To honor Roman Jakobson : essays on the occasion of his 70. birthday, 11. October 1966: Vol. 1 [Reprint 2018 ed.] 9783111604763, 9783111229584

Citation preview

TO HONOR ROMAN JAKOBSON

JANUA LINGUARUM STUDIA MEMORIAE N I C O L A I VAN WIJK DEDICATA

SERIES

MAIOR

XXXI

1967

MOUTON THE HAGUE • PARIS

ESSAYS ON T H E O C C A S I O N OF HIS S E V E N T I E T H BIRTHDAY

11 October 1966

VOLUME I

1967

MOUTON THE H A G U E • PARIS

© Copyright 1967 in The Netherlands. Mouton & Co. N.V., Publishers, The Hague. No part of this book may be translated or reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publishers.

LIBRARY OF C O N G R E S S CATALOG CARD N U M B E R 67-29070

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

T A B L E OF C O N T E N T S

A Bibliography of the Publications of Roman Jakobson

XI

ROBERT ABERNATHY

Rhymes, Non-rhymes, and Antirhyme

1

V. ADRIANOVA-PERETC 3aMeTKH o jieKCHKe CAO 6a o no my Hzopeee

15

EMILIO ALARCOS-LLORACH

El artículo en español

18

R. M. ALIEV SBOJHOUHH CTHXOCJIOaceHHH B nepCH^CKOH II033HH

25

W. S Y D N E Y ALLEN

Correlations of Tone and Stress in Ancient Greek

46

NIKOLAJ A N D R E E V J l H T e p a T y p a H HKOHOITHC£>: K HCTOPHH H f l e i í B MOCKOBCKOH

PycH .

.

63

The Grammatical Categories of the Indicative in the Contemporary Bulgarian Literary Language

82

HOWARD I. A R O N S O N

ROBERT AUSTERLITZ

Two Gilyak Song-texts

99

ROBERT AUTY

The Gospel and Psalter of Cherson: Syriac or Russian?

.

.

.

.

114

R. I. AVANESOV K HCTOpHH MOpcfjOHOJIOriIHeCKHX lepeflOBaHHH H (J)OHeMHOrO COCTaBa KopHeBMX MopcjjeM npH 06pa30BaHHH yMeHbiiiHTejitHLix cymecTBHTejibHMX B pyCCKOM H3bIKe

118

VI

TABLE OF CONTENTS

ANDREI AVRAM

De la langue qu'on parle aux enfants roumains

133

O. S. AXMANOVA HeKOToptie OCOGCHHOCTH rjiarojibHoö rnnepjieKceMbi B pyccKOM H3biKe

141

ROLAND BARTHES

Proust et les noms

150

C. E. BAZELL

On the Declension of Bats Substantives

159

IRENA BELLERT

A Semantic Approach to Grammar Construction

165

H R E I N N BENEDIKTSSON

The Pro to-Germanic Vowel System

174

E. BENVENISTE

Les relations lexicales slavo-iraniennes

197

KNUT BERGSLAND

The Eskimo Shibboleth inukfyuk

203

v . BESEVLIEV

Protobulgarica

222

H. H. BIELFELDT

Russisch Kycmdpb, xycmapA :

Kycm

wie Deutsch Stümper : Stumpf .

.

232

M A N F R E D BIERWISCH

Syntactic Features in Morphology: General Problems of So-called Pronominal Inflection in German

239

HENRIK BIRNBAUM

Predication and the Russian Infinitive

271

H. BLANC

The 'Sonorous' vs. 'Muffled' Distinction in Old Arabic Phonology .

.

295

.

309

MORTON W. BLOOMFIELD

The Syncategorematic in Poetry: From Semantics to Syntactics

.

TABLE OF CONTENTS

VII

P. G. BOGATYREV HMnpoBH3aima

noBecTeii

XVIII

h

hopmh

xyaoacecTBeHHwx

npneMOB

Ha

MaTepnajie

b., Ha^nHceH Ha jryöoHHMx KaptHHKax, CKa30K, h neceH

o EpeMe h $ 0 M e

318

DWIGHT BOLINGER

The Imperative in English

335

G. B O N F A N T E

Italia e Grecia

363

J. B R O N O W S K I

Human and Animal Languages

374

MARIA Z A G O R S K A B R O O K S

The Accusative-Genitive Contrast in Some Polish Constructions

.

.

395

E D W A R D J. B R O W N

Eugene Zamjatin as a Critic

402

D A L I B O R BROZOVKi;

Some Remarks on Distinctive Features Especially in Standard Serbocroatian

412

J. S. B R U N E R

The Ontogenesis of Symbols

427

Y U E N R E N CHAO

Chemical Analogies in Chinese Grammatical Structure

.

.

.

.

447

S U N I T I K U M A R CHATTERJI

Some Indo-European Tribal Names: Loans and Inheritances .

.

.

452

"But There Is Nothing I Have Is Essential To Me" (or "The Human Race Is Not A Club")

462

COLIN C H E R R Y

JANUSZ CHMIELEWSKI

Linguistic Structure and Two-valued Logic: The Case of Chinese .

475

P I E R R E COLACLIDfcS

Sur la forme d'un passage de Denys l'Areopagite

483

VIII

TABLE OF CONTENTS

E. COSERIU

Zur Vorgeschichte der strukturellen Semantik: Heyses Analyse des Wortfeldes 'Schall'

489

FRANTISEK D A N E S

Order of Elements and Sentence Intonation

499

PETAR DINEKOV

À travers les échanges littéraires bulgaro-polonais

513

STOYAN DJOUDJEFF

Esquisse d'une méthode musicologique pour l'étude des vers populaires

523

LUBOMÎR DOLEËEL

The Typology of the Narrator: Point of View in Fiction .

.

.

.

541

WITOLD DOROSZEWSKI

Sur quelques transformations structurales de mots en swahili .

.

.

553

LUBOMÎR ÖUROVLC

Das Problem der Morphonologie

556

FRANCIS DVORNIK

The Embassies of Constantine-Cyril and Photius to the Arabs .

.

.

569

C. L. EßELING

Historical Laws of Slavic Accentuation

577

WILLIAM B. EDGERTON

The Poetics of Periodical Names (from Aurora to Zar ja) .

.

.

.

594

VRATISLAV E F F E N B E R G E R

Journey to the Center of a Poem

615

VICTOR ERLICH

A Note on the Grotesque. Gogol: A Test Case

630

G U N N AR FANT

The Nature of Distinctive Features

634

C. A. F E R G U S O N

St. Stefan of Perm and Applied Linguistics

643

TABLE OF CONTENTS

IX

JAMES FERRELL

On the Prehistory of the Locative Singular of the Common Slavic Consonant Stems

654

RUDOLF FILIPOVlC

Compromise Replica and Phonemic Importation ELI

662

FISCHER-J0RGENSEN

Perceptual Dimensions of Vowels

667

A. V. F L O R O V S K I I IIAPEBHH

AjieKceii RIETPOBHH

B

1710 ROAY: 3anncKa Tp. T. B. Bejiiexa .

MEYER FORTES

Tallensi Riddles

672 678

HENRI FREI

Quasi-phrases et phrases-poteaux

688

L. G Â L D I

Le mètre et ses variantes typiques

692

A R N E GALLIS

Zu Syntax und Stil der gegenwärtigen russischen Zeitungssprache .

.

697

Kartvelian and Indo-European: A Typological Comparison of Reconstructed Linguistic Systems

707

T H O M A S V. G A M K R E L I D Z E

P A U L L. G A R V I N

Heuristic Syntax in Russian-English Machine Translation .

.

.

.

718

EMIL GEORGIEV B03HHKH0BeHHe

CTapoÖOJirapCKOH

(CTapOCJiaBHHCKOH)

CTHXOTBOpHOH

TpaaaiiHH

726

V L A D I M I R I. G E O R G I E V HHFLOEBPONEIICKHH TEPMHH

*deywos B

CJIABHHCKIIX JNTIKAX

.

.

.

734

MARIJA G I M B U T A S

Ancient Slavic Religion : A Synopsis

738

X

TABLE OF CONTENTS

ROBERT GODEL

Dorica castra: Sur une figure sonore de la poésie latine

760

ZBIGNIEW GOL4B

The Traces of vrddhi in Slavic

770

K O N R A D GÔRSKI

Quelques traits caractéristiques de la stylisation satirique dans l'œuvre poétique de Mickiewicz en 1832-1834

785

A. G R A U R

La langue de l'avenir

792

A. J. G R E I M A S

L'écriture cruciverbiste

799

M A R Y R. H A A S

Roger William's Sound Shift: A Study in Algonkian

816

A BIBLIOGRAPHY OF THE PUBLICATIONS OF ROMAN JAKOBSON

I. BOOKS AND PAPERS 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

"N. N. Durnovo, N. N. Sokolov, D. N. Usakov, Opyt dialektologiceskoj karty russkago jazyka v Evropé," Étnograficeskoe Obozrénie, CIX-CX, 1916, 102-107. Novejsaja russkaja poézija, Prague, 1921, 68 pp.; a part appeared in Slovak translation in Teória literatúry, edited by M. Bakos, Trnava, 1941, 63-73. Vliv revoluce na rusky jazyk (reprint with corrections and additions from Nové Atheneum, III, 1921, 110-114, 200-212, 250-255, 310-318), Prague, 1921, 32 pp. "O realismu v uméní," £ erven, IV, 1921, 300-304. "Brjusovskaja stixologija i nauka o stixe," Naucnye Izvestija, II, Moscow, 1922, 222-240. "Zamétka o drevne-bolgarskom stixoslozenii," Izvestija Otd. rus. jaz. i slov. Ross. Akad. Nauk, XXIV, No. 2, 1922, 351-358. [with P. Bogatyrev] Slavjanskaja filologija v Rossii za gody vojny i revoljucii, Berlin, 1923, 63 pp. (originally in Slavia, I, 1922, 171-184, 457-469, 626-634). "Nékolik zpráv o práci v oboru slovanské filologie na dnesní Ukrajiné," Slavia, I, 1922, 634-636. O cesskom stixe preimuscestvenno v sopostavlenii s russkim (= Sborniki po teorii poéticeskogo jazyka, V), Berlin-Moscow, 1923, 120 pp. "Starocesskie stixotvorenija, slozennye odnorifmennymi éetverostisijami," Slavia, III, 1924, 272-315. "O Králové Ceské prosodii," Kritika, II, 1925, 110-114. "F. Trávnícek: Príspévky k nauce o ceském prízvuku," Slavia, IV, 1926, 805-816. Základy ceského verse, Prague, 1926, 140 pp. (revised version of item 9). "Pro realizm u mystectvi," Vaplite, Kharkov, 1927, No. 2, 163-170. Spor duse s télem; O nebezpecném casu smrti (= Národníknihovna, IV), Prague, 1927, 111 pp. Fonetika odnogo severno-velikorusskogo govora s namecajuscejsja perexodnostju (mimeographed), Prague, 1927, 82 pp.

XII

17. 18.

19. 20.

21.

BIBLIOGRAPHY

"O hlàskoslovném zâkonu a teleologickém hlâskoslovi," Casopis pro moderni filologii, XIV, 1928, 183-184. "Quelles sont les méthodes les mieux appropriées à un exposé complet et pratique de la grammaire d'une langue quelconque?" Premier Congrès International de Linguistes, Propositions, Nijmegen, 1928, 36-39, and Actes du Ier Congrès International de Linguistes du 10-15 avril, 1928, 33-36. [S. Karcevski and N. Trubetzkoy added their signatures.] "K odstranovâni dlouhych souhlâsek v cestinë," Slavia, VII, 1928, 25-32. [with JU. Tynjanov] "Problemy izucenija literatury i jazyka," Novyj Lef, No. 12, 1928, 36-37, and Slovenské smery, V, 1938, 266-268, or Teôria literatury, edited by M. Bakos, Trnava, 1941, 101-103; also Readings in Russian Poetics, ed. by L. Matejka, Ann Arbor, 1962, 99-102. "Nové präce o üloze cirkevnëslovanskych prvkû v rustine," Casopis pro moderni filologii, XV, 1929, 174.

22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35.

36.

37.

