philadelphia. K A T Z , J. J., and P . M. POSTAL. 1964. An integrated theory of linguistic descriptions. Cambridge, Mass., M.I.T. Press. KIEFER, F . 1 9 6 8 . Mathematical linguistics in Eastern Europe. Amsterdam, Elsevier. ROSENKRANZ, D . J . 1 9 6 5 . Programmed grammars and classes of formal languages. JACM 16/1 (January). SCHUTZENBERGER, M.P. 1962. Certain families of automata and their decision problems. Proc. Symp. Math. Theory Automata, Vol. XII. Brooklyn Polytechnic Institute. . 1969. On context-free languages and push-down automata. I & C 1969.6-3. CHOMSKY, N .

ARTIFICIAL LANGUAGES: INTERNATIONAL (AUXILIARY) MARIO A. PEI HISTORICAL BACKGROUND

The episode of the Tower of Babel, in the First Book of Genesis, is probably the earliest recorded indication of man's recognition of the desirability of a universal tongue of common intercourse, viewed as something once possessed, then lost. The ancient Persians also conceived the history of the world as a series of evolutions in preparation for the reign of Ahura-Mazda, when men will live happily, and there will be one language, one law, and one government for all. The variety of forms of speech in the ancient world is fully attested, even if the forms themselves are not. The use of Akkadian as a common tongue in the Mesopotamian region, of Egyptian in the valley of the Nile, preceded the spreading of Greek throughout the Mediterranean basin, wherever the Greeks planted their farflung colonies. For a time Greek coexisted with the Punic of Carthage, the other great colonial power of antiquity. Then the rising star of Latin brought about a new symbiosis that overspread the ancient world, dominated by a Graeco-Roman culture which attracted and gradually absorbed speakers of other languages. The disappearance of the latter was a slow process, probably not completed until the Roman Empire was close to its downfall. There are indications that Etruscan and Oscan were still occasionally heard on Italian soil at the time of Christ, that Gaulish, Iberian, even the Punic of Carthage continued in use until at least the fourth century of the Christian era. But the Roman Empire was essentially a bilingual world, with Latin the language of the people in its western areas, Greek in the east, and a free interchange between the two by reason of the fact that the former was the universal tongue of military and civil administration, the latter the language of refinement and culture. The fall of the Roman Empire of the West in the fifth century brought about the fractioning of Latin into its Romance descendants, along with the entrance into the western world of new popular tongues, mainly of Germanic stock. Greek continued its unbroken evolution in the Roman Empire of the East for another thousand years, changing but slightly during its Byzantine period, then becoming a restricted vernacular after Asia Minor, the Balkans, and Constantinople itself fell to the Turkish invaders. Latin, though replaced in the mouths of the masses by its own offspring, con-

1000

MARIO A. PEI

tinued in use throughout most of Europe as the universal tongue of the western church, which meant in effect that it survived as a language in general use among churchmen, scholars, and scribes. When universities and centers of learning were again established, it was ecclesiastical or scholarly Latin in which their proceedings were carried on. Universality of language was limited to a minute fraction of the total population, but it still existed. For people of lesser learning there were curious creations of the pidgin variety that came into use, the Franco-Venetian by means of which French jongleurs were able to make the meaning of their chansons de geste accessible to Italian audiences, the lingua franca that united elements of all the Mediterranean tongues, including a newcomer, Arabic, and that was used in almost exclusively spoken form for purposes of commercial interchange, anticipating by centuries the Bazaar Malay of what is now Indonesia, the Hindustani of India, the pidgin English of the Chinese coast and the islands of Melanesia. But all these were and are symbols of limited, not general universality, extending over restricted areas. The Graeco-Roman world, and the Christian world that followed it, loved to think of themselves as universal, which, of course, they were not. Even when the Roman Empire had achieved its widest extent there were entire continents untouched by its influence — a far eastern world, dominated by Chinese culture; a great civilization that flourished in the Indian subcontinent; all of Africa south of the Sahara; and an as yet undiscovered and unsuspected Western Hemisphere. Yet the ideal of a universal Roman Empire, a universal Christian Church, a universal community of man under the fatherhood of God, endured throughout the Middle Ages, carrying as its corollary a belief in the universality of grammar based on Latin models, and expressed by Roger Bacon in his pronouncement that 'grammar is one and the same for all languages, though it may vary in particulars'. The breakdown of the universal idea came in part through the rise of modern national states with their centralizing tendencies which called for clear-cut separation from other national states; in part through the gradual rise of a new and affluent middle class which, especially after the invention of the printing press, insisted upon participating in the expanding life of the intellect, but on its own terms and in its own vernaculars, not in the difficult Latin of the clerics and scholars; in part through the great voyages of exploration and discovery, which revealed the existence of races and languages of a diversity hitherto undreamed-of. Latin fell into desuetude, though it remained the prerogative of the church and of the more philosophically inclined scholars. But now that very expansion of travel which had revealed the multiplicity of human speech aroused an unspoken desire for means of communication among the peoples of the earth. THE DAWN OF CONSCIOUSNESS

It was in this climate that Descartes, in 1629, advanced a proposal that was as

ARTIFICIAL LANGUAGES

1001

radical as it was novel. Beginning with an attack on the difficulties of national grammars, which prevent people from seizing the meaning of a passage by referring to the dictionary alone, Descartes went on to advocate the creation of a tongue with a grammatical structure so simple that it could be learned by anyone, being based on regularity and logic; along with a word-coining system of mathematical precision, with progression from the known to the unknown. This is the first proposal of a constructed language of the a priori ('from beforehand', or philosophical) type, with a regular, exceptionless grammar and a logically connected word-stock, something that does not coincide with the structure of any known tongue. Though Descartes supplied no sample of his ideal language, many of his seventeenth-century contemporaries did. Among them were Dalgarno, Urquhart, Wilkins, and Leibniz. Urquhart's Logopartdecteision of 1653 described the ideal parts of speech, genders, cases, etc., and stipulated that there should be at least ten synonyms for each word, and that each letter of each word should have a specific semantic value. Dalgarno's Ars Signorum of 1661 offered a logical classification of ideas, with seventeen chief classes designated by an equal number of letters of the alphabet, then Greek vowels to designate subclasses (K for political occurrences, Ke for judicial affairs, Kv for crimes, etc.). Bishop John Wilkins' Essay toward a real character and a philosophical language of 1668 offered a set of parts of speech and a grammatical framework (Z indicates animals, Za restricts the concept to fish, successive consonants and vowels further restrict the concept to particular classes and subclasses of fish). Leibniz contemplated the use of either numerals or letter syllables. All of these early a priori systems are concerned more with a common language of thought and philosophy than with one designed for peasants, or even merchants and travelers. A seventeenth-century Czech educator, Comenius, or Komensk^, advocated a radically different type of solution, utilizing existing languages as zonal tongues, with Russian serving eastern Europe, French and English western Europe. This is a solution which rejects complete universality, but recognizes existing linguistic reality, and is far more in line with the bilingual world of antiquity. The seventeenth century may be said to have witnessed the birth of the two main streams of thought on the subject of international languages, the one that favors the use and extension of existing national and natural tongues, singly or in combination, and the one that prefers a language of the constructed or artificial variety. These, with their numerous ramifications, will be discussed in turn. Together, they account for at least one thousand projects, of uneven merit, that have been advanced since the days of Descartes. TYPES OF SOLUTION

After the seventeenth-century flowering of the idea of constructed languages, the

1002

MARIO A. PEI

eighteenth century, and even the first half of the nineteenth, are somewhat disappointing. Faiguet's Langue Nouvelle of 1765, appearing in volume 9 of Diderot and d'Alembert's Encyclopedia, offers a plan of grammar to be applied to roots that can only be described as French, with a simplification of endings, particularly for verbs, later utilized in part by Esperanto. In a sense, this offering constitutes our first sample of an a posteriori constructed language, based not on logic and numbers, but on elements drawn from existing languages. By way of contrast, Delormel's Projet d'une langue universelle of 1795 is little more than a repetition of earlier ideas, with arbitrary arrangements of consonants and vowels. Most interesting among early nineteenth-century projects is Jean François Sudre's Solresol, based on the notes of the musical scale, which attracted great attention and found sponsorship among such names as Victor Hugo, Lamartine, Humboldt, and Napoleon III. It continued to have followers until the early 1900s. Sudre's creation of 1817, also known as Langue Musicale Universelle (1866), formed all words out of combinations of the syllables do, re, mi, fa, sol, la, si, yielding seven words of one syllable, 49 of two, 336 of three, 2,268 of four, 9,072 of five, along with shifts of stress which permitted functional change from noun to verb, adjective, or adverb. The language could be sung, played, or hummed as well as spoken; it could be written as music: knocks or colors could be substituted for the seven syllables for communication at a distance; it could even be tapped out by deaf-mutes. While the vocabulary was completely arbitrary, a certain amount of classification was perceptible, along with some attempt to show opposites by reversing syllable order (domisol 'God', solmido 'Satan'; sollasi 'to go up', silasol 'to go down'). The grammar was both arbitrary and complicated. From the middle of the nineteenth century on, international language offerings multiplied. They can be classified, but with the understanding that boundaries are often not definite, and that some plans may fit into either of two classes. 1) Select and use a natural language just as it stands, be it a large, official tongue with many speakers (English, French, Spanish, Russian, Chinese), or a small, even an obscure language with no imperialistic or nationalistic features (Finnish, Hungarian, Nahuatl, Choctaw). The language may even be classical (Latin, Greek, Hebrew, Sanskrit). 2) Use two or more national languages, either as zonal forms of communication to serve specific large areas (this would give us not an international language, but a series of geographically separated international languages); or to be learned and used bilingually or trilingually by all. 3) Take a natural language, living or classical, and modify it in one fashion or another, so as to make it more accessible to foreign learners, and even to its own speakers. Here we have proposals to restrict vocabulary (Basic English), or to revise and regularize spelling, pronunciation, or grammatical structure. 4) Take two natural languages and combine them in various proportions. Sometimes the mixture is simple, involving speaking the language without change, but

ARTIFICIAL LANGUAGES

1003

writing it with the script of another language (English written in Cyrillic characters); sometimes it is complex, with sound-schemes, grammars, and vocabularies fused together, after the fashion of Franco-Venetian or Pidgin English. 5) Create a fully constructed tongue, which may come close to the modified national language or mixed language (in which case it is described as a posteriori), or depart from all natural languages (a priori). Many mixtures of a priori and a posteriori features are possible, combining blends of many natural languages with arbitrary features of grammar and word-building (Esperanto).

THE NATURAL LANGUAGES

It is both natural and simplistic to say: 'Let us use my language as an international tongue.' The implications of such a statement are: 1) 'My tongue is best suited for use by all men'; 2) 'I won't have to learn it, because I already know it.' Both statements reveal a highly subjective state of mind, which has to be rationalized. Accordingly, English speakers point to the already widespread use of English, its industrial, commercial, scientific, and political importance, the fact that many nonspeakers seem disposed to accept it. French speakers point to the use of French in the past as the international tongue of diplomacy and culture. Italian speakers can point to the sonority and ease of pronunciation of their language, Spanish speakers to the relative grammatical ease of theirs, and so forth ad infinitum. Among the natural tongues most frequently advocated for international use are two classical ones, Greek and Latin, two modern western languages, French and English, and, by reason of recent political developments, two additional great tongues, Russian and Chinese. Classical Greek has had as its advocates numerous French and German Hellenists who flourished in the latter part of the nineteenth century. Advocacy of Greek in its modern spoken form (in its more literary variant, the katharevousa, as opposed to the more popular demotike, modern Greek does not differ too strikingly from classical Greek), the proposal was first voiced in a letter of Voltaire to Catherine II of Russia, suggesting that if the Empress were to become ruler of Constantinople, she could open a Greek academy and sponsor Greek as the universal language. More recently similar proposals have been advanced on grounds that Greek is harmonious, easy, and accessible to all who are acquainted with the thousands of Greek roots in all modern civilized tongues. The demand for classical Latin is far stronger. The Latin tradition continued unbroken from the days of the Roman Empire to the late Renaissance and beyond for what concerns scholarship, while it is still unbroken in the religious sphere for what concerns the Roman Catholic church. George Henderson in 1890 launched a magazine called Phoenix, or The International Latin Messenger, composed in a medieval Latin augmented by various coinages to refer to modern objects (naves