"Zur vergleichenden Forschung über die slavischen Zehnsilbler," Slavistische Studien — Franz Spina zum sechzigsten Geburtstag, Reichenberg, 1929, 7-20. [with P. Bogatyrev] "Die Folklore als eine besondere Form des Schaffens," Donum Natalicium Schrijnen, Nijmegen-Utrecht, 1929, 900-913. Remarques sur l'évolution phonologique du russe comparée à celle des autres langues slaves (= Travaux du Cercle Linguistique de Prague, II), 1929, 118 pp. "Über die heutigen Voraussetzungen der russischen Slavistik," Slavische Rundschau, I, 1929, 629-646. "Mladenov, Stefan: Geschichte der bulgarischen Sprache," ibidem, 682-684. "Jan Baudouin de Courtenay," ibidem, 809-812, and Portraits of Linguists, ed. by T. A. Sebeok, I, Indiana University Press, 1966, 533-537. Nejstarsi ceské pisnë duchovni (= Nârodni knihovna, VI), Prague, 1929, 48 pp. "Kus literârni pavëdy," Plan, I, Prague, 1930, 593-597. "O prekladu versû," Plan, II, Prague, 1930, 9-11. [with F. Slotty] "Die Sprachwissenschaft auf dem ersten Slavistenkongress in Prag vom 6-13 Oktober 1929," Indogermanisches Jahrbuch, XIV, 1930, 384-391. "Masaryk o jazyke," Central'naja Evropa, III, 1930, 270-276. "Von einer Generation, die ihre Dichter vergeudet hat," Slavische Rundschau, II, 1930, 481-495 (abridged version of item 34). "O pokolenii, rastrativsem svoix poètov," Smert' Vladimira Majakovskogo, Berlin, 1931, 7-45. "Jazykové problémy v Masarykovë dile," Masaryk a fee, Prague, 1931, 29-47 (revised and enlarged version of item 32) and in Vùdce generaci, Prague, 1931, 396-414. [with P. Bogatyrev] "K problème razmezevanija fol'kloristiki i literaturovedenija," Lud Slowianski, II B, 1931, 230-233, and Teôria literatury, edited by M. Bakos, Trnava, 1941, 104-106. "Pamjati Vjaceslava Vjaceslavoviëa Hanky," Central'naja Evropa, IV, 1931,

BIBLIOGRAPHY

XIII

268-275. "O russkom perevodöike poèmy Havlicka 'Krescenie sv. Vladimira'," ibidem, 327-337. 39. "Z fonologie spisovné slovenstiny," Slovenskâ miscellanea, presented to Albert Prazäk, Bratislava, 1931, 155-163. 40. "Les unions phonologiques des langues," Le Monde Slave, 1931, 388-395. 41. "O fonologiceskix jazykovyx sojuzax," Evrazija v svete jazykoznanija, Prague, 1931, 7-12 (revised version of item 40). 42. K xarakteristike evrazijskogo jazykovogo sojuza, Paris, 1931, 59 pp. 43. "Die Betonung und ihre Rolle in der Wort- und Syntagmaphonologie," Travaux du Cercle Linguistique de Prague, IV, 1931, 164-183. 44. "Über die phonologischen Sprachbünde," ibidem, 234-240. 45. "Prinzipien der historischen Phonologie," ibidem, 247-267. 46. "Projet de terminologie phonologique standardisée," ibidem, 309-323. 47. "Tolstoj o Masaryke," CentraVnaja Evropa, IV, 1931, 712-716. 48. "Masaryk vu par Tolstoi," Le Monde Slave, 1933, 384-391 (enlarged version of item 47). 49. "Der russische Frankreich-Mythus," Slavische Rundschau, III, 1931, 636-642. 50. "Odpovëd' dr. Frantisku Tichému," Slavia, X, 1931, 396-400. 51. "Neue cechoslovakische Arbeiten über die poetische Form (1929-1930)," Slavische Rundschau, III, 1931, 450-454. 52. "Neue Arbeiten zur Vergangenheit Karpathorusslands," ibidem, 512-515. 53. "Pesma o vojevanju Igorovu," ibidem, 617-618. 54. "Zur Struktur des russischen Verbums," Charisteria Gvilelmo Mathesio oblata, Prague, 1932, 74-84; A Prague School Reader in Linguistics, compiled by J. Vachek, Indiana University Press, 1964, 347-359; and Readings in Linguistics, II, ed. by E. P. Hamp al., Chicago University, 1966, 22-30. 55. "O jednom typu literérnich historikû," Jarnl almanach Kmene, Prague, 1932, 74-84. 56. Association internationale pour les études phonologiques, Bulletin d'information, No. 1, 1932, 6 pp., and in Casopis pro modernl filologii, XIX, 1932, 59-64. 57. "O dnesnim brusicstvi ceském," Spisovnâ cestina a jazykovâ kultura, Prague, 1932, 85-122. 58. "Neue Arbeiten über die südslavische dichterische Form," Slavische Rundschau, IV, 1932, 257-279. 59. "Arbeiten über die cechische dichterische Form," ibidem, 506-510. 60. "André Vaillant: Les chants épiques des Slaves du sud," Byzantinoslavica, IV, 1932, 194-202. 61. "Fonéma," Ottâv slovnik naucny, Dodatky II, 1932, 608. 62. "Fonologie," ibidem, 611-612. 63. "Über den Versbau der serbokroatischen Volksepen," Proceedings of the First International Congress of Phonetic Sciences, Amsterdam, July 3-8, 1932 38.

XIV

64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92.

BIBLIOGRAPHY

(= Archives Néerlandaises de Phonétique Experimentale, VII-IX, 1933), 44-53. "Bolgarskij pjatistopnyj jamb v sopostavlenii s russkim," Sbornik v cesi' na prof. L. Miletic, Sofia, 1933, 108-117. "Pravení o jinochu a dívce — Staroruská povídka," Milostny almanach Kmene, Prague, 1933, 94-101. "Úpadek filmu?" Listy pro uméní a kritiku, I, Prague, 1933, 45-49. "La scuola linguistica di Praga," La Cultura, XII, 1933, 633-641. "O predpokladech prazské linguistické skoly," Index, VI, Brno, 1934, 6-9. "Co je poesie?" Volné sméry, Prague, 1933-1934, 229-239 and Teoria literatúry, edited by M. Bakos, Tmava, 1941, 170-181. "Vers staroöesky," Ceskoslovenská vlastivéda, III, Prague, 1934, 429-459. "Slavische Sprachfragen in der Sovjetunion," Slavische Rundschau, VI, 1934, 324-343. "Metrica Slava," Enciclopedia Italiana, XXIII, 1934, 112-113. "Perpetuum mobile kyvadla," Listy pro uméní a kritiku, II, Prague, 1934, 73-79. "M. P. Stokmar: Bibliografía rabot po stixoslozeniju," Slavia, XIII, 1934, 416-431. "Slezsko-polská cantilena inhonesta ze zacátku XV století," Národopisny véstnik Ceskoslovansky, XXVII-XXVIII, 1934-35, 56-84. "K casovym otázkám nauky o ceském versi: I. Starocesky vers a Rukopisy; II. íesky vers pred tisici lety," Slovo a slovesnost, I, 1935, 46-53. "T. G. Masaryk, monografie Zdeñka Nejedlého," ibidem, 124-126. "Rub literární védy," ibidem, 130-132, and II, 1936, 133-135. "Ceské prvky v polské kultufe," ibidem, I, 1935, 132. "Poznámky k dilu Erbenovu: I. O mythu; II. O versi," ibidem, 152-164, 218-229. "Obecná linguistika v SSSR," ibidem, 187-188. Association internationale pour les études phonologiques, Information Bulletin, No. 2, Prague, 1935, 12 pp. "Linguistika," Ottùv slovník naucny, Dodatky III, 1935, 149-162. "Zur Geschichte der tschechisch-polnischen Beziehungen," Prager Rundschau, V, 1935, 50-54. "Les enclitiques slaves," Atti del III Congresso internazionale dei linguisti, Florence, 1935, 384-390. "Kontury Glejtu," Boris Pasternak, Glejt, Prague, 1935, 149-162. "Randbemerkungen zur Prosa des Dichters Pasternak," Slavische Rundschau, VII, 1935, 357-374 (enlarged versión of item 86). "Úvahy o básnictví doby husitské," Slovo a slovesnost, II, 1936, 1-21. "Památce Agenora Artymovyöe," ibidem, 63-64. "Bádání o òeském versi," ibidem, 121-122. "Kodañská prehlidka dnesního jazykozpytu," ìbidem, 166-169. "Um den russischen Wortschatz," Slavische Rundschau, VIII, 1936, 80-90.

BIBLIOGRAPHY

93. 94. 95.

96. 97. 98. 99. 100.

101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118.

XV

"Mathesius, Vilém," Ottùv siovník naucny, Dodatky IV, 1936, 132. "Metrika," Ottùv siovník naucny, Dodatky IV, 1936, 213-218. "Na okraj lyrickych básní Puskinovych," Vybrané spisy A. S. Puskina, edited by A. Bém and R. Jakobson, I, Prague, 1936, 259-267, and Listy pro umëni a kritiku, IV, 1936, 389-392. "Neues zur Geschichte der altrussischen Literatur," Slavische Rundschau, VIII, 1936, 255-262. "Památce G. I. Celpanova," Psychologie, II, Brno, 1936, 41-42. "Usmërnëné názory na staroöeskou kulturu," Slovo a slovesnost, II, 1936, 207-222. "O cestách k ôeské poesii gotické," Zivot, XIV, Prague, 1936, 57-63. "Beitrag zur allgemeinen Kasuslehre (Gesamtbedeutungen der russischen Kasus)." Travaux du Cercle Linguistique de Prague, VI, 1936, 240-288, and Readings in Linguistics, II, ed. by E. P. Hamp al., Chicago University, 1966, 51-89. "Antoine Meillet zum Gedächtnis," Slavische Rundschau, IX, 1937, 24-26. "Socha v symbolice Puskinovë," Slovo a slovesnost, III, 1937, 2-24. "Puskinovy bàsnë v prekladu Ilji Bárta," ibidem, 122-124. "Spornyj vopros drevnerusskogo pravopisanija," Beliéev zbornik, Belgrade, 1937, 39-45. "Über die Beschaffenheit der prosodischen Gegensätze," Mélanges de linguistique et de philologie offerts à J. van Ginneken, Paris, 1937, 25-33. "Z zagadnieñ prozodji starogreckej," Prace ofiarowane Kazimierzowi Wóycickiemu, Wilno, 1937, 73-88. "Die russische Totenklage," Slavische Rundschau, IX, 1937, 247-249. "Russische Klagen," ibidem, 404-405. "Základy stredovëku," Slovo a slovesnost, III, 1937, 187-189. "Staroëeskà píseñ o bitvë u Varny 1444," ibidem, 189-190. "Na okraj Eugena Onëgina," Vybrané spisy A. S. Puskina, edited by A. Bém and R. Jakobson, III, Prague, 1937, 257-265. "Glosy k legendë o sv. Prokopu," XLIX rocenka Chudym dëtem: ¿ivy Vrchlicky, Brno, 1937, 65-77. "L'histoire du tchèque et du slovaque littéraires," Le Monde Slave, 1937, 353-366. "Jazyk rusky," Pedagogická encyklopedie, I, 1937, 555-556. "Z dëjin staroëeského zpëvného básnictví," Slovo a slovesnost, IV, 1938, 41-44. "K Puskinovym ohlasûm lidové poesie," Vybrané spisy A. S. Puskina, edited by A. Bém and R. Jakobson, IV, Prague, 1938, 248-254. "Die Arbeit der sogenannten 'Prager Schule'," Bulletin du Cercle Linguistique de Copenhague, III, 1938, 6-8. "Není pravda, 2e... Odpovëd na broáuru K. Bittnera 'Deutsche und Tschechen. Eine Erwiderung'," Slovo a slovesnost, IV, 1938, 117-123.

XVI

119. 120. 121. 122. 123. 124.

125. 126. 127.

128. 129.

130. 131. 132. 133. 134.

135.

136. 137.

BIBLIOGRAPHY

"Franz Spina," Slavische Rundschau, X, No. 6, 1938, 1-5. "Die Reimwörter Cech-Lech," ibidem, 10-15. "K popisu Mâchova verse," Torso a tajemstvi Mâchova dila, Prague, 1938, 207-278. "Sur la théorie des affinités phonologiques des langues," Actes du IV Congrès International de Linguistes, Copenhague, 1938, 48-58. "Vyznam ruské filologie pro bohemistiku," Slovo a slovesnost, IV, 1938, 223-239. "Observations sur le classement phonologique des consonnes," Proceedings of the Third International Congress of Phonetic Sciences, Ghent, 1939, 34-41; abstract in Slovo a slovesnost, IV, 1938, 192. [pseudonym Olaf Jansen] "Cesky podil na cirkevnëslovanské kulture," Co daly nase zemë Evropë a lidstvu, I, Prague, 1939, 9-20. [pseudonym Olaf Jansen] "Cesky vliv na staropolské pisemnictvi," ibidem, 48-51. "Signe zéro," Mélanges de linguistique offerts à Charles Bally, 1939, 143-152, and Readings in Linguistics, II, ed. by E. P. Hamp al., Chicago University, 1966, 109-115. [pseudonym Olaf Jansen] "Sobaka Kaiin car'," Slavia, XVII, 1939, 82-98. "Le développement phonologique du langage enfantin et les cohérences correspondantes dans les langues du monde," Vme Congrès International des Linguistes, Bruxelles, 1939, Résumés des communications, Bruges, 1939, 27-28. "Nikolaj Sergejeviö Trubetzkoy," Acta Linguistica, I, 1939, 64-76, and Portraits of Linguists, ed. by T. A. Sebeok, II, Indiana University Press, 1966, 526-542. "van Wijk, N.: Phonologie. " ibidem, 123-129. "Das Nullzeichen," Bulletin du Cercle Linguistique de Copenhague, IV, 1940, 12-14. "Tatarische Nachklänge in der russischen Volksepik," Det Norske VidenskapsAkademi i Oslo, Ârbok 1940, Oslo, 1941, 4-5. Kindersprache, Aphasie und allgemeine Lautgesetze, Uppsala, 1941, 83 pp. (reprint from Sprâkvetenskapliga Sällskapets i Uppsala Förhandlingar, 1940-42 = Uppsala Universitets Ârsskrift, 1942:9). [with J. Lötz] Axiomatik eines Verssystems am mordwinischen Volkslied dargelegt = Thesen zu einem Vortrag im Ungarischen Institut, Stockholm, 1941, 7 pp. "Neizvestnye stixi Majakovskogo," NovoseVe, No. 2, New York, 1942, 57-62. "The Paleosiberian Languages," American Anthropologist, XLIV, 1942, 602-620.