1004

MARIO A. PEI

vaporariae for 'steamships', ferrae viae ordines for 'railroad lines'). In similar manner, the Praeco Latinus ('Latin Herald') was published monthly in Philadelphia between 1895 and 1902, and the Vox Urbis ('Voice of the City') was published twice a month in Rome. Ten Latin periodicals flourished between 1889 and 1914. More recently, there appeared R. Kent's Latin as the auxiliary language of 1922, as well as the Vatican's new Latin dictionary, brought up to date by Monsignor Antonio Bacci with such terms as ramenti sulphurati for 'matches' and radiophonica diurnorum actorum communicatio for 'daily radio newscast'. A UNESCO-sponsored Belgian film on 'Aspects of Imperial Rome in the fourth century' has its spoken commentary entirely in Latin. Perhaps the first hint that French might serve as an international language is Brunetto Latini's thirteenth-century statement that 'the French language is more delectable and more common to all men', made in justification of his having composed his Treasure of wisdom in French rather than in scholarly Latin or his native Italian. This pronouncement is all the more flattering to French because it comes from a foreign source. French advocacies of French begin in the sixteenth century, with du Bellay's Défense et illustration de la langue française and Henri Etienne's La Précellence du langage français. It is a fact, however, that even Marco Polo's account of his travels, dictated to Rusticiano da Pisa, appears in French, and that French was used during the Crusades at the courts of Jerusalem and Antioch, and at the English court until 1386. In 1784 the Academy of Berlin awarded a prize to Rivarol for his Discourse on the universality of the French language. French was used as the sole language of diplomacy from the seventeenth to the end of the nineteenth century, and was the official language of all treaties until Versailles, in 1919. It was only in 1920 that English was granted parity with French in the deliberations of the World Court. As late as 1902, H. G. Wells, writing in the Montreal Monde Illustré, suggested that we might all settle for French as a common tongue. The first advocacy of English was probably the letter written by David Hume to Edward Gibbon in 1767, in which he foretells the greater stability of English by reason of its firm grip on America. Many prominent linguists from non-Englishspeaking lands, including Jakob von Grimm and Otto Jespersen, advocated English during the nineteenth and early twentieth centuries. English became widespread during the eighteenth and nineteenth centuries by reason of British colonial and trade expansion, but it was the twentieth-century predominance of the United States, particularly after the second World War, that witnessed the greatest expansion of the language as a medium of international communication. English is today without doubt the language most widely studied outside of its own borders, as well as the tongue most widely used for purposes of inter-communication even by nations that deplore Anglo-American imperialism, as shown by the Bandung Conference of Asian and African nations, at which English was the primary official tongue. The advocacy of Russian and Chinese is somewhat muted, particularly in the

ARTIFICIAL LANGUAGES

1005

case of the latter, despite the fact that it is spoken by one out of every four persons in the world today. The Soviets, which encourage the study of English for practical purposes, view it with a jaundiced eye as an official international language, and have on occasion advanced the claims of Russian, describing it as 'the world language of socialism'. Many other languages, large and small, have had their advocates. The claims for German were rather strenuously advanced during the Hitler era, but have been somewhat subdued since 1945. Spanish, Italian, Dutch, Danish, Hungarian, even Malay and Pidgin English have at various times been suggested as the ideal solution to the world's language problems.

LANGUAGES IN COMBINATION

Comenius's seventeenth-century proposal to use English and French as tongues of common intercourse in western Europe and Russian in the east was only the first of many similar schemes, some of which, like Stalin's zonal language idea of 1950, are mere repetitions, others, like the proposal advanced by Pierre Fouche and L. Thommeret in 1951, far more elaborate, with four or more purely zonal languages (Russian for eastern Europe, Chinese for the Far East, Hindustani for southern Asia, Arabic for the Near East, Spanish for the Western Hemisphere), plus an over-all use of English and French throughout the world. This rather cumbersome system is more often replaced by the idea of a Monde Bilingue, in which English and French alone would function as international tongues, with all French speakers learning English, all English speakers learning French, and all others learning either English or French. This plan for a bilingual world, as originally launched in 1951 by Jean-Marie Bressand and Denise Poulain, carried with it the 'twinning' of English- and French-speaking cities (Coventry in England being paired off with St.Etienne in France, Lexington, Ky. with Montpellier, with the citizens of each city learning the language of and fraternizing with the citizens of the other). Something similar had been suggested as far back as 1901 by H. G. Wells. The latest suggestion of this type is the one advanced by George Hecht of Parents' Magazine to install English and Esperanto in all the world's schools.

MODIFIED NATIONAL LANGUAGES

The idea of taking a national language, but modifying it so as to make it more accessible to foreign learners, seems to have its starting point in Faiguet's 1765 work, already described, which makes use of a French vocabulary with arbitrary endings that give it an a priori touch. J. Schipfer's Communicationssprache of 1839 continues the use of a largely French vocabulary, on the ground that 'the

1006

MARIO A. PEI

French language is the best known and most widespread, both as a court and a colloquial tongue of the upper middle classes, to the point of being almost universal'. The grammatical framework, however, is largely of the a posteriori type, with no article, an invariable adjective that is compared Latin fashion (gran, grandior, grandiost), an adverb formed by adding English -ly, infinitives that end in -er, and other verb endings that are sometimes reminiscent of Latin, sometimes of French, sometimes of Italian. S. Bernhard's Lingua Franca Nuova of 1888 and his Welt-Italienisch Franca of 1891 offer a regularized Italian; R. de la Grasserie's Apolema of 1892 a regularized Greek. Lenz's Pasilingua Hebraica of 1887 and Nilson's Lasonebr of 1897 offer modified Hebrew, Ostwald's Weltdeutsch of 1916 and Baumann's Weltpitsch of 1928 modified German. Many projects involving simplified, regularized Latin have been offered through the centuries, starting with Teofilo Folengo's Macaronic Latin of the early sixteenth century, which had considerable vogue among students and scholars, despite its jocular aspects. Serious nineteenth-century projects include A. Volk and R. Fuchs' Weltsprache of 1883, E. Lauda's Kosmos of 1888, J. Lott's Mundolingue of 1888-1893, G. Henderson's Latinesce of 1901 and D. Rosa's Novlatin of 1890, A. Liptay's Langue Catholique of 1892, J. Stempfl's Communia of 1894, E. Beerman's Novilatiin of 1895, F. Kiirschner's Lingua Komun of 1900, F. Isly's Linguum lslianum of 1901, K. Frohlich's Reform-Latein of 1902. Most ambitious of all, perhaps, is G. Peano's Latino Sine Flexione of 1903, a fully classical Latin without classical endings. Additional offerings along this line continued through 1928. It will be noted that all attempts at modified natural languages so far described concentrate on modifying grammatical structure, leaving vocabulary, pronunciation, and spelling untouched. When we come to proposals to modify English, two new elements appear — modification of irrational English spelling and, most innovating of all, a restriction of English vocabulary so as to make a few words do the work of many. Advocacy of English spelling reform may, in a sense, be said to go back as far as the Anglo-Saxon Ormolum of about 1200, which shows a doubling of consonants after short vowels, indicating an attempt to reflect pronunciation. The sixteenth century offers attempts to 'improve' and regularize the written language on the part of such writers as Cheke, J. Hart, J. Bullokar, and Mulcaster. Samuel Johnson endeavored to promote the cause of 'correct' spelling, but without signal success. The real forces of true spelling reform appear first in modern times in America, with Benjamin Franklin and, particularly, Noah Webster. Only secondarily and remotely did these earlier attempts at English spelling reform aim at the use of English as an international language; they were mainly intended to facilitate the reading and writing process for native speakers of English. The twentieth century, however, shows the beginning of a link between the two ideas. Isaac Pitman's system of stenography, first launched in 1837, and the creation in 1888 of an International Phonetic Alphabet, based on Karl Lepsius's Standard Alphabet of 1856,

ARTIFICIAL LANGUAGES

1007

in which the sounds of all spoken languages would be phonetically portrayed, even Senator Robert Owen's Global Alphabet of the 1940s, still by-pass the question of internationality. But the use of phonetically spelled English as a universal language is at least implicit in the work of Britain's Simplified Spelling Board and America's Spelling Reform Association, and becomes far more explicit in the dozens of offerings that have appeared in recent times. Some are quite ingenious, but they are far too numerous to describe, or even enumerate. Several of their creators display the quaintly idealistic notion that it is only the complexities of present English spelling that prevent the enthusiastic, universal adoption of English as a world tongue. Far fewer in number are plans designed to phonetize other languages for possible international use. J. De Ria, in 1788, offered a form of phonetized French with a hint that it might become the world's choice, and Chappaz, in 1900, offered a Langage Instantané of the same description. One typical Anglo-Saxon creation is that of the language left undisturbed for what concerns spelling, pronunciation, and grammar, but used with a limited vocabulary. In the 1930s C. K. Ogden and I. A. Richards undertook to prove on the basis of a deep semantic study that it is possible to get along quite well in English with as few as 850 basic words, handled in normal English fashion, save for a restriction on the use of verb forms (use 'take part' instead of 'participate'; 'have hope' instead of 'to hope'). Churchill's advocacy of Basic English during the second World War led to widespread interest in the system for purposes of international use. Today it is regarded as primarily a learner's stepping-stone to the use of full English. In imitation of Basic English, G. Gougenheim more recently attempted a Basic French, with a similarly restricted vocabulary. There have even been attempts at Basic Spanish, but the structure of the Romance languages does not lend itself well to the form of modification advocated.

BLENDS OF TWO, THREE, OR MORE LANGUAGES

The fully a priori constructed language rejects all contributions from existing languages; the fully a posteriori tongue accepts such contributions from a variety of sources. Transitional systems usually display arbitrary features of grammatical structure coupled with vocabulary contributions from existing languages. In addition, there is a philosophical divergence between the a priori language, which normally bases itself on logic, and the a posteriori, which has the principle of the greatest ease to the greatest number. Lastly, if the a posteriori principle prevails, the question arises whether contributions should come from all existing languages, or only from the great tongues of western culture. The last-mentioned preoccupation has given rise to various controversies. Eichhorn's Weltsprache of 1887 shows concern for Polynesian habits of pronunciation,

1008

MARIO A. PEI

and reduces the number of consonants accordingly. Volapuk discards the letter and sound of r because of its supposed difficulty to Chinese speakers (interestingly, it does not discard I, which is equally difficult to speakers of Japanese). In 1888 the American Philosophical Society condemned Volapuk because of its 'non-Aryan' character, thereby clashing with the Philological Society of London, which deplored 'racialism' in language. At times Esperanto has been criticized on grounds that it shows undue favor to the western languages, while such writers as Albert Guerard and Alexander Gode point to the fact that a constructed language based on proportional representation would be a worse hodge-podge than the most a prioristic of a priori tongues. Many of the more recent constructed offerings frankly espouse the cause of a two or three-language blend. George Henderson in 1888 proposed a blend of Latin vocabulary with English syntax, and one year later displayed his versatility by suggesting an Anglo-Franca that ingeniously mixes English and French. J. Puchner's Nuove-Roman of 1897 is an almost perfect combination of Italian and Spanish. Vanghetti's Latin Esperanto of 1913 blends Latin vocabulary with Esperanto endings. There are numerous blends of various Slavic tongues (Slovan is a recent one), and of Germanic languages (Molee's Altutonish of 1912, for instance). One ingenious creation, W. Cheshikhin's Nepo of 1907 and 1915, blends Latin-Romance, Germanic, and Slavic. Lancelot Hogben's Interglossa of 1943 unites a Greek vocabulary with Chinese syntax. One might even say that the American Interlingua sponsored by IALA (International Auxiliary Language Association) is a blend of Pan-Romance, permitting the participation of English, German, and Russian only where they happen to coincide with the Latin-Romance majority of pilot tongues. Leidenfrost's very recent project for a Universal grammar suggests that it be based on ten representative languages: Arabic, Chinese, English, Hindustani, Hungarian, Indonesian, the Kpelle of Liberia, Russian, Spanish, and Swahili.

A PRIORI LANGUAGES

The earlier school of constructed languages based on Descartes' initial suggestion continued in the seventeenth century with Francis Lodwick, Cave Beck, and Leibniz, all of whom combined the use of letters and numerals in writing with the replacement of the numerals by spoken syllables in speech. When we examine Beck's pf2477, pronounced piftofosensen, or Leibniz's 81374, pronounced bodifalemu, we see why such systems did not take hold. The late eighteenth and early nineteenth centuries saw many offerings of the a priori type: Delormel's logical sequences of vowels and consonants (ava 'grammar', ave 'letter', avi 'syllable', avo 'accent', avau 'word'); A. Grosselin's 1,500 roots and 100 affixes capable of forming 150,000 words, offered in 1836; C. L. Letellier's device of 1852. Sotos Ochando's project of 1855, Maldant's Chabé Abane of 1886;

ARTIFICIAL LANGUAGES

1009

Dyer's Lingucdumina of 1889, G. A. Larsson's Luftlanda of 1893. In the early 1900s we have purely numerical systems, such as Tiemer's Timerio of 1921, and C. M. Gibson's Code; but also fully spoken languages built on logic, like Talundberg's Perio of 1904, or Foster's Ro of 1912, in which the root ta- covers expressions of time; then tab is 'moment', tac 'minute', tad 'hour', taf 'day', taj 'night', tak 'week', tal 'month', taq 'year'. Among European contributions to a priori systems are Gigli's Lingua Filosofica Universale of 1818, G. Ferrari's Monoglottica of 1874, Meriggi's Zimondal of 1884. Several a priori languages show the beginning of a shift to the a posteriori type. Nicolas's Spokil of 1900, for instance, makes use of gn (from Latin ignis) for words dealing with fire, dr (from Greek hydro-) for water, pn (from Greek pneuma) for air, /r for fruit, skr for write. The principle of mnemonic devices is carried still farther in Beatty's Qosmiani of 1922 and Hilbe's Zahlensprache of 1901. Among the most recent and purest offerings of the a priori type is Dr. Barnett Russell's Suma of 1957, recently revised. Stray samples of Suma are nea vuni puto 'disagreement' (not agree deed), moti pute 'beginner' (begin doer); kosa puti 'enlarge' (large do), toki kolo 'restaurant' (eat house).

MIXED SYSTEMS

The idea of combining simplicity and logic of grammatical structure with ease of sounds, fully phonetic spelling, and a familiar vocabulary was first advanced by Jakob von Grimm in 1860. The familiar vocabulary he presented as most desirable was that of Latin, because Latin is 'neutral' with respect to present-day languages, related to all languages of the Indo-European family, and familiar to the more cultured elements of all nations. It is easy to recognize in this pronouncement the principle of 'the greatest ease to the greatest number'. First to attempt to apply this principle in full was Monsignor Johann Schleyer with his Volapiik, presented in 1885. Volapiik spread like wildfire, and societies for its propagation appeared all over the world. But by 1890 the vogue of Volapiik had come to a standstill, and by 1911 its followers had shrunk in numbers to barely one-third of those who followed the green star of Esperanto. The basis of Volapiik was a highly modified English vocabulary (the name Volapiik itself is a combination of English 'World speak'), but with a grammatical structure strongly influenced by its creator's German-language habits, with four cases and very long compounds. Both noun and verb endings were quite arbitrary, in accordance with a priori principles. Volapiik's descendants and imitators were quite numerous. Among them may be mentioned such poetically named creations as Bopal, Spelin, Dil, Balta, Orba. Arnim's Veltparl of 1896 applied a declension system to the article rather than to the noun, and further distinguished itself by granting some sort of proportional representation to numerous lesser tongues, Hungarian, Hindustani, and Vietnamese

1010

MARIO A. PEI

among them. Marchand's Dilpok of 1898 and Bollack's Langue Bleue (1899) begin to show the preference for a mixed Romance-Germanic vocabulary which came to fuller fruition in later a posteriori systems. One rather recent sample of a mixed language is Kenneth Littlewood's Monling, which combines English roots with a Chinese syntactical arrangement.

A POSTERIORI SYSTEMS

The a posteriori language is by definition constructed on the analogy of existing tongues. This does not mean that it may not present arbitrary features, particularly in grammar. But it normally eliminates the irregularities of the natural languages on which it is based. Many of the earlier a posteriori languages were in reality mixed languages, to the extent that they combined vocabularies taken from many natural tongues, but tended in their structure to favor one or another of these. Among them may be listed de Rudelle's Pantos-dimou-glossa of 1858, Pirro's Universalsprache of 1868, Courtonne's Langue Internationale Néo-Latine of 1885, and Steiner's Pasilingua of the same year. Pasilingua made use of copious vocabulary synonyms, the choice of which was left to the speaker ('good' may be either bono or guto; 'language' may be either lingua or spracha). Most successful among a posteriori languages, having survived to the present day with a considerable number of speakers, a flourishing literature, both original and in translation, and a limited amount of international official recognition, is Zamenhof's Esperanto of 1887. Its author, inspired by highly idealistic principles, based his creation on absolute phonetization, easy pronunciation, complete regularity of grammatical forms and rules, and a vocabulary based on the languages most current in his days: Latin, Greek, English, German, and the Romance tongues; with these features he combined a system of word-formation that offers complete flexibility, predictability of forms you may never have seen or heard before, and the possibility of infinite expansion to accommodate all the needs of a rapidly changing world. The international popularity of Esperanto among both masses and scholars down to the outbreak of the first World War was such that many confidently predicted its international adoption. The unfortunate circumstances of that conflict were such as to interrupt its onward march, which was resumed with difficulty in the new world that emerged and that featured such absolutistic forms of government as the Communist, Fascist, and Nazi. Today Esperanto is without question the leader among constructed languages, being favored for its neutrality by numerous smaller nations that resent the predominance of large, aggressive national units and their languages, as well as by ever widening groups within the larger nations themselves. It seems reasonably safe to predict that if the ultimate choice is to fall upon a constructed rather than a national language, Esperanto will be that choice.

ARTIFICIAL LANGUAGES

1011

The appearance and success of Esperanto, however, were far from discouraging the creative spirit of hundreds who thought they could improve on it. First and foremost among these were the dissidents who branched away from the Esperanto movement with a slightly modified version called Esperantido or Ido. It would be idle to attempt to enumerate all the offerings that have been presented in the course of the twentieth century. A few that received some measure of attention before and between the two World Wars are Rosenberger's Idiom Neutral of 1902-1907, Molenaar's Universal of 1903, Michaux's Romanal of 1917, Jespersen's Novial of 1928. Cosmoglotta (or Interlingue) is a language that still flourishes in European circles. Interlingua is an American creation, concocted by a group of international linguists and designed to serve primarily as a written language for scientific congresses. It is overwhelmingly Pan-Romance in structure and vocabulary. THE PROBLEM IN THE FUTURE

Gallup and other polls taken within the last twenty years in various countries indicate that at least 75 % of those polled favor the idea of an international language to be placed in the school systems of all countries that boast such systems, on a basis of parity with the national tongues, so that the coming generations may grow up completely bilingual, with one of the two languages they speak with equal fluency having validity in all parts of the world to which they may stray. The real difficulty lies in the choice of the language that will serve as a universal tongue. Outside of the initial tendency of the speakers of each language to vote for their own, the polls indicate that English is the first choice among natural languages, with French a somewhat distant second. Esperanto is not merely the first, but practically the only contender among constructed languages. The merit of English lies in its already widespread use, based largely upon the achievements of its speakers in the fields of science and technology, industry and commerce, international politics, and military affairs. Its drawbacks he in its fluctuating and highly dialectalized forms, with no true spoken standard, its lack of fit between pronunciation and spelling, its over-numerous idiomatic constructions. To these is added the not very valid objection that English is the voice of Anglo-Saxon imperialism, commercialism, and colonialism. English is nevertheless a formidable contender, and many of its advocates feel that no particular effort is needed to establish it officially, since it will eventually become the universal language by a process of natural expansion and acceptance. This is perhaps an over-optimistic view of a situation that is subject to historical shifts, often of a sudden and drastic nature. French rests its claims mainly on traditional, historical and cultural grounds, as well as on its continued widespread use among the more educated classes of most nations. French stands on firmer ground than English to the extent that it possesses a thoroughly standardized form concerning which there is no uncertainty. In the

1012

MARIO A. PEI

matter of spread between pronunciation and spelling, it is almost, though not quite, as badly off as English. It is less open than English to the charge of being the mouthpiece of a way of life and world outlook that many nations view with suspicion and misgiving. While its position is weaker than it has been in the past, it is unwise to rule French out as a serious contender. Constructed languages (specifically, Esperanto) are challenged on the ground that they have no 'grass roots', no cultural tradition, and that their present speaking population is small in comparison with those of the world's linguistic giants. On the other hand, the best among them (again, notably, Esperanto) present an acceptable measure of 'neutrality', coupled with full standardization and phonetization. If there is to be a deliberate choice, it will have to be made by the world's governments, which alone have the power of implementation through their control of educational facilities. So far, the world's governments have not shown any great disposition to consider such a choice. The fact nevertheless remains that if world travel, world communications, the interchange of persons and products, continue to expand over the remaining years of our century at the same astounding rate at which they have expanded since 1900, an eventual choice will become imperative, under penalty of strangulation of world communications and intercourse.

BIBLIOGRAPHY

Three essential source-books are: L. Couturat and L. Léau's Histoire de la langue universelle (1903); A. Guérard's A short history of the international language movement (1922); and P. E. Stojan's Bibliografio de internada lingvo (1929). Couturat and Léau discuss constructed tongues up to 1903, along with polygraphy (pp. 1-10) and proposals to use dead languages (pp. 515-41). Each system discussed receives a detailed description and searching criticism. Bibliography is scattered throughout the footnotes. Guérard discusses artificial languages, proposals to use French, English, and Latin, and the possibility of a French-English condominium. A detailed, point-bypoint comparison of Esperanto, Ido, the older Interlingua, and Romanal appears. Appendix I (pp. 211-15, 'Bibliographical notes') presents a list of articles on the international language problem by such philologists and linguists as Bréal, Brugmann, Leskien, Courtenay, Jespersen, de la Grasserie, Meillet, P. Meyer, F. Miiller, Paul Régnaud, Schuchardt, and Sweet. Appendix II (pp. 216-19) contains 'A tentative list of artificial language projects up to 1913'. Stojan's work is a true bibliography, making a definite attempt at completeness and listing 321 artificial languages offered up to the time of its appearance. It includes a description of the work of polygraphists who tried to compose a 'universal character' by which people of different languages might be able to communicate

ARTIFICIAL LANGUAGES

1013

(Ramon Llull, L. Alberti, J. Trithemius, Nostradamus, G. B. Porta, T. Campanella, J. Wilkins, S. Hartlib, C. Beck, A. Kircher, J. J. Becher, Pierre Besnier, A. Kochanski, etc.). Listed also are attempts at a 'universal grammar' (Duns Scotus, G. S. Scaliger, J. Harris, A. Humboldt, A. Trombetti, etc.), and at universal alphabets (since 1412: C. Volney, Ellis, Lepsius, A. M. Bell, P. Rousselot, P. Passy, W. Viëtor, Jespersen, N. Marr), as well as international stenography since 1659, gestural languages for deaf-mutes since 1614, sign languages and symbolic codes since 1534 (including Dewey's Decimal classification [1885], now used by libraries). Cryptography and secret and commercial codes are also listed, as well as proposals to use Greek, Latin, French, English and other languages for international purposes, and opinions expressed by famous philosophers, writers, and linguists on the subject of a universal language. Constructed-language projects are listed, but only occasionally and very briefly described. More recent works of a bibliographical nature include: M. Monnerot-Dumaine's Précis d'interlinguistique générale et spéciale (1960); P. Ronai's Der Kampf gegen Babel (1969); and M. Pei's One language for the world (1961). Many of the following references are out of print. Where the publisher is not given, it is to be assumed that the author published his book privately. Only those works are listed which appear in the discussion above, plus a certain number of others that have seemed significant. W. VON. 1896. Entwurf einer internationalen Verkehrssprache, genannt Veltparl. Oppeln, Maske. BACON, R. [See R. E. Robins 1 9 5 1 . Ancient and medieval grammatical theory in Europe, 77. London.] BAUER, G. 1888. Spelin. Agram (Zagreb), Suppan. BAUMANN, A. 1925-26. Weltpitsch. München. BEATTY, W. 1922. Qosmiani. Washington, D.C. BECK, C. 1657. Universal character. London. BEERMANN, E. 1895. Novilatiin. Leipzig, Fock. BELLAY, J. DU. 1914. Défense et illustration de la langue française, ed. by A . Humbert. Paris, Garnier. BERNHARD, S. 1 8 8 8 . Grammatik der Lingua Franca Nuova. Wien. . 1891. Welt-Italienisch Franca. Wien. BOLLACK, L. 1 8 9 9 . La Langue bleue. Paris. BRUGMANN, K., and A. LESKIEN. 1907. Zur Kritik der künstlichen Weltsprachen. Strassburg, Trübner. CHAPPAZ, J. M. 1900. Langage instantané. Ville-la-Grande (Haute Savoie). CHAPPELIER, P. 1911. L'Espéranto et le système bilingue. Paris, Grasset. CHESHIKHIN, W. 1907. Nepo. Riga. . 1913. Gram. Ideografiya i Neosinografiya. Riga. CHRISTIAN, M. D. 1946. Greek as an international language. New York. ARNIM,