138. Cesstvi Komenského, New York, 1942, 8 pp. 139. Moudrost starych Cechä, New York, 1943, 240 pp. 140. "Polish-Russian Cooperation in Science of Language," Bulletin of the Polish Institute of Arts and Sciences in America, I, 1943, 970-974.

BIBLIOGRAPHY

141.

142. 143.

144.

145. 146.

147. 148. 149. 150. 151. 152. 153. 154.

155.

156.

XVII

"Some Russian Echoes of Czech Hagiography : I. The Translation of St. Venceslav's Relics ; II. The Hermit Ivan and John the Baptist ; III. Adolph the Martyr," Annuaire de l'Institut de Philologie et d'Histoire Orientales et Slaves, VII, New York, 1944, 155-180. "Saint Constantin et la langue syriaque," ibidem, 181-186. "Nécrologie: B. Baxa, J. Chlumsky, O. Fischer, V. A. Francev, A. Fuchs, V. Groh, O. Hujer, E. A. Ljackij, B. Ljapunov, G. L. Lozinskij, B. Mendl, A. Novák, D. Orel, J. Páta, D. Rasovskij, F. I. Scerbatskoj, J. Schránil, A. M. Seliscev, F. Siedlecki, N. S. Trubetzkoy, J. Uher, D. N. Usakov, M. Weingart, N. van Wijk, G. Winter", ibidem, 504, 509, 516-518, 520, 527, 530, 532-535, 537-541, 544-547. "Franz Boas' Approach to Language," International Journal of American Linguistics, X, 1944, 188-195, and Portraits of Linguists, ed. by T. A. Sebeok, II, Indiana University Press, 1966, 127-139. "Introduction to Russian by George L. Trager," The Slavonic and East European Review, XXII, 1944, 120-133. "A Note on Aleut Speech Sounds" and "A List of Works relating to the Aleut Language," Bulletin of the New York Public Library, August, 1944, and in the reprint, "Aleutian Manuscript Collection," by A. Yarmolinsky, New York, 1944, 9-12. "Slovo o polku Igoreve," Novosel'e, No. 14-15, 1944, 46-62. "The Significant Features of Indo-European," American Association for the Advancement of Romance Philology, Bulletin No. 3, May, 1944, 11. "Saussure," Yivo Bieter, XXIV, 1944, 67-78. "The Beginnings of National Self-Determination in Europe," The Review of Politics, VII, 1945, 29-42. "On Russian Fairy Tales," Russian Fairy Tales, Pantheon, New York, 1945, 631-656. "H. Grégoire: investigateur de l'épopée," Byzantina Metabyzantina, I, New York, 1946, 20-22, and Flambeau, XLVII, 1964, 330-336. "Polish Scholarship and Pushkin," The American Slavic and East European Review, V, No. 12-13, 1946, 88-92. "A List of Works relating to the Kamchadal Language and to the Language of Russianized Kamchadals," Bulletin of the New York Public Library, November, 1947, and in the reprint, "Kamchadal and Asiatic Eskimo Manuscript Collections," by A. Yarmolinsky, New York, 1947, 11-13. "Quelques remarques sur l'édition critique du Slovo, sur sa traduction en langues modernes et sur la reconstruction du texte primitif," La Geste du Prince Igor', sous la direction d'H. Grégoire, de R. Jakobson et de M. Szeftel (= Annuaire de l'Institut de Philologie et d'Histoire Orientales et Slaves, Université Libre de Bruxelles, VIII), New York, 1948, 5-37. "Edition critique du Slovo," ibidem, 38-78.

XVIII

157. 158. 159. 160. 161.

162. 163. 164. 165. 166.

167. 168. 169. 170. 171. 172. 173.

174. 175.

176. 177. 178. 179. 180.

BIBLIOGRAPHY

"Altérations du texte et leurs corrections," ibidem, 81-96 (plus 381). "Essai de reconstruction du Slovo dans sa langue originale," ibidem, 150-178. "Traduction du Slovo en russe moderne," ibidem, 181-200. "L'authenticité du Slovo," ibidem, 235-360 (plus 363-380). "Tasks and Aims of Comparative Slavic Literature (Abstract)," Program of the Sixty-Third Annual Meeting, Publications of the Modern Language Association of America, LXIII, Part 2, December, 1948, New York, 7. "Trudy otdela drevne-russkoj literatury, V" (review), The American Slavic and East European Review, VII, 1948, 375-376. "Zeitschrift für slavische Philologie, Vol. XIX" (review), ibidem, 383. "Slovo a slovesnost, Vol. X" (review), ibidem, 383-384. "Russian Conjugation," Word, IV, 1948, 155-167. "The Structure of Gilyak and Marr-ism," Mimeographed Program of the paper delivered on Sept. 8, 1949 at the XXIX International Congress of Americanists, New York, 1949, 5 + 2 pp. Slavic Languages, Columbia University, New York, 1949, 26 pp. Notes on General Linguistics: Its Present State and Crucial Problems (mimeographed), New York, 1949, 48 pp. "Notes autobiographiques de N. S. Troubetzkoy," Principes de Phonologie par N. S. Troubetzkoy, traduits par J. Cantineau, Paris, 1949, XV-XXIX. "Principes de phonologie historique," ibidem, 315-336 (revised version of item 45). "Sur la théorie des affinités phonologiques entre les langues," ibidem, 351-365 (revised version of item 122). "Les lois phoniques du langage enfantin et leur place dans la phonologie générale," ibidem, 367-379. [with M. Szeftel] "The Vseslav Epos," Russian Epic Studies, edited by Roman Jakobson and Ernest J. Simmons ( = Memoirs of the American Folklore Society, XLII, 1947), Philadelphia, 1949, 13-86. [with J. Lötz] "Notes on the French Phonemic Pattern," Word, V, 1949, 151-158. "The phonemic and grammatical aspects of language in their interrelations," Actes du Sixième Congrès International des Linguistes (Paris, Juillet, 1948), Paris, 1949, 5-18 and 601. "On the Identification of Phonemic Entities," Travaux du Cercle Linguistique de Copenhague, V, presented to Louis Hjelmslev, 1949, 205-213. "Slavic Mythology," Funk and Wagnalls Standard Dictionary of Folklore, Mythology and Legend, II, New York, 1950, 1025-1028. "O stixotvornyx reliktax rannego srednevekov'ja v cesskoj literaturnoj tradicii," Slavisticna Revija, Ljubljana, 1950, 267-273. "Les catégories verbales," Cahiers Ferdinand de Saussure, IX, 1950, 6. [with G. Ruziöic] "The Serbian Zmaj Ognjeni Vuk and the Russian Vseslav

BIBLIOGRAPHY

181. 182. 183. 184. 185. 186.

187.

188.

189. 190. 191. 192. 193. 194. 195.

196. 197.

198.

XIX

Epos," Annuaire de ¡'Institut de Philologie et d'Histoire Orientales et Slaves, X, Brussels, 1951, 343-355. "M. Vasmer: Russisches etymologisches Wörterbuch" (issues 1-3), Word, VII, 1951, 187-191. "On the Correct Presentation of Phonemic Problems," Symposium, V, 1951, 328-335. "The Puzzles of the Igor' Tale on the 150-th Anniversary of its First Edition," Speculum, XXVII, 1952, 43-66. "The Archetype of the First Edition of the Igor' Tale," Harvard Library Bulletin, VI, 1952, 5-15. "Langues paléosibériennes," Les Langues du Monde, Paris, 1952, 276-278 and 403-431. [with J. Lötz] "Axioms of a Versification System — Exemplified by the Mordvinian Folksong," Acta Instituti Hungarici Universitatis Holmiensis, 1952, 5-13 (revised version of item 135). [with C. G. M. Fant and M. Halle] Preliminaries to Speech Analysis ( = Acoustics Laboratory, Massachusetts Institute of Technology, Technical Report, XIII), January, 1952, 8 + 53 pp. [with C. G. M. Fant and M. Halle] Preliminaries to Speech Analysis, Second Printing, with Additions and Corrections, May 1952, 8 + 58 pp.; Third Printing, 1955; Fourth Printing, 1961; Fifth Printing, M.l.T. Press, 1963. "The Elementary Quanta of Language," The Journal of the Acoustical Society of America, XXIV, 1952, 581. "Studies in Comparative Slavic Metrics," Oxford Slavonic Papers, III, 1952, 21-66. "On Slavic Diphthongs Ending in a Liquid," Slavic Word, I, 2-6 ( = Word, VIII, 1952, 306-310). "Vestiges of the Earliest Russian Vernacular," Slavic Word, I, 46-51 ( = Word, VIII, 1952, 350-355). "New Slavic Etymological Dictionaries," Slavic Word, I, 83-90 ( = Word, VIII, 1952, 383-390). [with E. C. Cherry and M. Halle], "Toward the Logical Description of Languages in Their Phonemic Aspect," Language, XXIX, 1953, 34-46. Statements: lexicon meaning and grammatical meaning, 279-280; the cultural equivalent of the phoneme, 284-286; linguistic change, 292-293; pattern in linguistics, 310-314, in An Appraisal of Anthropology Today, Chicago, 1953. "The Kernel of Comparative Slavic Literature," Harvard Slavic Studies, I, 1953, 1-71. Results of the Conference of Anthropologists and Linguists (= Indiana University Publications in Anthropology and Linguistics, Memoir VIII, 1953), Chapter Two, 11-21. "Polska literatura sredniowieczna a Czesi," Kultura, No. 6/68, Paris, 1953,

XX

199.

200. 201. 202. 203. 204. 205. 206.

207.

208. 209. 210.

211. 212. 213. 214.

215. 216.

BIBLIOGRAPHY

27-42. "The Yiddish Sound Pattern and Its Slavic Environment," Yidishe Shprakh, XIII, 1953, 70-83, and Judah A. Joffe Book, Yivo Institute, New York, 1958, 207-220. "B. O. Unbegaun and J. C. G. Simmons, A Bibliographical Guide to the Russian Language," Slavic Word, II, 76-83 ( = Word, IX, 1953, 400-407). "A. V. Arcixovskij and M. N. Tixomirov, Novgorodskie gramoty na bereste," Slavic Word, II, 83-85 ( = Word, IX, 407-409). "Comparative Slavic Studies," The Review of Politics, XVI, 1954, 67-90. [with R. Poggioli] Cantare della Gesta di Igor, Testo critico annotato, Torino, 1954, 87-208. "St. Constantine's Prologue to the Gospels," St. Vladimir's Seminary Quarterly, New York, Summer, 1954, 19-23. "Minor Native Sources for the Early History of the Slavic Church," Harvard Slavic Studies, II, 1954, 39-74. "Tetrad' knjazja Belosel'skogo," Slovo o polku Igoreve v perevodax konca vosemnadcatogo veka (= Studies in Russian Epic Tradition, I), Leiden, 1954, 31-52. Ivan FedoroVs Primer of 1574: Facsimile edition, with commentary by R. Jakobson and appendix by W. A. Jackson, Cambridge, Massachusetts, 1955, 45 pp. and 24 plates. Preprinted from Harvard Library Bulletin, IX, 1955. Slavic Languages: a Condensed Survey (revised version of item 167), Columbia University, New York, 1955, 36 pp. "Unpublished Majakovskij," Harvard Library Bulletin, IX, 1955, 285-287 and 2 plates. "Aphasia as a Linguistic Problem," On Expressive Language, Clark University Press, Worcester, Massachusetts, 1955, 69-81, and Psycholinguistics, ed. by Sol Saporta, New York, 1961, 419-427. "The Origin of Russia. By H. Paszkiewicz," The American Historical Review, LXI, 1955, 106-108. "While Reading Vasmer's Dictionary," Slavic Word, IV ( = Word, XI, No. 4, 1955, 611-617). "K. Taranovski: Ruski dvodelni ritmovi, I-II" (review), Slavic Word, IV, (= Word, XI, No. 4, 1955, 644-647). Fundamentals of Language: I. [with M. Halle] Phonology and Phonetics; II. Two aspects of Language and Two Types of Aphasie Disturbances ( = Janua Linguarum, I), The Hague, 1956, X + 87 pp. "Die Verteilung der stimmhaften und stimmlosen Geräuschlaute im Russischen," Festschrift für Max Vasmer, Berlin, 1956, 199-202. "Novye stroki Majakovskogo: I. Tekst i primeöanija; II. Komentarij k pozdnej lirike Majakovskogo," Russkij literaturnyj arxiv, published under the auspices of the Harvard College Library and the Department of Slavic

BIBLIOGRAPHY

217. 218.

219.

220. 221.

222.

223.

224. 225. 226. 227. 228. 229.

230. 231.

232.

233.