1014

MARIO A. PEI

CoMENius

(KOMENSKy,

J.). 1641 (1668). Via Lucis. [Amsterdam, Cunradum,

1668.] E. 1885. Langue internationale néo-latine. Bulletin de la société niçoise des sciences naturelles. Nice, Visconti. COUTURAT, L . , and L . L É A U . 1 9 0 3 . Histoire de la langue universelle. Paris, Hachette. . 1907. Les nouvelles langues internationales. Paris, Hachette. DALGARNO, G. 1 6 6 1 . Ars Signorum. London. DELORMEL, J . 1 7 9 5 . Projet d'une langue universelle. Moniteur, 2 7 , Brumaire III. Paris. DESCARTES, J . 1897. Lettre au P. Mersenne (Nov. 20, 1629), ed. by AdamTannery, 1.76-9. Paris, Cerf. DEWEY, J . 1 8 8 5 . Decimal classification. Boston. DYER, F. 1889. The Lingualumina. London. EICHHORN, N. 1 8 8 7 . Die Weltsprache. Bamberg. ELLIS, A. 1891. On the conditions of a universal language. TPhS 1888-1890. 59-98. ETIENNE, H. 1850. La Précellence du langage français (1579). Paris, Delalain. FAIGUET, M. 1765. Langue Nouvelle. Encyclopédie of Diderot and d'Alembert, IX. Paris. FERRANTI, M. 1911. Simplo. Roma, Tipografia Italia. FERRARI, G. 1 8 7 4 . Monoglottica. 2nd ed. 1 8 7 7 . Modena. FOSTER, E . P. 1 9 0 8 . Ro. Cincinnati. FRÖHLICH, K . 1 9 0 2 . Reform-Latein. Wien. GAJEWSKI, B. 1902. Grammaire du Solresol. Paris. GIGLI, M. 1 8 1 8 . Lingua filosofica universale pei dotti. Milano. GODE, A . , and E . BLAIR. 195la. Interlingua: A grammar of the international language. New York, Storm. . 1951b. Interlingua-English dictionary. New York, Storm. GRASSERIE, R. DE LA. 1892. De la possibilité et des conditions d'une langue internationale. Paris, Maisonneuve. . 1907. Apolema. Paris, Leroux. GRIMM, J. VON. 1851. Ueber den Ursprung der Sprache. Berlin, Abhandlungen der KgJ. Akademie. GUÉRARD, A . 1 9 2 2 . A short history of the international language movement. London, T. Fisher Unwin. HEIMER, H . 1957. Mondial. Lund, Gleerupska. HENDERSON, G. (HOINIX). 1889. Anglo-Franca. London, Trübner. . 1901. Latinized English (Latinesce). The Referee (January). London. HILBE, F. 1 9 0 1 . Die Zahlensprache. Feldkirch. HOGBEN, L . 1 9 4 3 . Interglossa. New York, Penguin. COURTONNE,

ARTIFICIAL LANGUAGES

1015

1932. Letter to E. Gibbon (Oct. 24, 1767). Letters of D . Hume, ed. by J. Y. Grieg, H.170. Oxford, Clarendon. ISLY, F. 1901. Langue Isly. Paris, Richard. JESPERSEN, O. 1928. An international language (Novial). London, Allen & Unwin. KENT, R. G. 1922. Latin as the international auxiliary language. Classical Journal, Princeton, XVIII/1.38^4 (Oct.). KERCKHOFFS, A . 1 8 8 5 . Cours complet de Volapiik. Paris, Le Soudier. KÜRSCHNER, F. 1900. Die Gemeinsprache . . . . Lingua Komun. Orselina sur Locamo. LARSSON, G. A. 1893. (Ms.) Luftlandana. LAUDA, E . 1888. Kosmos. Berlin, Hennig. LAVAGNINI, A. 1 9 2 3 . Monario, Unilingue. Roma, Ars Nova. LEIBNIZ, G. W. (ca. 1679); [See Couturat, L., 1901. La Logique de Leibniz, eh. 3. Paris, Alcan; and . 1903. Opuscules et fragments inédits de Leibniz, Phil. VII, B, III. Paris, Alcan.] LENZ, F . 1 8 8 7 . Pasilingua Hebraica. Neuwied, Heuser. LIPTAY, A . 1 8 9 2 . Langue catholique. Paris, Bouillon. LOTT, J . 1 8 9 0 . Grammatik der Weltsprache, Mundolingue. Leipzig. MALDANT, E. 1886. Chabé Abane, la langue naturelle. Paris, Jouve. MARCHAND, J. A . 1 8 9 8 . Dilpok. Besançon, Jacquin. MEAZZINI, G. 1 9 2 8 . La lingua internazionale I.D.O. (reformed Ido). Giornale del Jonio (April). Catania. MERIGGI, C . 1 8 8 4 . Blaia Zimondal. Pavia, Fusi. MICHAUX, A . 1 9 1 7 . Romanal, une langue internationale anglo-latine. Boulognesur-mer, Boningue. MITROVIC, P. 1954. An essay on interlinguistics. Sarajevo. MOLEE, E . 1 9 1 2 . Altutonish. Tacoma. MOLENAAR, H. 1903. Panroman. Leipzig, Uhlig. MONNERET-DUMAINE, M . 1 9 6 0 . Précis d'interlinguistique générale et spéciale. Paris. MOSER, H . 1 8 8 8 . Die Weltsprache. Praha. NICOLAS, A. 1 9 0 0 . Spokil. Mémoires de la société nationale d'agriculture, science et art d'Angers. Angers, Lachèse. NILSON, A . 1 8 9 7 . Lasonebr. Gefle. NODIER, C . 1 8 3 4 . Du langage factice appelé macaronique. Paris, Techner. OGDEN, C. K . 1936. Basic English. 9th ed. London, Paul Trench, Trubner & Co. OSTWALD, J. 1916. Weltdeutsch. OWEN, R . L. Statement in Charter of United Nations, pp. 534-49 [See also Senate Documents 49, 133, 150 of the 78th U.S. Congress], PEANO, G. 1903. De Latino Sine Flexione. Revue des mathématiques Vili. . 1909. Vocabulario Commune ad Linguas de Europa. Torino, Bocca.

HUME, D .

1016

MARIO A. PEI

PEI, M. 1944. A universal language can be achieved. Town and Country. New York (September). . 1952. Six languages for one world. New York Times Magazine (Feb. 24). . 1961. One language for the world. New York, Devin-Adair. PIRRO, M. 1868. Universal-Sprache. Paris, Rétaux. PUCHNER, J. 1897. Nuove-Roman. Linz. REED, I. K. 1957. Esperanto and interlingua compared. New York, Esperanto Association of North America. RICHARDS, I. A., and C. M. GIBSON. 1945. Learning Basic English. New York, Norton. RIVAROL, A. DE. 1784. De l'universalité de la langue française. Paris, Bailly, Dessenne. RÓNAI, P. 1969. Der Kampf gegen Babel. München, Ehrenwirth. ROSA, D. 1890. Le Nov Latin. Bollettino dei musei di zoologia ed anatomia comparata della R. università di Torino, V / 8 9 . Torino, Clausen (Oct. 15). ROSENBERGER, W.

1902-07.

Grammatik und Wörterbuch der Neutralsprache

(Idiom Neutral). Leipzig, Haberland. RUDELLE, L. DE. 1858. Grammaire primitive . . . Pantos-dîmou-glossa. Paris, Delalain. RUSSELL, B. 1957. Suma (the 1000-word universal language). Gardena, Cal. SCHIPFER, J. 1839. Versuche einer Grammatik . . . Communications-oder Weltsprache. Wiesbaden. SCHLEYER, J . 1885. Grammatik der Universalsprache für alle Erdbewohner, Volapiik. Konstanz. SCHUCHARDT, H. 1894. Weltsprache und Weltsprachen. Strassburg, Trübner. SOTOS OCHANDO, B. 1855. Projet d'une langue universelle. Paris, Lecoffre. The Soviet linguistic controversy. 1951. Translated from the Soviet press by J. V. Murra, R. M. Hankin, and F. Holling. New York, King's Crown Press. STEINER, P. 1885. Elementargrammatik . . . Pasilingua. Neuwied, Heuser. STEMPFL, J. 1894. Communia. Kempten, Dobler. STOJAN, P. E. 1929. Bibliografio de internada lingvo. Genève, Universala Esperanto-Asocio. SUDRE, J . F. 1866. Langue musicale universelle. Paris. TALMEY, M. 1919. Ido. New York, Ido Press. TALUNDBERG, M. 1904. Perio. Elberfeld, Wasserloos. THIERFELDER, F. 1938. Deutsch als Weltsprache. Berlin, Kurzeja. THOMMERET, L. 1947. Pour l'organisation linguistique universelle. Paris, Rodstein. TIEMER. 1921. Timerio. A numerical language. Berlin. UNESCO. 1957. Scientific and technical translating and other aspects of the language problem. Paris and Genève. URQUHART, T. 1653. Logopandecteision. London, Calvert & Tomlins.

ARTIFICIAL LANGUAGES

1017

1913. Latin Esperanto. and R. FUCHS. 1 8 8 3 . Die Weltsprache. Berlin, Kühl. VOLTAIRE. 1 8 6 9 . Letter to Catherine II. Reprinted in Le Temps, Paris (May 2 0 ) . WAHL, E . DE. 1 9 2 5 . Radicarium Direktiv del Lingue International, Occidental. Tallin, Uehiselu. WEBSTER, N. 1806. Compendious dictionary of the English language. Hartford, Hudson and Goodwin. WELLS, H. G. 1 9 0 1 . Anticipations. London, Chapman and Hall. . 1902. La langue universelle. Monde Illustré, Montreal (March 15). WHORF, B. L. 1941. Languages and logic. Technology Review (April). WILKINS, J. 1668. Essay toward a real character and a philosophical language. London. ZAMENHOF, L. (Doktoro Esperanto). 1 8 8 7 . Langue internationale. Warszawa, Gebethner & Wolff. ZASLAVSKY, D. 1949. Veliky yazyk nashey epokhi. Literaturnaya Gazeta 1.3 (January 1). Moskva. VANGHETTI, G . VOLK, A . ,

COSMIC L A N G U A G E

H. FREUDENTHAL

1. Artificial languages are mainly of two kinds, general purpose languages, and languages restricted to the communication of specified subject matter or to communication within a specialized environment. Chemistry formulae, staves, and secret languages belong to the latter. The endeavour of the first is to reverse the legendary feat of the Babylonian linguistic confusion. For centuries investigations were made to restore the original language of mankind whatever this might have been; the stress shifted in the seventeenth century to the construction of artificial general purpose languages. Most of these languages borrowed their stems and all their syntax from existing languages, although authors and supporters of these languages are not likely to admit it. Since constructing such languages has never been preceded by a profound logical analysis, the authors usually do not know to what degree they depend on the natural languages they are acquainted with. These artificial languages offer greater regularity than the natural ones, but the attempts at regularity were restricted to wordbuilding from stems and to the arrangement of the words in a sentence. The proper syntactic means, however, such as connectives, quantifiers, and punctuation were, without much criticism, borrowed from existing natural languages. Often the artificial languages are more systematic than the natural ones but this need not be an advantage unless the creator has proved able to canalize his own systematism. Though they often claim to be, they are not usually more logical than the existing languages. 2. The use of mathematical formulae has been an old and conspicuous special purpose language. It freed itself from the vernacular and took its present shape about 1600 to be subsequently greatly amplified and perfected. Mathematical literature is generally written in a mixed language: the language surrounding the subject matter formulae is a vernacular, though a version that has been adapted to the special needs of the mathematical content. The first attempts to formalize the language around the formulae date from about a century ago. G. Peano was the first to undertake the creation of a formalized language designed to communicate mathematical investigations. Peano's idea has made very slow progress. Even now the overwhelming part of mathematics is still

1020

H. FREUDENTHAL

communicated in a more or less formalized vernacular. Rather than a vehicle of communication, formalized language became a device and the subject matter of investigations into the foundations of mathematics, which did not interest the overwhelming majority of mathematicians. The idea that formalized language could be a vehicle of communication is still far from being accepted, and it is rarely discussed although, in fact, more and more mathematicians grew accustomed to using at least some formalized language in communicating mathematics. Peano was an able mathematician, who knew perfectly what kind of formalization mathematics needed. The business of formalization was, however, seriously discredited by the intervention of G. Frege who created a forbidding formalism of his own, who connected the idea of formalization to mistaken ideas on the foundations of mathematics, and who cultivated such shallow opinions of contemporary mathematics that he bluntly and by ridiculous arguments rejected axiomatics, which meanwhile has pervaded all of mathematics. The most elaborated specimen of formalizing mathematical language is Russell and Whitehead's Principia Mathematica, though its scope was rather the foundations of mathematics. The use of formalism to serve the foundations of mathematics became still more manifest in Hilbert's work; by means of formalism Hilbert tried to solve such profound problems as that of the consistency of mathematics. Mathematical language has undergone momentous changes in the last forty years. The present set theory symbols, which were first propagated in the thirties, are now accepted by the great majority though quite a few mathematicians still adhere to older notations. The revision of the function notation (f instead of /(*)) took the same amount of time to become effective, though many problems concerning the function notation are still badly handled. Logical symbols are more and more often used in mathematical texts though this use is often spurious, for instance if '3' is naively understood as an abbreviation for 'there is', and the syntax around'3' is borrowed from the vernacular. What is more important is that mathematicians became more formalist minded; the process of adapting the vernacular to the needs of mathematical expression is now more conscious than it has ever been, and mathematicians are becoming accustomed to weighing more deliberately the pros and cons of linguistic solutions. 3. It is useful to explain the needs of mathematical expression in more detail and to explain how they were traditionally satisfied. The traditional objects of mathematics are numbers and geometric figures. Names of numbers are an early invention of mankind. A system to name all numbers is implicit in the use of the abacus and explicit in the Babylonian number notation. Naming geometric figures was not as easy. You cannot name every circle or triangle of the plane individually as you can men and integers. (Analytic geometry, that offers a means of naming, is a late invention.) In another respect circles and triangles present fewer problems in this regard than numbers: often you need not

COSMIC LANGUAGE

1021

name them since you can show them. Numbers can be named, geometrical figures can be shown. The teacher draws the triangle he refers to, with all auxiliary lines, and to make his arguments clear he points to them saying 'this Une' and 'that line', as he would do with 'this stone' and 'that stone', since stones, too, do not have individual names. But written communication of geometry asks for a more sophisticated notational system. We know this from Aristotle and Euclid, but the numbering or naming of the characteristic points of a geometrical drawing by letters and indicating lines or other parts of the figure by combinations of such letters is certainly much older. If 'A', 'B', 'C' are placed near points in a drawing, they are meant as conventional proper names of those points, by which these points can be quoted. Since, however, geometry is never about an individual triangle or circle, and since thanks to the generality of geometrical truth and arguments, the individual figure is only a paradigm, 'A', 'B', 'C'. . . develop into generic names. Though in an actual drawing 'A', 'B', 'C'... are proper names of specific points, they may and can be used as names of any point. This is the origin of variables as used in mathematics. Letters are applied to indicate not only points, but numbers and other mathematical objects as well. Of course, generic names such as 'man' and 'stone' are quite close to mathematical variables. 'Man' can mean any man, and 'stone' any stone, but as soon as many men or many stones are under consideration, these generic names, even with specifying adjectives, are insufficient. In mathematics however, we have created a rich stock of general purpose variables which according to the need can be restricted to some species of mathematical objects. 4. The rich variety of general purpose variables is one feature that distinguishes mathematical language from any vernacular. Another feature is the uniform and explicit means of binding variables. The following examples show how binding variables appear in the vernacular and are expressed in a mathematized language. A car is a vehicle — for all x, if x is a car, x is a vehicle [binding by the universal quantifier, a x F ( j c ) ] . I own a car — there is an x such that x is a car and I own x [binding by the existential quantifier, A, F(x)]. The car I own — the one x such that x is a car and I own x [article binding The number of cars in Paris is one million, or rather the number of the set of cars in Paris is a million — the set of x such that . . . [set-building binding, \x F(x)]. Being a car [functional binding, YxF(x)]. I own this car — no analogue in formalized mathematics [demonstrative binding].

1022

H. FREUDENTHAL

Which car is mine? — asked for the x such that x is a car and I own x [interrogative binding, ?x F(jc)]. The variable 'instant' if bound by universal quantification becomes 'always'; by existential quantification it becomes 'sometime', by article binding it becomes 'when', by set building binding it can become 'whenever', by demonstrative binding 'now' or 'then', by interrogative binding 'when'. In a mathematized language the indication of the means of binding would be more uniform: for all x if x is an instant . . . , there is an instant such that..., the x such that x is an instant and . . . , the set of all x such that x is an instant and . . ., asked for the x such that x is an instant and . . . The need for careful systematization in indicating the means of binding is felt in mathematics as soon as binding devices are to be applied in succession. Even a succession of two quantifiers can be a troublesome feature in the vernacular. Clearly the difference in succession between the universal quantifier of time and the existential quantifier of people is clearly indicated in 'someone has always been here' and 'always someone has been here', as is that between the existential quantifier of time and the universal quantifier of people in 'once everybody believed in God' and 'everybody once believed in God'. But more often the vernacular is utterly vague about such fine points. It can afford this lack of precision as long as meaning is guaranteed clear by the concreteness of the situation. Mathematical expression, however, demands more precision. Alternating successions of three or more quantifiers in one statement are not unusual in mathematics. The logical depth of mathematical statements has greatly increased in the course of a century and is still increasing; to understand such statements better, explicit quantifiers are a useful tool. 5. All natural languages I am acquainted with have developed an involved system of referring to a single object by a variety of variables. In the sentence He said she left her umbrella in his car 'he' and 'his' are variables bound to the same person, as are 'she' and 'her'. A more involved structure is I told the man who the day before on Main Street had sold my daughter his car she would have it picked up there tomorrow by her husband. Here T and 'my' are bound to the speaker, 'the man', 'who', 'his' are bound to the seller, 'daughter', 'she', and 'her' are bound to the speaker's daughter, 'car' and 'it' are bound to a certain car, 'on Main Street' and 'there' are bound to the same spot, 'the day before' and the tense of 'to sell', as well as 'tomorrow' and the tense of 'to pick up' include a reference to an instant which is implicit in the past tense of 'to tell'. For a long time the communication of mathematics by means of such linguistic

COSMIC LANGUAGE

1023

tools has been attempted. Up to a few decades ago it was still the approved style. But in modern algebra, topology, and, in particular, functional analysis, mathematics needed so much more involved linguistic structures that the traditional tools of the vernacular could no longer be usefully applied. In the traditional setting a certain situation with mappings of sets would be described as follows: Let a set be mapped into another and the last one into still another, then the mapping of the first into the second and that of the second into the third can be combined to give a third mapping. The subject is three sets and three mappings which are referred to by numerous variables — different variables for the same object but also occasionally the same variable to refer to different objects. The current style would be Let A, B, C be three sets, f a mapping of A into B, g a mapping of B into C; then / and g possess a composed mapping gf of A into C. If this transformation were applied to our earlier examples, they would read: x said (y left y's umbrella in *'s car). x says at the time ti: {* at a time to before ii says to y [y at a time one day before to at the place z sells q (q is / s car) to u (u is *'s daughter)]: [u at some time orders v (v is u's husband): v one day after to at the place z picks up q]}. 6. The last few examples should evoke a vague idea about how variables are dealt with by mathematically influenced language. Another feature of the texts is the occurrence of a large number of brackets and braces — actually in a more consequential formalization there could have appeared many more. Brackets and braces, which are rare in the vernacular, are indispensable and widely used elements of the language of mathematical formulae. Without brackets one would not know in which order the operations should be carried out in expressions like (a+b)

(a-b)

Punctuation devices as brackets and braces are the structuring backbone of mathematical formulae. Punctuation fulfills the same function in highly formalized texts around the formulae. In the written vernacular punctuation (except the period) is a redundant element that fulfills no essential tasks. Instead, the text is structured by means of relative pronouns and conjunctions. More often structuring devices are lacking, and the structure has to be derived from the content. Admittedly, the classical example pretty little girls school with its seventeen meanings is an ad hoc construction but the clause It was the directness, genuineness, and sincerity of the mother that I saw

1024

H. FREUDENTHAL

reflected in the daughter, and which first gave to me the promise and the certainty that this was the girl for me is not constructed but taken from the top of the page at which I opened the first English book I took from my shelf to find an example of vernacular syntax. Among other things consider the striking fact that the first relative pronoun (that) may refer to the immediate predecessor 'mother' whereas the next (which) clearly points farther back, to 'the directness . . .'. Moreover experience rather than syntax tells you the referent of 'this'. In oral expression little boys and girls and little boys and giants would show a difference in intonation to account for the divergent structure, which however is not indicated syntactically in the written record. The vernacular can afford to structure by means of the content rather than by syntactical devices because meaningfulness is a strongly selective criterion in all contexts where human behavior plays a part. Expressions, however, like a-\-ba-b, a-\-b(a-b),

(a-\~b)(a-b),

which can be distinguished from each other only by the position of the brackets, are all meaningful though their meanings differ. 7. Another feature of variables in the vernacular is that most of them can also be used as names of their domain. 'Car' is a generic name in 'my car' but it is the name of a genus in 'Cadillacs are cars'. Man is a bound variable in 'this man'; it means mankind in 'the ancestry of man'. There is, however, no 'carkind' in order to say 'Cadillacs belong to carkind'. As far as mathematics is expressed in the vernacular, we still cultivate these confusing ambiguities. 'A topological space' is an element of the category 'Topological Space'. If an article begins with the definition that R is a connected regular topological space with a countable basis, then subsequent objects like R can be referred to as Rs, which means that the variable R is used as the name of its domain. Formalist minded mathematicians, today, will avoid such bad habits, which thirty years ago were still considered good usage. The double status of variables in the vernacular is closely connected to some ambiguities of the verb 'to be'. Its occurrence in Socrates is a man and Apes are mammals are formalized in mathematics by the symbols g and c:, the first meaning the

COSMIC LANGUAGE

1025

adherence of an element to a set, while the second is used for the inclusion of one set in another. The status of 'to be' in Amsterdam is the capital of the Netherlands is again different; here it indicates equality, which in mathematics is formalized by '='. In It is hot and There is a cat in the doorway it indicates existence, which in formalized language is dealt with as a quantifier. 8. For a century set theory has pervaded all of mathematics and determined its structure from its base to its most detailed branchings. Mathematical language, whether lightly or strongly formalized, has duly accounted for this fact, although often set theory is a linguistic rather than a material means. Set theory terminology in the vernacular is an extremely artificial element; all set theory terms had to be created by assigning to expressions of the vernacular a set theory meaning. It is most characteristic that now in the various natural languages the concept of set is named by terms with a wide range of vernacular meanings. Of course set theory models are not as powerful or as largely useful as might be believed because of their generality; a genus or species is more than a set and 'part' may be a different concept from 'subset', but even where set theory applies, the vernacular expression greatly differs from the set theory language saying that something is an element of a set or a set is a subset of another set. The most striking example is the way the vernacular states the cardinal of a set. The fact that number is a property of sets rather than of individuals had to be ascertained by a profound analysis because it is hidden by the vernacular expression. It is true 'the number of eggs in the basket' adequately describes the situation that this number is a property of a set of eggs, but the attributive position of cardinals and quantifiers in 'ten eggs', 'many eggs', 'all eggs', 'some eggs', creates a seeming analogy to 'fresh eggs' as though an egg besides being an egg and fresh, could also be ten, many, all, and some. To the mathematical analysis ten-ness, many-ness, all-ness, some-ness are not — unlike freshness — properties of eggs but rather properties of sets of eggs. 9. I could multiply indefinitely these examples, which are intended to show how fundamentally the syntax of mathematical language, even in a moderately formalized state, differs from that of vernacular, including such regular vernaculars as Chinese and artificial languages as Esperanto. Non-mathematicians usually are not acquainted with this fact, and even many mathematicians are not conscious of it.