XXI

Languages and Literatures of Harvard University, I, New York, 1956, 173-206. "Red a pisemnictvi Seskych zidfi v dobe premyslovske," Rok 1957, Moravian Library, New York, 1957, 35-46. "Serge Karcevski," Cahiers Ferdinand de Saussure, XIV, 1956, 9-16, and Portraits of Linguists, ed. by T. A. Sebeok, II, Indiana University Press, 1966, 533-537. "Izufienie slavjanskix jazykov i sravnitel'noe slavjanovedenie v Soedinennyx Statax Ameriki za poslevoennoe desjatiletie," Beogradski medunarodni slavisticki sastanak (15-21.IX.1955), Belgrade, 1957, 415-428. "Balladic Byliny Recorded in the South-Ladoga Basin," Journal of the American Folklore Society, No. 273, 1956, 236-238. [with J. Besharov and H. A. Wolfson] "An Old Russian Treatise on the Divine and Human Word," St. Vladimir's Seminary Quarterly, IV, No. 1-2, New York, 1956, 45-50. Fragmenta Chilandarica. Facsimile edition prefaced by R. Jakobson (= Monumenta Musicae Byzantinae, V, A. Sticherarium, 'Preface', 7-9; B. Hirmologium), Copenhagen, 1957. Shifters, Verbal Categories, and the Russian Verb. Russian Language Project, Department of Slavic Languages and Literatures, Harvard University, 1957, 14 pp. "The Relationship between Genitive and Plural in the Declension of Russian Nouns," Scando-Slavica, III, presented to Olaf Broch, 1957, 181-186. "Mufaxxama — the 'Emphatic' Phonemes in Arabic," Studies presented to Joshua Whatmough, The Hague, 1957, 105-115. "Stixotvornye citaty v velikomoravskoj agiografii," Slavisticna Revija, X, Ljubljana, 1957, 111-118. "StaroSeskd Msen o prvotnim hrichu," Listy filologicke, V (LXXX), Prague, 1957, 204-210. [with M. Halle] "Phonology in Relation to Phonetics," Manual of Phonetics, ed. by L. Kaiser, Amsterdam, 1957, 215-251. [with G. Hiittl-Worth and J. F. Beebe] Paleosiberian Peoples and Languages, A Bibliographical Guide, New Haven, 1957, 222 pp. with "A Short Sketch of the Paleosiberian Peoples and Languages" by R.J., 218-222. "Notes on Gilyak," Studies Presented to Yuen Ren Chao (= Bulletin of the Institute of History and Philology, Academia Sinica, XXIX, 1957), 255-281. "The Cardinal Dichotomy in Language," Language: An Inquiry into Its Meaning and Function, ed. by R. N. Anshen, New York, 1957, 155-173 (a variant of item 214, II). "Typological Studies and their Contribution to Historical Comparative Linguistics," Proceedings of the VIII International Congress of Linguists 1957, Oslo, 1958, 17-25. "IzuCenie Slova o polku Igoreve v Soedinennyx Statax Ameriki," Trudy Otdela

XXII

234.

235. 236.

237. 238. 239.

240. 241. 242. 243. 244. 245. 246. 247. 248. 249.

250.

251.

BIBLIOGRAPHY

drevnerusskoj literatury AN SSSR, XIV, 1958, 103-121. "Morfologiôeskie nabljudenija nad slavjanskim skloneniem," American Contributions to the IVth International Congress of Slavists, The Hague, 1958, 127-156 (and separately — 30 pp.) with an English summary "Morphological Inquiry into Slavic Declension," 154-156. "Medieval Mock Mystery (The Old Czech Unguentarius)," Studia philologica et Utteraria in honorem L. Spitzer, Bern, 1958, 245-265. "Powstanie pojçcia fonemu w lingwistyce polskiej i swiatowej," PAN, Sprawozdania z Prac Naukowych Wydzialu Nauk Spolecznych, I, No. 6, Warsaw, 1958, 48-53. "Autobiographische Notizen von N. S. Trubetzkoy," N. S. Trubetzkoy, Grundziige der Phonologie, Gôttingen, 1958 and 1962, 273-288. "On Linguistic Aspects of Translation," On Translation, Harvard University Press, 1959, and New York, Oxford University Press, 1966, 232-239. [with M. Halle] "Supplementary Note" to E. Petrovici, "La distinction entre trois sortes de n et 1 — non diésés, diésés et palatals — en roumain et en slave," International Journal of Slavic Linguistics and Poetics, I/II, 1959, 192-194. "Marginalia to Vasmer's Russian Etymological Dictionary (R-JA)," ibidem, 265-278. "A. R. Luria and F. Ja. Judovic : Re£' i razvitie psixiôeskix processov u rebenka" (review), ibidem, 279-280. "R. L. Osvalt : Russian Loan Words in Southwestern Pomo" (review), ibidem, 286. "R. I. Avanesov, Fonetika sovremennogo russkogo literaturnogo jazyka" (review), ibidem, 286-289. "I. P. Eremin and D. S. Lixaôev, Xudozestvennaja proza Kievskoj Rusi" (review), ibidem, 304-305. "Za i protiv Viktora Sklovskogo," ibidem, 305-310. "Boris Viktoroviô Tomasevskij," ibidem, 313-316. "Gyula Laziczius," ibidem, 348. "Linguistic Glosses to Goldstein's 'Wortbegriff'," Journal of Individual Psychology, XV, dedicated to Kurt Goldstein, 1959, 62-65. "Boas' View of Grammatical Meaning," The Anthropology of Franz Boas. Essays on the Centennial of His Birth (= American Anthropologist, LXI, 5, part 2; Memoir 89, 1959), 139-145. Trois conférences données à Bucarest les 3 et 6 oct. 1958. Bucarest, Institut de linguistique de l'Académie de la République Populaire Roumaine, 1959, 93 mimeogr. pages: Les problèmes les plus actuels de l'étude des sons du langage; Discussions sur la linguistic mathématique; Sur les méthodes d'analyse de la langue. Description and Analysis of Contemporary Standard Russian, I, II, and III. Ed. by R. Jakobson and C. H. van Schooneveld, The Hague, 1959-61.

BIBLIOGRAPHY

252.

253. 254. 255. 256. 257. 258.

259. 260. 261. 262.

263.

264.

265. 266.

267.

268. 269.

XXIII

"Introductory Note" by R. Jakobson and C. H. van Schooneveld, I, 5-8. "Linguistics and Poetics," Style in Language, ed. by T. Sebeok, New York, 1960, 350-377, and Essays on the Language and Literature, ed. by S. Chatman and S. R. Levin, Boston, 1967, 296-322. "Poetyka w swietle jçzykoznawstwa," Pamiçtnik Literacki, LI, 1960, 431-473. "Stroka Maxi o zove gorlicy," International Journal of Slavic Linguistics and Poetics, III, 1960, 87-108. "Z zagadnieñ struktury czeskiego poematu romantycznego," Pamiçtnik Literacki, LI, 1960, 389-409. "Velikaja Moravija ili Velikaja nad Moravoj," Ezikovedsko-Etnografski izsledvanija v pamet na akad. St. Romanski, Sofia, 1960, 483-486. "Why 'Mama' and 'Papa'," Perspectives in Psychological Theory, Essays in Honor of Heinz Werner, New York, 1960, 124-134. "Nepovsimnuté filiace" : I. Kunhutina skladba a modlitby Miliôovy ; II. Slovanské duchovni dëjiny v pojetí Jana Amóse Komenského, Scando-Slavica, VI, 1960, 26-34. "Metajçzyk w krçgu zagadnieñ lingwistycznych," Sprawozdania Poznanskiego Towarzystwa Przyjaciól Nauk, LIII, 1960, 157-158. "The Gender Pattern of Russian," Omagiu lui Al. Graur (= Studii si Cercetári Lingvistice, XI, 1960), 541-543. "Kazañska szkola polskiej lingwistyki i jej miejsce w swiatowym rozwoju fonologii," Biuletyn Polskiego Towarzystwa Jçzykoznawczego, XIX, 1960, 3-34. [with M. Halle] Grundlagen der Sprache, ed. by G. F. Meier, Berlin, 1960, V I I I + 74 pp. ( = Schriften zur Phonetik, Sprachwissenschaft und Kommunikationsforschung, I). German translation of item 214. "Wklad jçzykoznawstwa do krytycznej analizy tekstu Slowa o wyprawie Igora," Zeszyty Naukowe JJniwersytetu Jagielloñskiego, Prace Jqzykoznawcze, No. 3, 1960, 301-302. Structure of Language and its Mathematical Aspects, ed. by R. Jakobson (= Proceedings of Symposia in Applied Mathematics, XII, American Mathematical Society, 1961), V I + 2 7 9 pp.; Introduction by R. Jakobson pp. V-VI. "Linguistics and Communication Theory," ibidem, 245-252. "The Slavic Response to Byzantine Poetry," XIIme Congrès International des Etudes Byzantines, Ochride, 1961, preprint, Belgrade-Ochride, 1961, and definitive, corrected edition, 1963, 249-265. Tônnies Fenne's Low German Manual of Spoken Russian, Pskov, 1607, Vol. I, ed. by L. L. Hammerich, R. Jakobson, E. van Schooneveld, T. Starck, and Ad. Stender-Petersen, Royal Danish Academy of Sciences, 1961. "Foreword" by R. Jakobson and E. van Schooneveld, 5-31. "Struktura na poslednoto Botevo stihotvorenie," Ezik i literatura, XVI, 1961, 1-14. [with M. Halle] "Phonemic Patterning," Psycholinguistics, ed. by Sol Saporta,

XXIV

270. 271.

272. 273. 274. 275. 276. 277. 278. 279. 280. 281. 282.

283. 284.

285. 286. 287.

288.

BIBLIOGRAPHY

New York, 1961, 346-350 (re-ed. of item 208, Part I, §§ 4.1-4.16). "Poèzija grammatiki i grammatika poèzii," Poetics, Poetyka, Poètika, Warsaw, 1962, 397-417. Studies in Russian Philology: I. K lingvistiöeskomu analizu russkoj rifmy; II. O morfologiöeskom sostave drevnerusskix otöestv ( = Michigan Slavic Materials, ed. by L. Matejka, I, Ann Arbor, 1962), 26 pp. "Deux aspects du langage et deux types d'aphasies," Les Temps Modernes, XVII, No. 188, 1962, 853-880 (French translation of item 208, II). [with C. G. Fant and M. Halle] "Vvedenie v analiz reci," Novoe v lingvistike, II, Moscow, 1962, 173-230 (Russian translation of item 184). [with M. Halle] "Fonologija i ee otnosenie k fonetike," ibidem, 231-278 (Russian translation of item 214,1). [with E. Cherry and M. Halle] "K voprosu o logiöeskom opisanii jazykov v ix fonologiöeskom aspekte," ibidem, 279-298 (Russian translation of item 194). "Russkij istoönik öesskoj komedii: Bratri Capkové, Ze iivota hmyzû," Studii in onore di Ettore Lo Gatto e Giovanni Maver, Rome, 1962, 331-335. "Diskussionsbeitrag," Zeichen und System der Sprache, II, Berlin, 1962, 50-56. [with C. Lévi-Strauss] "Les Chats de Charles Baudelaire," VHomme, II, 1962, 5-21, and II Corpo, I, 1965, 43-59. "O sootnosenii mezdu pesennoj i razgovornoj narodnoj reö'ju," Voprosy jazykoznanija, XI, No. 3, 1962, 87-90. "On the Rumanian Neuter," Mélanges linguistiques offerts à Emil Petrovici (= Cercetäri de Lingvisticä, III, Supliment, 1958/1962), 237-238. [with A. Sommerfelt] "On the Role of Word Pitch in Norwegian Verse," Studia gratulatoria dedicated to A. W. de Groot (= Lingua, XI, 1962), 205-216. Selected Writings, I: Phonological Studies, The Hague, 1962, X + 678 pp. With unpublished papers: "Zur Struktur des Phonems," 280-310; "Die urslavischen Silben ür-, ül-," 546-549; "Retrospect," 629-658. [with B. Casacu] "Analyse du poème Revedere de Mihai Eminescu," Cahiers de linguistique théorique et appliquée, I, Bucharest, 1962, 47-54. "Morfologiöeskie nabljudenija nad slavjanskim skloneniem" (Tezisy; zakljuôitel'noe slovo), IV Mezdunarodnyj s"ezd slavistov, II, Moscow, 1962, 31-33, 47-48. "Fonetika i fonologija slavjanskix jazykov," IV Mezdunarodnyj s"ezd slavistov, II, Moscow, 1962, 308-310. "Parts and Wholes in Language," Parts and Wholes, ed. by D. Lerner, New York-London, 1962, 157-162. Discussions, Actes du Colloque international de civilisations, litératures et langues romanes, 1959, Bucharest, 1962, 201-202: Atlas linguistiques; 236-238: Phonologie. "O xudozestvennom realizme," Readings in Russian Poetics (= Michigan Slavic Materials, ed. by L. Matejka, II, Ann Arbor, 1962), 29-36 (Russian original

BIBLIOGRAPHY

289. 290. 291.

292. 293. 294. 295. 296.

297. 298. 299. 300. 301.

302. 303. 304. 305.

306.

307.