1026

H. FREUDENTHAL

Every mathematician, if made aware, would agree that mathematics can be taught in every vernacular, as well as taught or at least understood without any vernacular. Mathematical language in its most formalized version does not depend on any vernacular, and by its syntactical means it fundamentally differs from all vernaculars. Linguists often deny this. I quite often read or hear the opinion uttered that mathematical languages were a code grafted upon Western languages — whatever this may include — and that there is a difference in principle between groups of natural languages whereas mathematical languages are closely related to the socalled Western languages. Apart from the different spirit of other languages of the manner of thinking of their users, the strongest arguments adduced to support this thesis are such as the different kinds of numerals that exist in some languages to count different kind of objects, or the distribution of nouns over a large number of genera, or the existence of a great number of genera of modalities of the verb, or a great number of cases of the noun, or a strange order of the words in a sentence, or the dependence of the form of speech on the rank of the person addressed. People expressing such opinions generally are not well acquainted with mathematical language or are at most able to read such texts in a vernacular interlinear version. If they see such an expression as 'there is an x such that x . . . ' they forget that this is just an English version of a fundamental term of mathematical language, which in other vernaculars could read quite differently, and that the use and understanding of such an expression does not presuppose anything with respect to the knowledge of terms like 'there', 'is', 'an', 'such', 'that', or that of syntactical sorts and devices such as adverbs, copulae, articles, conjunctions and so on. They behave as though they read a Chinese text in an interlinear English version and conclude that Chinese is nothing but a strange kind of English. It is strange that people who by their very profession would mistrust every translation of a text as long as its original is available and who would object to replacing the 'inch' in 'every inch a king' by '2.54 cm', believe they can understand and judge mathematical text by interlinear versions and explanatory glossaries. 10. Mathematical language as described till now is rather a syntactical pattern, which needs a vocabulary to be filled in. The lexicographic needs of mathematics, however, are modest. Its fundamental vocabulary does not contain much more than the names of the natural numbers, which, though an infinity, are brought back algorithmically to a finite number of principles. It was Peano's desire to create a general purpose language. If this is to be satisfied, then, given that the general syntactic means in mathematics can be put to wider use, a general purpose vocabulary should be added to the inventory; of course if new syntactical problems arise with respect to the use of the terms of the vocabulary, they should be solved within the frame of the general syntax. Creating a vocabulary should be considered a serious problem. If the formalism of mathematical language is accepted as a stringent syntactical pattern, it would be

COSMIC LANGUAGE

1027

ridiculous to fill it out by a lexicographical activity which simply replaces with new linguistic terms those of a natural language. As far as I know, little analysis has preceded the lexicographical creations of artificial languages; in fact, all of them have preserved the features of homonymy that are common to the Western languages, and even quite a lot of idiomatic homonymies. I would not deny that some of these languages may be excellent means of communication between people who are accustomed to our customary languages, but this is not what I would hope for when filling the syntactical pattern of mathematical language with a vocabulary adequate to this syntactical pattern. Along with the syntax the vocabulary should share the virtue of independence from the natural languages. Of course, when creating an artificial vocabulary, it is no easy job to avoid the unconscious influences of the natural vocabularies one is acquainted with. A way to fight these influences is to tie the creation of the vocabulary to a communication problem that fundamentally differs from those which are solved by our usual languages. On the one hand the devices of mathematical language should be put to use in the communication problem (this was Peano's premature idea); on the other hand this communication problem should be so difficult that it cannot be overcome by our normal means of communication. This consideration led to formulating the problem of creating a language to communicate with partners that do not know any of our usual natural or artificial languages and that are not even acquainted with their syntactical structures. In order to exclude merely formal linguistic systems that can be handled by formal criteria, it is urged that the language be used as a vehicle of more than formal, that is realistic knowledge. Communicating with partners that are not acquainted with any of our usual languages is, however, an everyday activity we perform whenever we are dealing with little children who still have to learn their mother language. The device by which in this communicative process the linguistic symbols are connected to the realistic data is showing. To exclude extra-linguistic means as much as possible, the opportunity to show should be sharply restricted. On the basis of these principles, I set out to construct a language. More concretely, I had in mind a language with which to communicate with inhabitants of distant worlds, which, of course, must be supposed humans or humanoid as to their mental states and experiences, though not necessarily with regard to their anatomy and physiology. A modest part of this language, called Lincos (lingua cosmica) has been constructed and published.1 The messages of this language should travel on radio waves of varying frequency and duration, but this basic assumption is not truly essential. My publication (1960) is a printed version of Lincos where instead of the Lincos words as their printed representatives symbols or series of symbols occur, as available in the letter 1 H. Freudenthal, Lincos: Design of a language for cosmic intercourse, foundations of mathematics (Amsterdam 1960).

I. Studies in logic and

1028

H. FREUDENTHAL

case of a printer of mathematical texts. At this stage of Lincos it is immaterial how the symbols are to be 'pronounced'. Before being transmitted the texts should be transposed into suitable radio sounds. I did not, and will not, discuss technical problems like this and others. Our radio messages can now reach distant planets and nearby stars, but nobody knows how far away intelligent neighbors live. Maybe a language like Lincos has already been constructed somewhere in the universe and messages in that language are already crossing space. Mathematicians can never tell whether their abstract thoughts will ever be applied to concrete situations and so I cannot predict whether Lincos will ever be used for cosmic communication. I believe, however, that it can be a useful tool in linguistic analysis and in the analysis of mental behavior. 11. Lincos is a program connecting the teaching of a language to the communication of relevant facts of scientific knowledge and human behavior. The amount published thus far covers every detail though in every series of messages the number of examples should be sensibly multiplied if the program were actually to be transmitted. Maybe there are more important gaps in the program, but up to now none has been indicated to me. Receivers will try to decipher the program, to understand its language and contents. At least intelligent humans would do so if they listened to such a program from space, and we agreed that our distant partners are mentally humanlike. In any case if we were to receive such a transmission from outer space, we would, without any doubt, recognize that this was a message and not noise. It would be interesting to isolate the criteria by which we would distinguish message from noise; as far as I see information theory does not provide the clues to make this distinction. On the evidence of quite a few, though rather informal tests, I feel entitled to believe that decoding Lincos would be an easy job. That cryptographies are hard to decipher is no counter-argument, because such codes are designed to cover secrets. Lincos, however, as a code, has been constructed with the explicit goal to be easily broken by any intelligent receiver. Of course, we cannot predict how a receiver would actually proceed to understand our messages, but as the reader can verify, every single message and its situation in the contexts is designed to make the meaning of its components clear. There is a great wealth of clues, and this redundancy could easily be multiplied. 12. Since showing should be excluded as much as feasible, the Lincos programme must start with the most abstract subject, which is mathematics. The first messages are simple arithmetic like

. >

< ..

COSMIC LANGUAGE

1029

+ .. = The number n is ideophonetically shown by n peeps; these 'ostensive' names are one of the few opportunities in the Lincos program where a concept is 'shown'. The receiver should learn from the context that n peeps means the number n and the meaning of the words > , < , = , + , - . The next step is to replace the ostensive numerals by algorithmic ones which are more easily handled. There are reasons to prefer the dyadic number system, which now is introduced by a catalogue like . = 1 . . = 10 ... =

11

. . . . = 100 =

101

= 110 and so on. The foregoing is an example of how the language of mathematics is taught to the receiver: the receiver is supposed to know mathematics, to identify the transmitted mathematics with his own, and in this way to understand the syntax and the vocabulary of the message. The mathematics of the message is quite informal, only moderately formalized. It is on this level a non-axiomatic approach to mathematics, taught by examples which compel the receiver to experimental inductions. The first chapter of the Lincos program, called Mathematics, introduces mathematics and its language. Its Lincos version does not differ too much from that usually found in the textbooks, though its degree of formalization is much higher. There are a few noteworthy divergences. One of them is punctuation. I earlier stressed that in the language of mathematical formulae, rather than by relative pronouns and conjunctions, syntactic structure is created by punctuation, and that this is also the principle of the language surrounding mathematical formulae as far as it has been formalized. Russell and Whitehead devised a system of punctuation using separators of different weight, instead of brackets and braces. This system has been adopted and perfected in Lincos. In Lincos punctuation of different weights is expressed by pauses of varying length; in the written code of Lincos the pauses have been expressed by a system of square points. Lincos punctuation can again be considered as an ideophonetic ostensive device. 13. Mathematics has been chosen as a way to teach the general syntax of Lincos; if no ostensive or linguistic teaching vehicle is available, this is very likely to be the only possible approach. In our first chapter demonstration was restricted to intro-

1030

H. FREUDENTHAL

ducing the natural numbers ostensively, and to punctuation. I think these devices are rather unproblematic. In the second chapter, 'Time', ostensiveness is resumed, beginning with certain signs, the so-called time signals. They are even more ideophonetic than the ostensive numerals of Chapter I. Those signals meant numbers, whereas the time signals only mean themselves; they share the property of meaning themselves with some noises produced once in a while in the program. Time signals of varying duration and frequency are transmitted to tell the receiver their duration and frequency, and by this means, let the receiver guess what the Lincos words for duration and frequency mean, and which time unit the sender is using. By time signals one shows not only the time unit but also the order relation of time; at this opportunity words like 'second', 'before', 'after' are introduced. A clock starts ticking in the program, and the receiver is instructed how to read this signal; a system of quoting earlier Lincos program events by means of their time datum is developed. A rather circumstantial analysis was needed to establish the quoting status of variables in a quoted event. We will return to this delicate point when we deal with reported speech. 14. The Chapter 'Time' is an indispensable basis for the next one, entitled 'Behavior'. Rather than by general or quasi-general rules as with mathematics and chronometry, human behavior is staged by generalizable examples. Fictional actors perform fictional actions. Since at this stage the Lincos vocabulary is not sufficient to introduce the bodies of the actors, the only action that can be staged is their communication with one another. Since behavior has to be displayed, actions have to be valuated; it starts crudely with two values such as 'good' and 'bad', which should not be confused with the logical values of 'true' and 'false' introduced earlier. It is astonishing that such words that seem to be beyond the scope of mathematics can be introduced by mathematical examples. The human actions shown in the greater part of the Chapter on 'Behavior' are dialogues. Since no other subject is available, these are dialogues on mathematics: class talks, as it were, between teachers and students. From this starting point the program progresses gradually to more involved texts, and to exhibiting more sophisticated and diversified plays. The spoken matter changes from mathematics into talk, and at the end of the chapter into playing games, which leads to an operative definition of probability. The valuation system exhibited by these actions is more and more refined. Rather than going into details, I will give a few indications as to the vocabulary available at the end of the third chapter. Beside mathematical, logical, and chronometric terms there are quite a few particles like 'but', 'thus', 'because', 'supposing that', 'in the way of', instead of', 'omitting', 'adding'; verbs such as 'saying', 'answering', 'knowing', 'thinking', 'analyzing', 'counting', 'proving', 'seeking', 'finding', 'perceiving', 'understanding'; another group is 'nearly', 'error', 'much', 'little', 'too', 'now', 'short time back', 'long ago'; modalities like

COSMIC LANGUAGE

1031

'necessary', 'possible', 'wishing', 'allowing', 'promising', 'being obliged', 'being allowed', 'it is courteous', 'it is convenient', 'difficult'. 15. The fourth chapter, called 'Space, motion, mass', is mainly about mechanics though it contains substantial sections on behavior. The fundamental notions of mechanics are introduced behavioristically according to the usual learning procedure although afterwards they are embedded in a mathematical system. It starts with explaining differences of place and distance by the delay of messages; the set of places is then mathematically described by Euclidean space. Movements are defined as changes of place; and an explanation is given how movement can be caused. Movements, up to vibrations, are put into a mathematical frame; vibration phenomena as communication devices leads to light, light velocity, and the numerical definition of the length unit by the Rydberg constant. Mass is introduced to measure the difficulty of moving things, and more precisely by the elastic collision law; later on the mass unit is provided by atomic data. Humans, which were merely ghosts as actors in the third chapter's plays, are credited with spatial dimensions and mass. More sophisticated mechanical concepts and laws are communicated; planetary motions, the structure of the solar system, and the neighbourhood within the universe are described. Aggregate states, instruments and processes of measurement are explained, and finally the fundamentals of relativity are introduced. New words have been added to the vocabulary, which attains a few hundred at the end of the fourth chapter. Further chapters on 'Matter', 'Earth', 'Life', and a second chapter on 'Behavior' have been planned but are not yet realized. 16. Designing such a program has again and again required inventing situations by which some concept and word can be introduced. This generally is easier with notions from mathematics and physics than with behavioristic notions, because teaching mathematics and physics is a more conscious activity than teaching behavior; we know better how people learn sciences than how they grasp the meaning of 'saying', 'but', 'good'. On the other hand, since we know that behavioristic terms cannot possibly be put into an axiomatic system (though it has been tried again and again with modalities), we are not likely to try to introduce them more or less axiomatically, whereas our knowledge about the kind of axiomatic state into which we can put mathematics and some parts of physics may seduce us to teach these sciences axiomatically. I have tried to avoid this danger and to approach all new notions of mathematics and physics behavioristically, that is, in the manner a good teacher would use with his students, rather than according to a preconceived, more or less axiomatic system. This means that concepts are not introduced once definitely, but gradually along a sequence of levels. Provisional definitions are reconsidered and refined, with as much repetition as needed, and at the most suitable places. For this reason a program like Lincos is not, and cannot be, too systematic.

1032

H. FREUDENTHAL

17. Many special syntactic problems had to be solved to attain the present state of Lincos. I restrict myself to mentioning two of them. In the Introduction to Lincos I said: 'Hitherto in logistics little attention has been paid to sentence patterns other than the propositional one. The interrogative mode must however be accepted if one considers mathematics not as a stock of true propositions, but as an art of discovering. Moreover language as a means of communication cannot dispense with the interrogative mode.' My statement that logistics neglected all others but the propositional mode has been challenged by E. E. Dawson,2 who quoted as counterexamples two papers, one of which appeared three years after Lincos, and the other 'forthcoming'. I would gladly withdraw my statement if among the many thousands of titles on logistics before 1960, Dawson could have uncovered, let us say, ten that contributed to the subject of interrogative sentences; post-1960 references do not convince me that I was wrong then. Even now, ten years later, a look at the literature shows that my statement is still valid. I know that linguistic habits in mathematics change slowly so I am not astonished that my notation for interrogative binding — the only one ever proposed in mathematics — has been adopted only by a small minority. My analysis showed that the interrogative mode can be formalized by a certain way of binding a variable, which I called interrogative binding. It is indicated by a particle '?'. A pattern like

should be read asked for an x with the property . . . This particle, indeed, suffices to form all interrogative pronouns and adverbs, as well as the simple factual interrogative mode. For instance, a sentence pattern such as at the time t, on the place I, x killed y with z because of u gives rise to the interrogative particles It 11 Ix ly 1z ?«

: : : : : :

when where who whom by which means why

Even the factual question can be formulated by this device: w is the truth value of p * See fn. 6, p. 1038.