XXV

of the Czech and Ukrainian versions: items 4 and 14). "Anthony's Contribution to Linguistic Theory," Ruth H. Weir, Language in the Crib, The Hague, 1962, 18-20. "Struktura dveju srbohrvatskih pesama," Zbornik za filologiju i lingvistiku, IV-V, Novi Sad, 1961-62, 131-139. "Circular Number One," May, 1960, Proceedings of the Fourth Congress of Phonetic Sciences held at the University of Helsinki 4-9 September 1961, The Hague, 1962, pp. V-VI. "Concluding Remarks," ibidem, pp. XXV-XX1X. "The Phonemic Concept of Distinctive Features," ibidem, 440-455. "Implications of Language Universals for Linguistics," Universals of Language, Cambridge, Mass., 1963, 208-219, and 1966. "Boris Mixajloviô Èjxenbaum," International Journal of Slavic Linguistics and Poetics, VI, 1963, 160-167. "On the so called vowel alliteration in Germanic verse," Zeitschrift fur Phonetik, Sprachwissenschaft und Kommunikationsforschung, XVI: O. von Essen gewidmet, 1963, 85-94. With "A Note on the Dissimilation of Vowels in Icelandic Alliteration" by Sigrid Valfells. Essais de linguistique générale, ed. by N. Ruwet, Paris, 1963, 260 pp. "Toward a linguistic typology of aphasie impairments," The CIBA Foundation, Disorders of Language, Abstracts of Papers, London, 1963, 5-6. "Tipologiôeskie issledovanija i ix vklad v sravnitel'no-istoriceskoe jazykoznanie," Novoe v lingvistike, III, Moscow, 1963,95-105 (translation of item 232). "St. Constantine's Prologue to the Gospels," St. Vladimir's Seminary Quarterly, VII, No. 1, 1963, 14-19 (revised version of item 204). "Tainaja slufba Konstantina Filosofa i dal'nejsee razvitie staroslavjanskoj poèzii," Mélanges G. Ostrogorski, I ( = Zbornik radova Vizantoloskog instituta, VIII, Belgrade, 1963), 153-166. "Izbytocnye bukvy russkogo alfavita i smeznye orfografiôeskie voprosy," Ivsicev zbornik, Zagreb, 1963, 143-152. "'Prszesziosc' Cypriana Norwida," Pamiçtnik literacki, LIV, 1963, 449-456. "Ktoz jsu bozi bojovnici," International Journal of Slavic Linguistics and Poetics, VII, 1963, 108-117. "Opyt fonologiôeskogo podxoda k istoriceskim voprosam slavjanskoj akcentologii," American Contributions to the V International Congress of Slavists, The Hague, 1963, 153-176. English Summary: "A Phonemic Approach to the Structure and Evolution of the Common Slavic Prosodie Pattern," 176-178. "Efforts toward a Means-End Model of Language in Interwar Continental Linguistics," Trends in Modern Linguistics, II, Utrecht, 1963, 104-108, and A Prague School Reader in Linguistics, compiled by J. Vachek, Indiana University Press, 1964, 481-485. [with D. Worth] Sofonija's Tale of the Russian-Tartar Battle on the Kulikovo

XXVI

308. 309. 310. 311. 312. 313. 314. 315. 316. 317. 318.

BIBLIOGRAPHY

Field, The Hague, 1963, 71 pp. + 49 tables. "Analiza gramatyczna poezji slowianskiej," Zeszyty Naukowe Uniwersytetu Jagiellonskiego, LX, 1963, 429-432. "O lingwistycznej analizie rymu," Prace Filologiczne, XVIII: Witoldowi Doroszewskiemu, part 1, 1963, 47-52 (Polish version of item 271, I). "Patronimika w Slowie o polku Igoriewie," Studia linguistica in honorem Taddei Lehr-Splawinski, Cracow, 1963 (Polish version of item 271, II). "Towards a Linguistic Typology of Aphasie Impairments," CIBA Foundation Symposium on Disorders of Language, London, 1964, 21-46. [with M. Halle] "Tenseness and Laxness," In Honour of Daniel Jones, London, 1964, 96-101. "The Prosodie Questions of Slavic Historical Phonology Restated," M.I.T., R.L.E. Quarterly Progress Report, No. 72, 1964, 216-218. "Postscript," Two Essays on Poetic Language by O. M. Brik ( = Michigan Slavic Studies, V, Ann Arbor, 1964), 77-81. "Results of the Congress," Proceedings of the Ninth International Congress of Linguists, Cambridge, Mass., 1962, The Hague, 1964, 1135-1142. "A IX. Nemzetközi Nyelvészkongresszus zârôeloadâsa," Különlenyomat a Nyelvtudomânyi Kôzlemények, 1964, 1-8 (Hungarian translation of item 315). "Nachruf auf N. S. Trubetzkoy," Wiener Slavistisches Jahrbuch, XI, 1964, 22. "On Visual and Auditory Signs," Phonetica, XI: E. Zwirner gewidmet, 1964, 216-220.

319. [with M. Halle] "The Term Canaan in Medieval Hebrew," For Max Weinreich on His Seventieth Birthday, The Hague, 1964, 147-172. 320. "Lingvisticä si poeticä," Probleme de Stilisticä, Bucharest, 1964, 83-125 (Rumanian translation of item 252). 321. [with M. Halle] Podstawy jçzyka, ed. by L. Zawadowski, Wroclaw, 1964, 140 pp. (Polish translation of item 214 with passages from 188). 322. "Language in Operation," Mélanges Alexandre Koyré, I: Uaventure de Vesprit, Paris, 1964, 269-281. 323. [with E. C. Cherry and M. Halle] "K logickému popisu jazykû v jejich fonologickém aspektu," Theorie informace a jazykovëda, ed. by L. Dolezel, Prague, 1964, 100-113 (Czech translation of item 194). 324. "The Grammatical Structure of Janko Kräl"s Verses," Sbornik filozofickej fakulty Univerzity Komenského, XVI, 1964, 29-40. 325. "O latinizacii mezdunarodnyx telegramm na russkom jazyke," Voprosy Jazykoznanija, XIV, 1965, 111-113. 326. "List badacza polskiego," Kultura i Spoleczenstwo, IX, 1965, 13-21. 327. "Der grammatische Bau des Gedichts von B. Brecht 'Wir sind sie'," Beiträge zur Sprachwissenschaft, Volkskunde und Literaturforschung, W. Steinitz dargebracht, Berlin, 1965, 175-189. 328. "Information and Redundancy in the Common Slavic Prosodie Pattern,"

BIBLIOGRAPHY

329.

330. 331. 332.

333. 334. 335. 336. 337.

338. 339.

340. 341. 342. 343. 344.

345. 346. 347.

XXVII

Symbolae Linguisticae in honorem Georgii Kurylowicz, Cracow, 1965, 146-151. "Znaöenie lingvisticeskix universalij dlja jazykoznanija," Istorija jazykoznanija XIX-XX vekov v ocerkax i izvlecenijax, ed. by V. A. Zvegincev, II, Moscow, 1965, 383-395 (Russian translation of item 294). "Vystuplenie na 1-m mezdunarodnom simpoziume 'Znak i sistema jazyka'," ibidem, 395-402 (Russian translation of item 277). "Lingvistika i teorija svjazi," ibidem, 402-435 (Russian translation of item 266). "Razrabotka celevoj modeli jazyka v evropejskoj lingvistike v period mezdu dvumja vojnami," Novoe v lingvistike, IV, Moscow, 1965, 372-377 (Russian translation of item 306). "Itogi devjatogo kongressa lingvistov," ibidem, 577-588 (Russian translation of item 315). "Usöekotal" skaöa," Lingua viget: Commentationes Slavicae in Honorem V. Kiparsky, Helsinki, 1965, 83-89. "O budowie ukrainskiego rozkaznika," Studia z Filologii Polskiej i Siowianskiej, V: Zdzislawowi Stieberowi, 1965, 213-218. "Szczupak po polsku," Prace Polonistyczne, XX: Stefanii Skwarczynskiej, 1965, 132-141. "Methodius' Canon to Demetrius of Thessalonica and the Old Church Slavonic Hirmoi," Sbornik praci filosofické fakulty Brnënské university, F9: J. Rackovi, 1965, 115-121. "Poesie der Grammatik und Grammatik der Poesie," Mathematik und Dichtung, ed. by H. Kreuzer, Munich, 1965,21-32 (reshaped German version of item 270). "Vers une science de l'art poétique," Théorie de la littérature: textes des formalistes russes, ed. by Tzvetan Todorov, Paris, 1965, 9-13, and Les lettres françaises, No. 1118, Feb. 10-16, 1966, 3. "Du réalisme artistique," Théorie de la littérature, Paris, 1965, 98-108 (French translation of item 288). [with Ju. Tynjanov] "Les problèmes des études littéraires et linguistiques," ibidem, 138-140 (French translation of item 20). [with G. Fant and M. Halle] Onsei Bunseki Josetsu, Tokyo, 1965, 118 pp. (Japanese translation of item 188 by Shigeru Takemura and Osamu Fujimura). "Notes préliminaires sur les voies de la poésie russe," La Poésie Russe, ed. by Elsa Triolet, Paris, 1965, 2-28. "An Example of Migratory Terms and Institutional Models (On the Fiftieth Anniversary of the Moscow Linguistic Circle)," Omagiu lui Alexandru Rosetti, Bucharest, 1965, 427-431. "À la recherche de l'essence du langage," Diogène, 51, 1965, 22-38 (French version of item 346). "Quest for the Essence of Language," Diogenes, No. 51, 1966, 21-37. "Igor' Tale Reconstruction," appendix to T. Ciievska, Glossary of the Igor' Tale, The Hague, 1966, 390-403.

XXVIII

348. 349.

350.

351. 352.

353. 354. 355. 356. 357.

358. 359. 360. 361.

362. 363. 364. 365. 366.

BIBLIOGRAPHY

"The Byzantine Mission to the Slavs," Dumbarton Oaks Papers, XIX, 1966, 257-265. "The Grammatical Texture of a Sonnet from Sir Philip Sidney's 'Arcadia'," Studies in Language and Literature in Honour of M. Schlauch, Warsaw, 1966, 165-174. Selected Writings, IV : Slavic Epic Studies, The Hague, Paris, 1966, xii + 751 pp. With unpublished papers: "Za solomjanem / Za Solomonom," 534-539; "Vlijanie narodnoj slovesnosti na Trediakovskogo," 613-633; "Retrospect," 637-704; "Postscript," 738-751. [with JU. Tynjanov] "Problems of Literary and Linguistic Studies," New Left Review, XXXVII, 1966, 59-61 (English translation of item 20). "The Role of Phonic Elements in Speech Perception," XVIII International Congress of Psychology, Symposium 23 : Models of Speech Perception (preprint: Salk Institute for Biological Studies, San Diego, Calif., 1966), 12 pp. "Aspetti linguistici della traduzione," Il Verri, XIX, 1966, 98-106 (Italian translation of item 238). "Devuska pela," Orbis scriptus D. Tschizewskij zum 70. Geburtstag, Munich, 1966, 385-401. "Henry Sweet's Paths Toward Phonemics," In Memory of J. R. Firth, London, 1966, 242-254. "Grammatical Parallelism and Its Russian Facet," Language, XLII, presented to Yuen Ren Chao, 1966, 399-429. "Linguistic Types of Aphasia," Brain Function, III: Speech, Language, and Communication ( = Forum in Medical Sciences, No. 4, Berkeley and Los Angeles, 1966), 67-91. [with P. Colaclides] "Grammatical Imagery in Cavafy's Poem 'Remember, Body'," Linguistics, XX, 1966, 51-59. "K védè bàsnického umèni," Flamen, No. 8, Prague, 1966, 95-96 (Czech translation of item 339). "Glossolalie," Tel Quel, No. 26, 1966, 3-9 (French translation of a part of item 348: "Retrospect"). "L'architettura grammaticale della poesia Brechtiana 'Wir sind sie'," Paragone, NS XVIII, No. 198, Milano, 1966, 3-22 (Italian translation of the corrected text of item 325). Saggi di linguistica generale, ed. by Luigi Heilmann, Milan, 1966, XXXVI + 219 pp. Author's Preface: 1-2. "Razbor tobol'skix stixov Radisöeva," XVIII vek, VII : RoV i znacenie literatury XVIII veka v istorii russkoj kul'tury, Leningrad, 1966, 228-236. Lingvistika ipoetika, ed. by Milka Ivic, Belgrade, 1966, 327 pp. "Relationship between Russian Stem Suffixes and Verbal Aspects," Studia Slavica Academiae Scientiarum Hungaricae, XII, 1966, 203-206. [with P. Valesio] "Vocabulorum constructio in Dante's Sonnet 'Se vedi li

BIBLIOGRAPHY

367. 368. 369. 370.

371. 372.

373. 374.

XXIX

occhi miei'," Studi Danteschi, XLIII, Florence, 1966, 7-33. "Gramatická struktúra versov Janka Krála," Slovenská literatura, XIV, No. 1, 1967, 67-78 (Slovak translation of item 324). "Une microscopie du dernier Spleen dans les Fleurs du Mal," Tel Quel, No. 29, 1967, 12-24. "L'importanza di Kruszewski per lo sviluppo della linguistica generale," Ricerche Slavistiche, XIV, 1966, 1-20. [with M. Halle] "Fundamentals of Language," Language and Thought, ed. by D. C. Hildum, Princeton, N.J., 1967, 51-61 (Excerpt from item 214, I, §§ 4.1-4.3). "The Twofold Character of Language," ibidem, 171-176 (Excerpt from item 214, II). "About the Relation between Visual and Auditory Signs," Models for the Perception of Speech and Visual Form, ed. by W. Watten-Dun, Cambridge, Mass., 1967, 1-7. "Gengo Honshitsu no TankyQ," Diogenesu, No. 1, Tokyo, 1967, 36-51 (Japanese translation of item 346). Selected Writings, II: Word and Language, The Hague, 1967.

II. M I S C E L L A N E A ( N E W S P A P E R ARTICLES, LETTERS, FOREWORDS, A N D POEMS)

1. Letter to V. Xlebnikov (Feb. 1914) about new experiments in poetry, Majakovskij — Materialy i issledovanija, Akademija Nauk SSSR, Moscow, 1940, 385-386. 2. [pseudonym Aljagrov] Zaumnaja kniga, poetry edited by A. KruCenyx, Moscow, 1914. 3. [pseudonym Aljagrov] Poem (1914), Zaumniki, ed. by A. KruCenyx, Moscow, 1921, 16. 4. Translation of Majakovskij's poem, "Ni£ego ne ponimajut," into Old Church Slavonic (1918), Tridcat' dnej, Moscow, 1940, No. 9-10, 104. 5. "Futurizm," Iskusstvo, Moscow, Aug. 2, 1919. 6. "Stixi v Cukokkale" (1919) in K. I. Cukovskij, "Cto vspomnilos'," Prometej, I, Moscow, 1966, 239. 7. "Prof. Sachmatov," Cas, Prague, Aug. 31, 1920. 8. "V. Chlebnikov: Z poematu 'Sestry blyskavice'" [translation], Den, Prague, Dec. 27, 1920. 9. "Prof. V. N. Sfiepkin," Cas, No. 45, Prague, 1921. 10. "Konec basnickeho umprumactvi a zivnostnictvi," Pasmo, Brno, May 1925. 11. "Stanislav Neumann: Vesennie ritmy" [translation], Nove Rusko, I, Prague, 1925, No. 5-6, 164.

XXX

12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41.

BIBLIOGRAPHY

"Ein neuer Beitrag zur slavischen Verslehre," (Die altslawischen Verstexte von Kiew und Freising, by E. Sievers), Prager Presse, July 5, 1925. "K diskusi o òeské prosodii," Närodni osvobozeni, Prague, Aug. 8, 1926. "Nikolaj Durnovo," Prager Presse, Sept. 5, 1926. "Vladimir Majakovskij," Närodni osvobozeni, Prague, Apr. 26, 1927. "Wandlungen in der Sprachwissenschaft. Vom I. Internationalen Linguistenkongress in Haag," Prager Presse, Apr. 25, 1928. "Dem Gedächtniss Jan Wiktor Porzezinskis," Prager Presse, Apr. 17, 1929. "Romantické vseslovanstvi — nova slavistika," Cin, Prague, Oct. 31, 1929. "Vladislav Vanöura: Markéta Lazarovä," Literärni noviny, Prague, No. 9, May 1931. "Der Genfer Linguistenkongress," Prager Presse, Sept. 13, 1931. "Prof. Friedrich Slotty," Prager Presse, Sept. 18, 1931. "Tolstoj über Masaryk," Prager Presse, Sept. 28, 1931. "Tolstoj o Masarykovi," Literärni noviny, No. 18, Sept. 1931. "Neue Arbeiten N. van Wijks über das Altkirchenslavische," Prager Presse, Apr. 14, 1932. "Die Lautwissenschaften," Prager Presse, July 24, 1932. "Prof. Vilém Mathesius," Prager Presse, Aug. 3, 1932. "Die entschwindende Welt. Ein Film der sterbenden Folklore," Prager Presse, Sept. 11, 1932. "Musikwissenschaft und Linguistik," Prager Presse, Dee. 7, 1932. "Ukäzky z chystané monografie o slägrech V & W," Tucet melodii z Osvobozeného divadla, Prague, 1932. "Aleksandr Matvejeviö Peskovskij," Prager Presse, Apr. 11, 1933. "Aus der deutschen Russistik," Prager Presse, July 21, 1933. "Die Slavistik auf dem Linguistenkongress in Rom," Prager Presse, Oct. 13, 1933. "Cesko-ruské styky v minulosti," Lidové noviny, Brno, May 26, 1935. "Spoleönä ree kultury," Zemé Sovetü, Prague, July 1935, 109-111, and Prager Presse. [with P. Bogatyrev] "Aufzeichnungen über Tolstoj. Aus unveröffentlichen Papieren von Albert Skarvan," Prager Presse, Nov. 17, 1935. [with P. Bogatyrev] "Unbekanntes über und von Tolstoj," Prager Presse, Nov. 20, 1935. [with P. Bogatyrev] "Tolstojs Wildheit und Sanftmut," Prager Presse, Nov. 24, 1935. "Ein Grundwerk zur tschechoslovakischen Sprachgeschichte," Prager Presse, Jan. 5, 1936. "Aktuelle Aufgaben der Bylinenforschung," Prager Presse, Apr. 26, 1936. "Der IV. Internationale Linguistenkongress," Prager Presse, Sept. 12, 1936. "Kolybel'nyje" [translation from Czech into Russian], Vàclav Kapräl,

BIBLIOGRAPHY

42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62.

63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74.

XXXI

Uspävanky, Brno-Prague, 1936. "Policejni konfidentka opevovanä Puskinem a Mickiewiczem," Lidove noviny, Brno, Jan. 3, 1937. "Puskin v realistickem svStle," Program D 37, VI, Jan. 23, 1937, 133-136. "Nespoutany Puskin," Lidove noviny, Feb. 14, 1937. "Nikolaj Konstantinovic Nikolskij," Lidove noviny, Mar. 24, 1937. "Veda a noviny v pojeti Mägrove," Lidove noviny, Apr. 6, 1937. "Holger Pedersen sedmdesatnikem," Lidove noviny, Apr. 7, 1937. "Dopis Jifimu Voskovcovi a Janu Werichovi o noetice a semantice svandy," 10 let Osvobozeneho divadla, Prague, 1937. "Frank Wollmans Ideen zur slavischen Literaturgeschichte," Prager Presse, Apr. 29, 1937. "Ljubomir Miletiö zemrel," Lidove noviny, June 3, 1937. "Olaf Broch — 70 Jahre," Prager Presse, Aug. 4, 1937. "Neznäme Bulharsko," Lidove noviny, Sept. 5, 1937. "Rusko-cesky slovnik," Lidove noviny, Sept. 12, 1937. "Z nejstarsich dSjin polskeho a ceskeho bäsnictvi," Lidove noviny, Oct. 7, 1937. "K dejinäm cyrilometodejstvi," Lidove noviny, Oct. 19, 1937. [pseudonym O. Jamal] "Je zeme blizko polu," Lidove noviny, Nov. 7, 1937. "Prehlednutä pamätka velkomoravskä," Lidove noviny, Dec. 23, 1937. "Svaty Alexius v slovanskych literaturäch," Lidove noviny, Dec. 24, 1937. "Ruske vypravy do budoucna," Lidove noviny, Jan. 1, 1938. "Srdce ceskoslovenskeho trelistu," Lidove noviny, May 7, 1938. "Prispevek k diskusi," Ctvrtletnik skupiny Blok, I, Brno, 1938, 86-87. [with B. Havränek and F. Trävniöek] Sborupro vyzkum Slovenska a Podkarpat. Rusi pri Slovanskem üstave (Vyvoj soucasneho spisovneho jazyka na Podkarpatske Rusi), Prague, 1938, 4 pp. "Podvrzenä milenka bäsnikova," Lidove noviny, Mar. 13, 1938. "Nikolaj Trubeckoj zemrel," Lidove noviny, June 29, 1938. "Mezinärodni sjezd pro fonetiku," Lidove noviny, July 29, 1938. "Professor Frantisek Trävniöek 50 Jahre," Prager Presse, Aug. 17, 1938. "Vasilij Zlatovlasy, krälevic öeske zemö," Lidove noviny, Aug. 28, 1938. "Fr. Spina jako vedec," Lidove noviny, Sept. 20, 1938. "Cesta sv. Doroty na Rus," Lidove noviny, Feb. 5, 1939. Letter to E. Emsheimer about the etymology of Gilyak names for the Jew'sHarp. Ethnos, Nos. 3-4, Stockholm, 1941, 119 (note 52). "The City of Learning: The Flourishing Period of the Jewish Culture in Medieval Prague," American Hebrew, Dec. 5, 1941. "Cesky pfedek Puskinüv," Obzor, II, London, Feb. 1942. "Oldrich Hujer," New-Yorske listy, June 28, 1942. "Svedsky prispSvek k dSjinäm staroöeske kultury" (A. Baecklund, Das Stockholmer Abecedarium), New-Yorske listy, Oct. 7, 1942.

XXXII

75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92.

93. 94.

95.

BIBLIOGRAPHY

"Fr. Boas, hrdina präce," New-Yorske listy, Dec. 27, 1942. "Russian Books," Rare Books (Published for the Friends and Clients of H. P. Kraus), II, No. 4, New York, 1943, 1-6. Letter to E. Emsheimer about the Gilyak name for the Shaman Drum. Ethnos, Nos. 3-4, Stockholm, 1944, 145. "Two Words on Bywords," The Russians Say It This Way (compiled and edited by D. Gurvitch and A. Herenroth), New York, 1945, 5-6. "Foreword," Taschenwörterbuch der russischen und deutschen Sprache, International University Press, New York, 1945, III-IV. "Foreword," Taschenwörterbuch der spanischen und deutschen Sprache, International University Press, New York, 1945, VII-VIII. Letter to J. Hadamard about internal speech. J. Hadamard, The Psychology of Invention in the Mathematical Field, Princeton, 1945, 96-97. "Study of Armenian History Held Vital," The Armenian Mirror-Spectator, Aug. 21, 1948. "Preface," U. Weinreich, College Yiddish, Yiddish Scientific Institute, New York, 1949, 7-8. "A Handbook of Slavic Studies," Saturday Review of Literature, New York, June 11, 1949, 18, 40. Letter to Gladys A. Reichard about Language and Synesthesia. Word, V, 1949, 226, 230-231. "Foreword," W. E. Harkins, The Russian Folk Epos in Czech Literature, New York, 1951, V-VI. "Preface," K. H. Menges, The Oriental Elements in the Vocabulary of the Oldest Russian Epos, the Igor' Tale, New York, 1951, V-VI. Letter about the original character of the Charles University, Kultura, No. 12/74, Paris, December 1953, 156-157. "Foreword," M. Souckovä, A Literature in Crisis: Czech Literature 1938-1950, New York, 1953, V-IX. "Preface," Serbocroatian Heroic Songs, collected by Milman Parry, edited and translated by Albert B. Lord, I, Cambridge and Belgrade, 1954, XI-XII. "O pozycji i perspektywach wspölczesnej poetyki," Nowa Kultura, X, No. 46, Warsaw, November 15, 1959, 3. Letter to Gordon Wasson about Holger Pedersen's etymology of the fungal name in the Indo-European languages, Botanical Museum Leaflets, Harvard University, XIX, No. 7, 1961, 150. "Amerikanskite slavisti i Sofijskijat kongres," Slavjani, XIX, No. 5, Sofia, 1963, 21-22. "Relevance of Linguistics for Psychoanalysis," Minutes of Study Group in Linguistics and Psychoanalysis, ed. by Drs. Rosen and Edelheit, The New York Psychoanalytic Institute, May 11, 1964, 6 pp. "Excelentä activitate §tiin{ificä," Contemporanul, No. 34, Bucharest, 1964, 14.

BIBLIOGRAPHY

96. 97. 98. 99. 100. 101.

XXXIII

"La linguistique va-t-elle devenir la science des sciences? — Un entretien de Claude Bonnefoy avec Roman Jakobson," Arts, No. 20, Feb. 9-16, 1966, 10-11. "Un incontro con Romàn Jàkobson," interview with Aldo Rossi, L'Approdo Letterario, XII, NS 33, 1966, 93-96. "V Rimé o Praze," interview with A. M. Ripellino, Literârni noviny, XVI, Prague, Feb. 18, 1967, 1, 3. "Questionner Jakobson," Jean Pierre Faye, Le récit hunique, Paris, 1967, 273-285. First version in Lettres françaises, Nov. 30, 1966. "Lingvistica rivoluzionaria e avanguardie russe," Rinascita, No. 12, Mar. 24, 1967, 23-24. "Intervista," Cinema e Film, No. 2, 1967.

ROBERT

ABERNATHY

RHYMES, NON-RHYMES, A N D ANTIRHYME

A

'ACT seldom explicitly brought out (perhaps just because it is taken for granted as obvious) is that the role played by rhyme in the design of verse is determined no less by the distribution of rhymes in a text than by that of its non-rhymes: both positive and negative constraints are operative in poetry. Clearly this has to be so if rhyme, viewed as a paradigmatic relation, is to perform effectively a syntagmatic organizing function. Otherwise, the positive pattern would be subject to obliteration by casual accumulation of structurally irrelevant factors — the signal would tend to be drowned out by the noise.