COSMIC LANGUAGE

1033

gives rise to the particle ?w : whether. The identity question (who is Caesar? what is an integer?) is somewhat more difficult; it requires a lengthy analysis, which can be found in Lincos 3022-3. Another problem is pseudo-interrogative speech (I told him who was coming, what has happened, and so on), which was satisfactorily tackled in Lincos 3170. The status of 'to know' with respect to the interrogative mode was analyzed in Lincos 3130. In many languages the semantic domain of 'to know' is claimed by two clans of words; the difference can be circumscribed in Latin by 'scire' and 'noscere', in French by 'savoir' and 'connaître', in German by 'wissen' and 'kennen'; the English translation would be 'to know a fact' and 'to know a thing'. Lincos's sticking to one word is legalized by the rule that a sentence depending on the Lincos word for 'to know' is to be constructed as a question. So intermediate translations of sentences to be translated into Lincos would be: 'Ich kenne Spiegelberg' -> Ich weiss: wer ist Spiegelberg?' 'I know the solution of 3x = 9' 'I know: (?*) (3* = 9)' Notice that 'I know that 3 is the solution of 3x = 9' should rather be translated by '3 is the solution of (3*) (3x = 9); and I know: (?*) (3* = 9). Likewise 'I know a Roman citizen' may not be translated by 'I know: who is a Roman citizen?' but by 'There is an x such that x is a Roman citizen and such that I know: is x a Roman citizen?' 18. I will now come back to the linguistic problems of reporting events and quoting speech. Formalist semantics has been saddled with the pseudo-problem of intentionalism. Since 'London' and 'the capital of England' are names of the same object, as are '2 + 2' and '4', they can, according to the so-called extensionalist thesis, be substituted for each other. This works very well in texts like Trafalgar Square is in London

1034

H. FREUDENTHAL

and 1 2 - 4 = 8, but it does not work in He said (knew, asked whether, dreamt, believed) London is the capital of England, 2 + 2 = 4. since He said (knew, dreamt, believed,. ..) London is London, 4 = 4 would obviously mean something different. This 'intentionalist paradox' is a consequence of considering language as a formal system. As long as the communicative character of language is ignored, it makes little sense to consider reporting an activity that should be expressed linguistically; then such a paradox cannot even be formulated. If, however, the communicative character of language is accepted, it cannot be decided on formalist evidence whether one report is good or bad, or better or worse than another report. If a sound or a speech is to be reported, it depends on the circumstances whether the new message, to be trustworthy, should be an exact, or a more or less fashioned, copy. To speak about an object, I need a name for it; if no individual name is available, it may be a generic name (a variable), but generics are often not indicative enough. If the object which I want to deal with is itself a linguistic or at least an acoustic event, it is a habit in everyday language to use a copy of this event instead of a name, or rather as a name. The relation between Did you hear the prt-prt-prt? and Did you hear that noise? is the same as between Did you hear him saying: I love you? and Did you hear his words? The 'prt-prt-prt' as well as the 'I love you' are copies of acoustic events, used, if needed, as names of these events. In formalist semantics autonymity, that is, using a copy of an acoustic (or in written language of a textual) object as a name of this object, leads to paradoxes. In Lincos 2041-50,1 explained that if a language is to serve as a vehicle of com-

COSMIC LANGUAGE

1035

munication, it cannot avoid autonymity, and that the classical devices of formalist semantics to circumvent autonymity are both insufficient and superfluous. I cannot repeat here the circumstantial argumentation that led me to this conclusion. I justified autonymity as an indispensable tool in Lincos syntax. But at the same time I had to clarify the status of variables in autonymous texts as met with in reports and quotations. (See also Lincos 3015.) Natural languages have the pattern of oblique speech: John said to me: I am ill is transformed into John told me that he was ill where the variable T is transformed into the variable 'he'. Or John said to me: Mary is ill is transformed into John told me you were ill (if said to Mary), where the constant 'Mary' is transformed into the variable 'you'. It is an interlinguistic feature that in oblique speech we find variables that are not genuine, but belong to the discourse of the reporter. For Lincos I had to decide how to deal with variables and which characteristics should determine to which discourse a variable belongs. The rule I finally set, was that a variable always belongs to the discourse in which it is bound or, if it is not explicitly bound, as happens in universal binding, the discourse in which it occurs first and that hidden variables such as 'I', 'now', 'here' are considered to belong to the discourse in which they occur. These are rules that have worked very well. 19. The Lincos program addresses itself to humanoid receivers. It has sometimes been doubted whether even this condition would guarantee the decodability of the program. The only skeptical argument that struck me was that of Otto Neugebauer when I gave a talk at Brown University. He said that before hieroglyphs and cuneiform texts were decoded, people firmly believed that all those texts were of a religious kind since there was hardly anything else ancient sages could have bothered to think about. Now imagine your Lincos program received by people in the same state of mind; for such people messages from outer space would be godly revela-

1036

H. FREUDENTHAL

tions. It would be a sacrilege to interpret > .... as mathematics; maybe they would read it 'six gods are mightier than four gods'. I admit Neugebauer may be right. He may be right, but not for long. Hieroglyphs and cuneiform inscriptions were finally deciphered; it was discovered and recognized that besides religion they deal with history, economy, and even mathematics. Being humanoid is, of course, a large presupposition. We do not know what it includes, because it is difficult to analyze. Distant worlds may be inhabited by intelligent beings that are only partially humanoid. The Lincos program starts with mathematics as a universal ability of intelligent beings. In the third chapter separate individuals occupy the stage; in an otherwise intelligent world which consists of one single individual, this would be hard to understand. The actors of the Lincos program converse with each other, which would be unintelligible in a world of many persons with one consciousness. In the fourth chapter the human body is described; maybe elsewhere all have one body though separate minds. Lying and gambling, which is not unusual in the Lincos program, might be entirely unknown elsewhere in the universe. At many points in the Lincos program one could raise the question: which quality would enable a being to understand this point? All those qualities together characterize the being that understands the total Lincos program, that is, man. Maybe an undertaking such as Lincos can help us to better understand what makes man a man. 20. When, after ten years, I survey the reactions to Lincos, I am a bit disappointed. Almost all reviews were highly favorable, though I would have preferred to learn from expert criticism, in particular with respect to the many details of the program and its syntactic devices. On the other hand I am glad that my ideas have in general been well understood. However, in public lectures on Lincos, I was sometimes struck by a lack of abstractive power, which seems unbelievable, at least to the mind of a mathematician. I mean situations like the following. Early in the Lincos program I introduced the word which in the vernacular means 'less or equal'. Somewhat later the particle 'v' meaning 'or' entered. At this moment a listener interrupted me: There is something wrong; you used the 'or' already in 'less or equal'. The interrupter could not grasp 'less or equal' as one Lincos word; to him it made sense only as a combination of three meaningful components, deriving its meaning from each partial meanings. Most misunderstandings with regard to Lincos are of this 'interlinear version'. I will deal with two reactions to Lincos in more detail, the first objective though skeptical, the second betraying irritation and apparently intentional misunderstanding. 21. R. M. W. Dixon's review may be characteristic of a linguist's attitude, viewing

COSMIC LANGUAGE

1037

an attempt such as Lincos as a transgression. . . a linguist has to be keenly aware of the current opinions of philosophers, literary critics, mathematicians and anthropologists. However much he deplores these opinions, the fact that they are at present many times better known and understood than his own forces him to argue with respect to them, if he is to succeed in changing them.' It does not matter whether such opinions might be right or even whether it might pay to learn from them; they are deplorable and must be changed.3 'From the way Lincos is constructed it is plain that aliens are expected not only to be aware of the same universal truths as us, but to express them in more-or-less the same way.' The first is true, the second is plainly wrong. It is the mistake of the interlinear version. The language, the formal presentation, and even the basic concepts of their mathematics, chronometry, and mechanics could be entirely different from ours (as were the language, formal presentation, and basic concepts of mathematics three centuries ago); and nevertheless those aliens would be able to equate the two systems (as we can map old mathematics adequately into ours). 'In common with many mathematicians Freudenthal fails to realise that mathematical formulae are abstractions from Indo-European syntactic patterns, and that they can only be explained to, and understood by, someone who is familiar with one of a certain set of natural languages.' Every mathematician will absolutely deny this. I do not understand what Dixon means here by 'abstraction'; mathematicians have quite different ideas of what an abstraction is. He probably means no more than that mathematical formulae have arisen from Indo-European syntactic patterns. This may be partially true, but it is irrelevant. It is mistaken historicism. 'Indo-European' means a family in the sense of descendence, not of similarity of syntactic patterns. 'Indo-European' fortunately covers the greater part of languages in which mathematics has been produced, but it excludes the first mathematicians, who spoke Sumerian, which looks like Lincos, and it excludes the Arabic influence on the development of algebraic language, the Japanese and Chinese who created a mathematics like ours though less developed, and the Hungarians who contributed greatly to mathematical style. Reasoning like Dixon, one should also conclude that since French arose from Latin, it can only be explained to and understood by people who know Latin, and since Latin is an Indo-European language, you have first to study Sanskrit, which was long believed to be closer to the original Indo-European idiom. In fact, such historicisms have for a long time dominated linguistics and influenced teaching languages. Though they would deny it, there are still people who believe in word magic, that is, that things can be shaped by language, and that to shape things you have to speak some language. 'Thus the world view of modern science arises by higher specialization of the basic grammar of the Western Indo-European languages', to cite Whorf.4 It 9 4

Linguistics 5.116-18 (1965). B. L. Whorf, Language, thought and reality (Cambridge, Mass., 1956).

1038

H. FREUDENTHAL

is the 'post-hoc — propter hoc' of the historist. But even historically, science and language were never magically interrelated. Language is an easily adaptable vehicle. Primitives, who counted up to two or three, develop a far reaching number system as soon as they make commercial contacts. In all languages zero and fractions were created as soon as they were needed. Mathematical language has developed through a continuous adaptation of carrier languages to becoming a vehicle that has wholly emancipated itself from its origin. Understanding mathematical language only in an interlinear version prevents linguists from grasping the fundamental syntactical divergence between the so-called Indo-European languages and mathematical language. This leads to opinions on mathematical language like the above, which to mathematicians look as absurd as would be opinions on Chinese expressed by English speaking people who know Chinese only in an English interlinear version.5 22. Never has anybody tried to misunderstand me as obstinately as has E. E. Dawson.® It is a pity that I could not answer his criticism earlier. Nobody likes to be seen through a distorting mirror.7 In Lincos I dealt with a communication problem. This is clear from the subtitle; it was clearly stated in the introduction and for a while it should have been clear to Dawson, since even he quotes my purpose 'to design a language that can be understood by a person not acquainted with any of our natural languages or even of their syntactic structures'. But if Dawson ever realized what this meant, it was soon forgotten. There is no trace that he ever considered Lincos in this regard. He time and again deals with Lincos as an attempt to design a fully formalized language. 'Freudenthal has at least partially fallen into the elementary error of supposing that one achieves formalization and rigour simply by introducing a symbolic vocabulary. . . . This attitude seems particularly bad in a work of a mathematician, since mathematics is the paradigm example of formalization achieved by having rules giving a rigorous syntax.' According to Dawson, because mathematics can be fully formalized, a mathematician should deal with mathematics only in a fully formalized state. I do not know how Dawson arrived at this rule of behavior he is setting up for mathematicians, and suspect his credentials. There are very few mathematicians who have chosen mathematical language as a subject matter. The greater part use it to communicate mathematics to other mathematicians or to students. This means that in every single case the state of formalization has to be adapted to that specific situa5 W. A. Verloren van Themaat, La Monda Lingvo-Problemo 1.167-82 (1969), analyzed excellently this attitude of linguists. e Mind 73.299-302. I came across Dawson's review when I set out to write the present paper; no copy of the review has been received by the publisher of Lincos. 7 This is how I seem to have been seen by L. Aqvist who in the bibliography of A new approach to the logical theory of interrogatives, part 1 (Uppsala 1965), in lieu of Lincos, quoted Dawson's review, in order to wholly dismiss Lincos as irrelevant.