Before examining some implications of this observation, it will be well to introduce a measure of formalism into the discussion, since terms such as 'rhyme' are commonly employed in diverse and inconsistent ways. Of several possible explications suggested by informal usage,1 the most fruitful approach for present purposes is to consider rhyme as a binary relation R on a vocabulary V (consisting of 'words', used here as a term of convenience without prejudice as to the desirability of specifying V by means of a more sophisticated notion of rhyming 'foci' or the like).2 That is, R is some subset of the Cartesian product V x V (the set of all ordered pairs of words belonging to V). On this basis one can inquire into the logical properties of R as defined by various traditional or theoretical canons of rhyme. A natural starting-point for such investigation is the intuitively more or less plausible conception of R as an equivalence relation, satisfying the requirements of reflexivity, symmetry, and transitivity.3 It can 1 On the diversity of such usage, cf. Arthur Melville Clark and Harold Whitehall, article "Rhyme" in Encyclopedia of Poetry and Poetics, Alex Preminger, ed. (Princeton, New Jersey, 1965), 705-710. 8 For a recent effort at setting up such a 'focus' concept, cf. LfcHKo MjiaaeHOB, "EejiexcKH 3a PHMOBOTO ABHXEHHE B CTHXOBaTa pen", CjiaenncKa UAOAOZUH, 4 (COHH, 1963), 4 0 9 - 4 2 5 . s I.e., R is such that: (1) For all x belonging to V, the pair (x, x) belongs to R. (Reflexivity: Every word rhymes with itself.) (2) If (x, y) belongs to R, then so does (y, x). (Symmetry: Any word which rhymes with another is rhymed with by that other.) (3) If both (x, y) and (y, z) belong to R, then so does (x, z). (Transitivity: Words which rhyme with the same word rhyme with one another.) For these and other terms of relation theory, cf. e.g. Patrick Suppes, Axiomatic Set Theory (Princeton, New Jersey, 1960).

2

ROBERT ABERNATHY

be shown, however, that many rhyme-prescriptions fail one or more of these requirements : thus, for example, Kondratov notes in a recent study 4 that Russian 'inexact rhyme' as he defines it is intransitive, 5 and similar demonstrations can be carried out also in connection with the other two fundamental properties mentioned. The present paper will be particularly concerned, among other things, with the case of rhymerelations which fail of reflexivity. For the special case when R is an equivalence relation, it may be noted also that the partition of V which it induces can be identified with the 'rhyming dictionary' of the language involved, and, conversely, that any such 'dictionary' determines an equivalence relation. But in case R is not of this kind, no 'rhyming dictionary' in the usual sense exists, and it is also worth noticing that in this event the customary algebraic notation for rhyme-schemes (expressions such as abab, abccb, etc.) becomes unusable or usable only with important reservations about its adequacy.® The foregoing considerations can readily be generalized to the case of coexisting rhyme-prescriptions Rj, ..., R n (comprehending, e.g., various relations such as 'exact rhyme', 'assonance', 'near rhyme', 'alliteration', simultaneously admitted but 4

A. M . KoHflpaTOB, "CraTHCTHKa thiiob pyccieoft pnMi.r, Bortpocbi 96-106. 6

HSUKOSHOHUH,

12.6 (1963),

That transitivity does not always hold even for traditionally recognized relations of the kinds called 'rhyme' has on occasion been quite explicitly, if casually, noted by literary scholars. Cf. B. de Selincourt, "Rhyme in English Poetry", Essays and Studies by Members of the English Association, VII, John Bailey, ed. (Oxford, 1921), 7-29 : "There is also a disposition to allow that words which rhyme with the same word rhyme with one another, as sigh and jollity both rhyming with cry. But this only holds under limitations, for sigh and tree, both allowed with jollity, would not be allowed together." Similarly, Ph. Martinon, in the introductory treatise attached to the Larousse Dictionnaire des rimes françaises (Paris, 1962), remarks (p. 50) that, as a consequence of the relaxation of certain constraints of French rhyme in cases where one or both of the rhyming words are monosyllables, "il arrive assez souvent que des polysyllabes qui n'iraient pas ensemble vont avec le même monosyllabe". An intriguing example with implications for the historical dynamics of evolving rhymecanons is pointed out by Roman Jakobson, Studies in Russian Philology (= Michigan Slavic Materials, 1) (Ann Arbor, n.d.): in the development of Russian rhyme at the beginning of the nineteenth century, historical and dialectal factors had created the possibility of rhyming words ending in a morphophonemic {g} with ones ending either in {k} or in {x}, and hence (pp. 9f.) "Ecjih to ace aTHMOJiorHiecKoe Z coieTaercH oaHOBpeMeHHO h c K, h c X . . . to ecrecTBeHHO B03HHKaK>T h Tanne He3aBHCHMLie pn(J)Mbi cjiob Ha x h Ha K, xaK dyx—pyK y JlepMOHTOBa hjih ne.ween—ecex y IlacTepHaKa." The effect of introducing rhymes of the latter kind is to restore transitivity or at least to remove one impediment to it. 6 Consider, for example, the variety of rhyme-schemes for typical quatrains (subject to the condition that each line-end rhyme with at least one other, thus leaving out of account such things as the "Omar Khayyâm" stanza aaba). If R is an equivalence relation, there are just four possible schemes of this kind, which are adequately represented in the algebraic notation and recognized in descriptive nomenclature: aabb (couplets), abab (alternating rhyme), abba (enveloping rhyme), and aaaa (monorhyme). But if the transitivity requirement is abolished, a number of additional possibilities exist: e.g., it may be that the first line-end rhymes with the second, the second with the third, and the third with the fourth, but not the first with the third or fourth nor the second with the fourth. When R is intransitive there are in all 41 distinct rhyme-schemes for quatrains subject to the indicated limitations, and 37 of these are unrepresentable in the standard notation. There is room to wonder, in such a case, how many 'je ne sçai quoi' judgments on the part of critics may reflect merely inadequacies of explicit descriptive techniques for dealing with purely formal devices intuitively understood and applied by poets.

RHYMES, NON-RHYMES, AND ANTIRHYME

3

differentiated as to function or esthetic value within a single system of prosody), which may form coherent wholes, e.g., such a set of relations may be a lattice under inclusion. 7 Now the remark with which this paper began can be made more precise by saying that the organization of verse through rhyme means not only that certain pairs of words (occupying determinate positions with respect, say, to a fixed metrical pattern : for purposes of illustration here I will use mostly examples uncomplicated by scansion problems) are members of R, but also that other, likewise locatable, pairs belong to its complement (V x V) — R, i.e., the set of non-rhymes of the given vocabulary with respect to the given rhyme-prescription. Some very strict 'rhyming' verse forms are at the same time obligatorily not 'rhyming' with reference to specific positions: for example, in Chinese Regulated Verse in one of its varieties, the stanza consists of four five-syllable lines in which the second and fourth line-ends must rhyme with one another and just as definitely must not rhyme with the third line-end, and moreover, by virtue of the traditional tone-pattern constraints, rhyme is excluded e.g. between the second and fourth, the second and nineteenth, or the fourth and ninth syllables (characters) — in all, this stanza is required to contain 19 non-rhymes as against a single rhyme, so that its description as a 'rhyming' form seems curious, in a sense.8 In many cases, where metric limitations are more flexible, it can be said only that there is a high probability that given pairs will be non-rhymes, so that exceptions to the 'rule' have a high surprisal value and tend to focus attention on themselves as incidental ornament. A largely neglected corollary of these considerations is that, for verse of kinds commonly and ambiguously styled 'unrhymed' (e.g., English blank verse, Old English alliterative and Biblical Hebrew parallelistic poetry, Old Russian, classical Greek and Roman forms), a distinction should be drawn between what might be called, for this distinction's sake, simply 'rhymeless' verse — meaning that to which no rhyme-canon is systematically relevant 9 — and 'antirhymed' verse, characterized primarily by negative constraints (e.g., by the requirement that adjacent line-end pairs belong obligatorily to the complement of R, for some definite non-trivial rhyme-prescription 7

For an elementary introduction to lattices (one of the important unifying concepts of modern mathematics), stressing graphic presentation, see now Roy Dubisch, Lattices to Logic (New York, 1964). A lattice is, roughly speaking, a 'hierarchy' of elements of any kind. Instances of classifications of coexisting rhyme-canons which are or can easily be construed as systems of this kind are offered by Stephen L. Mooney, "New Devices in Sound Repetition", Word Study, 24.4 (1949), 1-4, or in the treatment of "wspólbrzmienia" elaborated on the basis of the ideas of Kazimierz Nitsch by Mieczyslaw Giergielewicz, Rym i wiersz (Londyn, 1957). Detailed examination of questions of this order is matter for a separate study. 8 For specifications of the Chinese Regulated stanza, cf. James J. Y. Liu, The Art of Chinese Poetry (Chicago, 1962), and, for a succinct formal characterization of these patterns, now also Roman Jakobson, "The Prosodie Design of Chinese Regulated Verse", forthcoming in Mélanges Claude Lévi-Strauss. • More formally, this notion of 'rhymelessness* can be brought into the relation-theoretic framework of explication here employed by assuming that the applicable R is the trivial one by which any word rhymes with any other word.

4

ROBERT ABERNATHY

R). What tends to be overlooked is that a composition of the latter kind presupposes, no less than the most elaborate rhyme-scheme, the existence of an accepted canon of rhyme with respect to which the poem has its form. One might anticipate, for instance (and research seems to bear the expectation out), that Victorian blank verse will be 'unrhymed' — i.e., antirhymed — precisely in terms of the rules which typically apply to Victorian rhymed verse. Conversely, it makes very questionable sense e.g. to sift (as modern scholars have done) the half-line ends of the Old English Beowulf in order to catalog pairs which are rhymes by 19th- or 20th-century standards.10 In critical literature it is quite exceptional to find the above-noted ambiguity of the 'unrhymed' label resolved, even by indirection. One case in which a writer has come to grips with the question can be found in Lowell's 1872 essay on Milton,11 where Lowell remarks with reference to the versification of Paradise lost that "Milton's ear has tolerated not a few perfectly rhyming couplets, and others in which the assonance almost becomes rhyme, certainly a fault in blank verse ..." and concludes that such instances "serve to show that Milton's ear was too busy about the larger interests of his measures to be always careful of the lesser. He was a strategist rather than a drill-sergeant in verse. ..." The factual observation is correct, and can be supported by numerous examples besides those Lowell cites.12 But the conclusion is askew — it rests on the gratuitous assumption that when Milton, in the language of his famous and often-quoted preface, set about writing "English Heroic Verse without Rime", he in fact undertook to compose Victorian antirhymed blank verse, an enterprise at which he succeeded only indifferently. On the face of the evidence it would seem more reasonable to conclude that Paradise lost is really just rhymeless. But Lowell should be credited,13 at all events, with making explicit a matter too often left tacit, and he moreover in effect raises (with the comment about 'assonances' which are uncomfortably close to rhyme) a question worthy of further investigation, viz., that of avoidance or non-avoidance of inexact rhymes of various kinds in antirhymed verse. As an example of a large-scale composition which seems to be clearly antirhymed, Book VII of Browning's The ring and the book exhibits the 'rhyme scheme' a 1 a 2 ...ij 1828 , subject just to the constraint a\ # ai+i: that is, of the 1828 lines no two adjacent 10

There seem to be less than a score of passable examples in some 6000 half-lines; the question which should really be put to this material is not whether the Beowulf poet was using end-rhyme (in some sense close to or remote from its Victorian definitions), but rather whether he was making any determined attempt to avoid it. 11 James Russell Lowell, Among My Books, III ( = Lowell's Works, V) (Boston-New York, 1904), 243-312. 12 Cf., e.g., Paradise lost 2:220f. (light, flight), 2:709f. (right, might), ll:597f. (express'd, blest). 13 One more recent writer who makes the point at issue explicit (following Lowell) is C. E. Andrews, The Writing and Reading of Verse (New York-London, 1934). But the example which Andrews gives (pp. 215 f.) of what he takes to be "unintentional introduction of rime or assonance" in blank verse is not convincing, since what he regards as exact rhymes and assonances are not such at all by the strict canons of English versification (e.g., he calls two lines "perfectly rimed" which are linked by the matching (thee, memory) — a mere courtesy-rhyme!). Here, as it so often falls out, it seems that the critic, as critic, is wrong, and the poet, as poet, right — which implies nothing about what may happen if they change places.