COSMIC LANGUAGE

1039

tion. In kindergarten and elementary school formalization will be restricted to a few number symbols, and maybe, a few operation symbols; it will include a few symbols for algebraic variables in the higher grades of the primary and lower grades of the secondary school; as time goes on, reasoning will be formalized to a certain degree. It is true there are university textbooks that stick to the same degree of formalization from the first to the last page, but they are belied by good courses that show a gradual strengthening of the formalist attitude. It is true there are textbooks that confront 11-12 years olds with a highly formalized geometry and even with definitions of a definition, a theorem, a proof; there are university teachers who start with categories because this is the most general subject matter. None of them goes as far as Dawson who seems to believe that you can communicate with somebody who does not even know your language in a fully formalized discourse, and that as a mathematician you are not allowed to compromise in this difficult communication problem the state of formalization of your discourse which, according to Dawson, should in cosmic communication be higher than has ever been assumed in terrestrial mathematical communication. Dawson failed to remark that there are different levels of formalization and that in every single case you have to adopt the one that is most adaptable to the particular communication problem; if there is no communication problem, if nothing has to be communicated in the language, you can choose full formalization. How far Dawson is from grasping what a communication problem is is shown by the following example. He quotes the remark at the end of Lincos 3241, 'Imperceptible speaking (thinking)' and adds 'he assumes a Watsonian view of thinking'. I said in the Introduction I espoused no philosophy, behaviorism included. This should have been a warning. Lincos's vocabulary thus far contains no words for thinking, mind, soul, love, God, and so on, because I would not know how to introduce them. Clearly, rather than the text 3241, Dawson only read its free translation at the end. The truth about it is dead simple: Ha communicates to Hb four time signals and asks him to tell how many there were. Hb answers: 4. Ha asks Hb how he knew it. Hb says he counted. Ha objects he observed him but did not hear him counting. Hb explains that he counted in a way that Ha could not observe. This is the whole story, I think its meaning is perfectly clear. From Ha and Hb's behavior the receiver of the Lincos program can learn that a human can utter silently to himself words that others cannot perceive. This is an important feature which is often repeated in the Lincos program. Terrestrial receivers would call it thinking. I did not define thinking, and if I were to do so, I would wait for a situation where I can make clear to the receiver that there also exists non-linguistically-carried thinking. Nevertheless, if a small child would ask me what thinking is, I would neither translate it by 'frowning one's eyebrows', nor by quoting Hegel, but by telling the same kind of story as I did in Lincos 3241. This, rather than by Hegelian or Watsonian philosophies, is the way people deal with communication problems.

1040

H. FREUDENTHAL

I am glad Dawson's review provides laughing matters as well. 'And there is a fairly elaborate message (4101) for "whistling one's dog"! One might indeed suppose that there are better uses than this for "a language for cosmic intercourse", but perhaps the inclusion of such instructions as how to utter "one's wishes in a more circumstantial manner" or "how to whistle for one's dog" are justified by the hope that Lincos will "be continued so as to cover the whole of human experience".' Well, there are neither whistles nor dogs in the Lincos program. 'Whistling one's dog' is a tongue-in-cheek translation. It was not meant as a trap; it is too transparent to be misleading. But I did not realise that irate people are caught in any trap, however obvious it may be. I will not explain this very short story — it is not at all elaborate. Simply by looking at this text, Dawson could have easily ascertained that it has nothing to do with 'circumstantial speaking' or 'covering the whole of human experience', but that it had to prepare the introduction in 4111 of a word, meaning 'concrete things' (in contrast to living beings); criteria are formulated through an elaborate complex of stories around 'moving (neutre)' and 'moving (active)'. According to Dawson I am wrong to use words like 'to mean', 'to designate', 'necessary', 'possible' as they are used in the vernacular. In fact, I have not introduced Lincos words for 'to mean' and 'to designate', but I explained circumstantially why in the vernacular I did not follow formalist semantics at this point: the reason is that I could not and that nobody, Dawson included, can, as long as he uses language as a communication medium. I am sure, however, that if I had adopted that odd meaning of 'meaning', Dawson would triumphantly have declared me a Fregeian, though I refused to espouse any philosophy. 'Necessary' and 'possible' somehow occur in Lincos; I explained that I could not use the formalist approach to modality, again for the reason that nobody, Dawson included, can as soon as he uses language as a communication medium. Dawson seems to praise Lincos, when he states 'perhaps the most interesting philosophical problem involved in the construction of Lincos is the expressibility of syntactical properties of the language in the language.' Since it is the only praise Lincos got from Dawson, it is a pity that I have to admit honestly that Lincos did not merit it. The only syntactic theory developed on Lincos is metaprogrammatical. I even stressed that it is much too early to write a syntax of Lincos in Lincos. Dawson's statement proves a gross misunderstanding of my elaborate discussion of reporting events and quoting speech. The Lincos actors can speak about speech (Ha says that Hb says that. . .); they can even use syntactic terms like 'his saying', 'proposition', and values with respect to such terms as 'good' and 'bad'. But they never speak about syntax, or syntactic properties. They use the Lincos syntax as every language user uses the syntax of his language. Dawson claims: 'Freudenthal varies between these two courses, sometimes showing the metalinguistic properties by examples, and sometimes stating them in the language. Thus the syntactical properties of "Sci" ("knows"), "Enu" ("counts"), "Ani" ("perceives") and so on,

COSMIC LANGUAGE

1041

are shown ( . . . "a certain behaviour that will be exhibited by means of examples"), while the properties of "Ver", "Fal", etc. are expressed in the language ("We now introduce a few words that will enable the user of Lincos to make statements in Lincos about some kind of Lincos expressions": 1 + 1 = 10. e Ver, a + b b + a e Fal are examples).' I am not responsible for this mess. First, Dawson vacillates between 'metalinguistic' and 'syntactical'. There is a lot of metalinguistics in Lincos (Ha says that Hb says t h a t . . . ) but syntax has till now neither been shown nor expressed in Lincos. Second, to prove my inconsequence, Dawson claims that I used another policy in the case of 'Ver' and 'Fal' than in those of 'Sci', 'Enu', 'Ani', and to prove this latter fact, he quotes a place where I explicitly state the same policy. I have kept Dawson's grossest derailment for the end of this evaluation. The reader may recall that I built written Lincos in a rather arbitrary way without caring about how transmitted Lincos would finally be pronounced. According to Dawson, pronunciation is so essential that since I omitted it, it would have been better to have built Lincos from English words: 'why not just put the written Lincos down in English? The present procedure renders the book well-nigh impossible to read (the "Register at the end contains some 304 entries . ..").' Well, since we do not know how Homer pronounced Greek, stop reading the original. Whatever you think about the conclusion, you must agree that the argument at least is irrelevant. Moreover, according to Dawson I should have written 'less or equal' instead of ' Ï 20 10

14 15;

15

50

20 10

m

:i6

70

:i5

9i

m

10

II

'AV,

I I

Ü 4- /7 V ^ l

2

2

i 3 3

4- /7 %-3/i,

m

1 /4-%

4

Fig. 3. The proportion by age and sex of correct responses to reversible passive sentences (e.g. 41a). Each child received only one reversible passive sentence. (The numbers in each bar indicate the number of children interviewed at that age. The data is broken down by 4-month age groups for purposes of the discussion below.)

cannot understand. Thus, the basic linguistic capacity evidenced by the two-yearold child includes the notion of reference for objects and actions, the notion of basic functional internal relations and at least a primitive notion that there are different sentence structures. Beilin and Spontak (1969) used the emergence of the capacity to recognize the relationship between the active and the passive construction as a measure of the development of the child's capacity to produce linguistic intuitions. (I should emphasize that while the facts are due to Beilin, the interpretation is not necessarily his.) Beilin shows that the child does not appear capable of appreciating the regularity of the relationship between active and passive sentences until about age 7-8, which is also the age at which the child is alleged to have developed the integer concept. Indeed it is support for my claim that the adult numerical and linguistic phenomenologies are the same type of cognitive phenomena that Beilin finds a correlation between the child's ability to deal correctly with numerical and linguistic transformations (e.g. recognizing that changing the array in 2a to that of 2b doesn't change the number of circles in any of the rows, or recognizing the relation of the active and passive). As in the case of the development of the child's capacity to make judgments of relative numerosity, the linguistic behavior between the ages of 2 and 6 displays

1184

T. G. BEVER

a period of relative dependence on perceptual generalizations. Consider first the early appearance of the basis for Strategy B, that the first N . . . V . . . (N) sequence is the main clause. In a recent study we have asked young children to act out sentences like 42a. Presumably because of memory limitations children often 42

a) The cow that jumped, walked away, b) The cow jumped and walked away.

act out only one of the two clauses of such sentences. Which clauses they act out gives us a measure of which clauses they consider the most important when they hear it. Our results show that children between and 2\ who perform poorly on acting out both actions in 42b act out only the first action (the subordinate verb) in 42a; children who do well on sentences in 42b act out the second action (the main verb) in 42a. That is, children at the beginning of language comprehension pick the first 'N. . . V' sequence as the most important part of a sentence; they follow Strategy B completely; more advanced children learn to discriminate the main verb from the subordinate verb and consider the main verb to be the most important action. Consider now the development of the basis for a semantic strategy like Strategy C, involving probabilistic constraints. We examined the development of this in the course of the same experiment outlined above by including simple active sentences which either followed 43a or did not follow 43b probabilistic constraints. Figure 4 shows the relative sensitivity to the semantic constraints at each age (i.e. the percent correct performance on sentences like 43b subtracted from the percent 43

a) The mother pats the dog. b) The dog pats the mother.

correct on sentences like 43a — a large number indicates a high dependence on semantic constraints). Figure 4 shows that this dependence undergoes a marked increase during the third year. These results were found initially with only 2 sentences of each type, but have been replicated in a second experiment with five sentences of each type (Figure 5). A similar result is reported by Sinclair and Broukhardt (1972). These experiments show that the two-year-old child is relatively unaffected by semantic probabilities. The implication of this is to invalidate any theory of early language development which assumes that the young child depends on contextual knowledge of the world to tell him what sentences mean, independent of their structure. It is obvious why the very young child cannot make use of contextual probabilities: he does not have enough relevant experience to know what the probabilities are. For example, the young child may know the meaning of the word 'pat' but may not have heard it enough, or done it enough, to know that usually people pat dogs and not the reverse. Thus, it is not until the third year of life that the kind of contextual probabilities which provide the basis for Strategy C in adult perception develop as guide for sentence comprehension.

PERCEPTION AND LINGUISTIC

1185

STRUCTURES

100

80 QJ tn c o S" 60 t/>

o

í: 4 0 o o 20 0 I

I 83

I 79

I 77

1 24

1

Number

Fig. 4. The proportion by age of correct responses to probable and improbable sentences, in which each child received one sentence of each kind (roughly half the children at each age received one probable and one improbable active sentence and roughly one-half received one probable and one improbable passive sentence). The bottom line represents for each age the difference in correct responses between the probable and improbable sentences, and thus is a measure of the children's dependence on probability as an aid to correct performance.

T

% 100

1

1

1

1

1

r

80 tn 0)

i/i § 60 o. t/>

0)

I 40

Probable - improbable

o o 20

(2.0-2.7) (2.8-3.3) (3.4-3.7) (3.8-3.11)14.0-4.3) (4.4-4.7) (48-5.7) i

i 25

i 34

i 32

Age

i 34

i 35

i 25

i 43

i

Number

Fig. 5. The same data display as in Fig. 4 except from a different experiment with different children in which each child received three probable sentences (balanced within each age group so that the same number of actives and passives of each kind were responded to).

1186

T. G. BEVER

Shortly after this development the child goes through a phase in which he depends relatively heavily on something like Strategy D for the comprehension of sentences which do not have semantic constraints. This is brought out by his performance on acting out passive sentences like those in 35 (see Figure 3). The most important feature of these results is the steady increase in performance until age 3/8 for girls and 4/0 for boys when there is a sharp (temporary) drop in performance. These results were obtained with only four sentences (of which each child acted out only one) so a larger experiment was run (again by a different experimenter and in a different city) in which twelve reversible passive sentences were used (of which each child acted out three). The results for the passive sentences in this group are presented in Figure 6. Again the same brief decrease appears at the same ages (although in these materials, the decrease starts at the same time for boys and girls, but lasts to a later age in boys than in girls). Finally, in a separate experiment we have studied the performance of the child on cleft-sentence con-

Fig. 6. The proportion of correct performance by age and sex to reversible active and passive sentences in which each child acted out three sentences of each kind. The bottom line in each graph represents the difference between the performance on actives and passives and thus is a measure of the children's dependence on the 'actor-action-object' order as an aid to correct performance.

PERCEPTION AND LINGUISTIC STRUCTURES

1187

structions which can reverse the subject object relation without reversing subject verb order, as in 44a, b. 44

a) It's the cow that kisses the horse (actor first) b) It's the horse that the cow kisses (object first)

Figure 7 presents the tendency to perform correctly on sentences like 44b. Again, the same decrease in performance appears at about age 4. % 100

50

0) to

tO

c o

oa>