RHYMES, NON-RHYMES, AND ANTIRHYME

5

lines rhyme, with respect namely to the usual canons of exact rhyme in English. On the other hand, it is easy to show that this inhibition does not extend over longer intervals — cf., e.g., rhymes with only one intervening line: (air, prayer) in lines 611-3, (deceit, meet) 766-8, (die, sly) 1192-4, (fly, die) 1235-7, (me, he) 1257-9, across two intervening lines (I, untie) in lines 514-7, (true, review) 627-30, (review, true) 630-3, (ear, austere) 1001-4, etc. Neither is there any conspicuous tendency to avoid inexact rhymes even between adjacent lines, cf. (apart, are) 85f., (almost, last) 738f., (inside, mankind) 852f., (stare, stage) 950f. The immediately evident pattern of non-rhymes here is very simple, but it remains possible that more thoroughgoing analysis would elicit less obvious but still structurally significant features or at least statistically demonstrable proclivities, either positive or negative in terms of relevant rhymecanons. That either sort of constraints can be important is the fundamental point at issue here. More complex and delicate problems are posed by mixed forms which oscillate between deliberate use of rhyme and its just as deliberate avoidance. A classic example is the Elizabethan dramatists' assignment of specialized functions to rhyming ornament in predominantly blank-verse texts: painstaking studies have been carried out, in particular, of Shakespeare's use of such devices (the rhyming couplet(s) as a 'closing bell' signaling scene-changes, etc.). 14 Such investigations ought in principle to be supplemented by methodical analysis of the author's use of non-rhymes: a major obstacle is, of course, the sheer bulk of material requiring to be examined, but modern data-processing techniques promise to overcome such difficulties, and the work might well be repaid by new light upon some of the many textual, critical, and linguistic problems outstanding in connection with the Shakespearean corpus. Reciprocal interest of the literary historian and the linguist in these matters is evident — rhymes, certainly, constitute where available a valuable means to reconstruction of earlier stages in the phonology of a language, and conversely knowledge about pronunciation is needed to decide when problematic matchings are or are not rhymes. 15 In this connection it is interesting to note that in Puskin's deliberately 'Shakespearean' drama Boris Godunov the Elizabethan mixed versification has been adapted to a Russian climate, but in the process has undergone noteworthy modifications of 14 Cf. Frederic W. Ness, The Use of Rhyme in Shakespeare's Plays (New Haven-London, 1941), who distinguishes nine functional categories of such use, plus a small but annoying residue of 'unintentional' cases. The 'closing bell' phrase is adopted from John Ciardi ("Manner of Speaking", Saturday Review for October 31, 1964, pp. 16, 18), who suggests as a general principle in this connection that "form tends to conclude itself by some increase in formality". 16 An outstanding paradigm of the cooperation needed between the sometimes unduly compartmented disciplines concerned with language and literature was given early in this century by the integrally related development in Russia of Formalist poetics and structural linguistics (on which cf. Victor Erlich, Russian Formalism: History — Doctrine (= Slavistic Printings and Reprintings, IV) (The Hague, Mouton, 1955), and carried forward since in the works of Roman Jakobson, whose assessment of the results one can scarcely do better than to quote in this connection: by the 1960s "a linguist deaf to the poetic function of language and a literary scholar indifferent to linguistic problems and unconversant with linguistic methods are equally flagrant anachronisms" ("Linguistics and Poetics", in Style in Language, Thomas A. Sebeok, ed. (MIT-Wiley, 1960), 350-377).

6

ROBERT ABERNATHY

detail: the role which in Shakespeare almost always belongs to one or more couplets (e.g., at scene ends) is by Puskin usually assigned rather to alternating quatrains, a practice rare though not wholly unknown in Shakespeare.16 This looks like a case of 'dialect borrowing' in a sense which Jakobson has suggested.17 A really complicated and intriguingly nuanced example of the interplay of relations covering a whole spectrum from exact rhyme to non-rhyme (cf. remarks above on systems of coexisting canons) is offered by Blok's curious 'dream' poem "HoHHaa 3>HajiKa". This is a work of just over 300 lines which, unlike most of Blok's poetry, fails to exhibit any simple and regular stanzaic pattern or any prevailing rhymescheme. Kemball18 analyzes it as consisting of three main parts: lines 1-44, which are "without rhyme" and metrically "mixed", lines 45-287 identified as "anisometric non-rhyming anapaests", and finally lines 288-304, forming a rhymed finale (iababcdcdejfecgcgc, where moreover b = d by the canons of inexact rhyme frequently observed by Blok) which also closely approaches complete metrical regularity. Such a lapidary description (Kemball, to be sure, treats rhyme-schemes only incidentally to his primary interest in meter) falls far short of doing justice to the subject — is, indeed, even inaccurate on its own terms, since there are sporadic occurrences of exact rhymes before the 17-line coda mentioned. But relations weaker than exact rhyme are also unmistakably significant: viewed as a whole, the poem displays a remarkable progression from what could here appropriately be called, with 2irmunskij, 19 "embryonic rhyme", through successive stages of approximation until a definite exact-rhyme scheme crystallizes in the end. This evolution is interwoven with that of the metrical design. If, with a somewhat arbitrary precision, we take to be significant just relations between line-finals separated by 0, 1, or 2 lines (as suggested by the coda), then the poem has — discounting lines 1-5, which form a species of preface — an initial portion quite saturated with line-end assonances, though devoid of exact rhymes and metrically quite uneven: cf. the abab assonancescheme of lines 29-32, or the more striking aabb of lines 35-38: H ocTaBHji MeHfl oflHoro (*IeM H 6MJI HecKa3aHH0 AOBOJieH, H6o HTO ace NPHHTHEN Ha CBeTe,

HeM yTpaTa jiyniiiHx jrpy3eH?).

" Cf. Ness, op. cit., p. 72. " P. .HKO6COH, Hoeeutuanpyccnan nos3un, Ha6pocoK nepBbifi: Bunmop XM6HUKOS (Ilpara, 1921), 5-6: "Pa3BHTHe TeopHH nosTHHecxoro «Mica 6y.neT BO3MO»CHO jrann. Toraa, Korjja no33iw 6y«eT TpaKTOBaTbc«, KaK coimajn>Hhifi aKT, Kor.ua 6yner co3#aHa cBoero po.ua nosTmecKa« fluajieKToJioriw." Jakobson goes on to suggest parallelisms between the borrowing, assimilation and adaptation of poetic devices and comparable facts about the propagation of dialect features. If the example here adduced is correctly interpreted, it serves also to show that neither the geographic nor the linguistic distance between English and Russian is sufficient to constitute a barrier to this kind of 'dialect borrowing'. 18 Robin Kemball, Alexander Blok: A Study in Rhythm and Metre (The Hague, Mouton, 1965), 413, 420, 441 f. 18 B. SCnpMyHCKHfi, Pu$Ma, ee ucmopun u meopiut (= Bonpocu no3muKu, 3) (IleTporpafl, 1923), 20. Zirmunskij, to be sure, uses this term with a historical interpretation in mind, and with reference more to optional vi. obligatory than to weaker vs. stronger nature of rhyme.

7

RHYMES, NON-RHYMES, AND ANTIRHYME

(Note the 'rich' assonance of the a lines, where congruence in the three-phoneme sequence /avo/ would, if word-final in both cases, be more than sufficient to determine an exact rhyme.) Thereafter, the establishment of a more consistent metrical pattern, identified by Kemball as beginning with line 45, coincides quite closely with the abrupt and almost total disappearance of assonances for about twenty lines — it is as if, at this point, one kind of structure had been supplanted by another. 20 Next, stressed-vowel echoes begin to reassert themselves, beginning with ( C T O J I H U H , 3 K y p < J ) H K C O B ) in lines 74-76 and (ao6pe, 6ece,a) in 77-79, and become rapidly more insistent, until in lines 102-108 we have, in terms of assonances now isosyllabic and alternating between masculine and feminine line-ends with unstressed-vowel agreement also in the latter case — thus by a stricter rhyme-canon than before — the pattern abcbcdc: B

¿YIHHHOH,

HH3KOH

H36E

n o

CTEHAM

HeyKJHoacne J I E B K H C T O 5 U I H . Ha oflHoii — nepefl A J I H H H M M C T O J I O M MojinaJiHBo CHflejia 3a npaaceft, OnycTHB Haa pa6oToft npoSop, HeKpacHBaa fleByuiKa C

HENPHMETHBIM

—

JIHUOM.

(Here one might note also that the first and third lines are linked by consonantrepetition of a kind studied by Brik 21 in Russian poetry, resembling the cynghanedd groes of Welsh bardic tradition, and that the assonance of the first and last c lines is in fact an exact rhyme.) Meanwhile, poor rhymes (masculine rhymes which would qualify as exact in English, but do not in Russian) have put in their appearance, first two lines apart: (rjia3a, a) in 86-89, then one line apart: (ce6a, aoayw) in 95-97, and the passage quoted above is immediately followed by one which commences with a poor-rhyme couplet (109f.) that is also in /a/: 51 HE 3 H a i o , 6biJia JIH OHa

Mojio,na M b crapa, and which is dominated throughout (lines 109-123) by just such rhymes and by feminine assonances likewise congruent in /a/. This buildup of progressively more formal structure continues, with some fluctuations and reversions to stages earlier reached (cf. in lines 182-187 the assonancepattern ababab, where the first and third a lines form an exact rhyme). Alternating exact rhyme first emerges with lines 214-216: (cica3aHHH, T K a H e S ) . Finally in 240-249 " One might regard the persistent lack of assonances (over a span of four lines) as reflecting a negative constraint. It should be kept in mind that the intrinsic value of stressed-vowel echoes as such in Russian is not very high, since any text of 6 or more lines must contain at least one such echo in line-finals. 81 Ocm M A K C H M O B H H E P H K , "3ByKOBwe noBTopw (aHajiro 3ByxoBofi crpyKTypu cnixa)", in Two Essays on Poetic Language (= Michigan Slavic Materials, 5) (Ann Arbor, 1964), 1-45 (original publication: 1917).

8

ROBERT ABERNATHY

a sequence of assonances in /6/ culminates with an exactly rhyming abba quatrain in completely regular anapestic trimeter: flajibiue, a a j i b m e — 6e33ByiHO n p j w r , H npa^eT, H npiifleT KopoJieBHa, OnycTHB Haa pa6oToS npoôop, Cjia^KHM C H O M oaypMaHHjia Hac, OnoHJia Hac 3ejibeM 6 O J I O T H M M , OKpy»HJia Hac C K A 3 K O I I H O H H O H , A caMa Bcë U B C T Ê T H UBerer, H 6oJiOTaMH flbmiHT OnajiKa, H 6e33ByHHaa KpyxcHTC« npajiKa, H n p j w r , H npa^eT, H npjmeT.

This is, I think, the end of the poem. Since there are, after all, 55 more lines, this remark calls for justification. First, it is clear that with this passage the process of increasing formalization has reached its natural climax with the assertion of a regular rhyme-scheme and meter: in these respects, the remainder of the composition does no more than recapitulate this evolution after a fashion and provide a longer rhyming conclusion (four quatrains plus an 'odd' line to the last of them, echoing significantly the last line of the passage above quoted (H Ho«mafl OnajiKa UBeTeT.)), but adding nothing particularly new. The definitive air possessed by 240-249 is moreover enhanced by their relentless insistence on certain word- and phoneme-repetitions,22 especially those of the sequence /ot/ which here (repeated eleven times!) functions unmistakably as a sort of keynote,23 echoing the theme-word 6 O J I O T O and with it the motif of the 'swamp', at once depressing and poisonously alluring, which dominates this and other poems composed by Blok during 1905.24 Naturally enough the concluding rhyme is congruent in /6t/. So much for internal evidence, though of course much more could be said on this subject. In the second place, we happen to be unusually well-informed about the " Note also that line 242 is a repetition of line 106 in the subclimactic passage earlier quoted. This line has moreover been identified as a 'reminiscence* of one in a poem by Fet, cf. AiieKcannp E J I O K , CoôpaHue covuneuuu e eocbMu moMax, 2 (M.-JL, I960), 388. But it has been reshaped to fit the assonance-pattern of its context in the present composition. " Elsewhere, in connection with another work by this same poet, I have discussed the possibility of temporary semanticization of small units, such as even single phonemes, within the framework of a particular poem on the pars pro toto principle, via their introduction in affectively charged and strategically placed key words ("A Vowel Fugue in Blok", International Journal of Slavic Linguistics and Poetics, VII (1963), 88-107). The present example seems to furnish an excellent illustration of this point. " On the 'swamp' theme in Blok's writings of this period cf. Sophie Bonneau, L'univers poétique d'Alexandre Blok (= Bibliothèque russe de l'Institut d'Études Slaves, XX) (Paris, 1946), 219-221. Its comparison with a Bunyanesque Slough of Despond (F. D. Reeve, Aleksandr Blok: Between Image and Idea (New York-London, 1962), 80) is as misleading as it is obvious — Blok's swamp is quite a different region. One ingredient in its 'poison', alluded to in this poem, is certainly alcohol — the word used (3ejibe) has this meaning in a jocular sense, and Blok's contemporary notes leave small doubt that the dream which he eventually fashioned into a poem originally was part of a wretched hangover (cf. reference in fn. 22, pp. 385-388).

RHYMES, NON-RHYMES, AND ANTIRHYME

9

origins of this poem: its subtitle "A dream" is meant literally, since by Blok's own testimony it is based on an actual dream he experienced in November 1905 and on his written notes made a day and a half later, still under the influence of the powerful dream-impression. This preliminary sketch of the poem survives, and comparison of it with the final published text makes plain that the latter is directly founded on this 'dream-stuff' only as far as line 249. Lines 250-304 lack such antecedents, and in their content reflect quite a different mood, dispelling the nightmare atmosphere of the swamp (dawn breaks, birds cry, ships come into port, "Joy arrives"; the gloomy Viking warrior's panoply crumbles and a "cheerful little mouse" scampers out of the wreckage; the forbidding "old man" and "old woman" of the dream are no longer king and queen; and the Night Violet herself, earlier smelling ominously of the "swamps" and their "poison", is subtly transformed into a symbol of hopeful constancy). It is reasonable to see in this a product of the artistic reworking, completed only in May 1906, of the original dream-nucleus. What is interesting from the present standpoint is that, from a formal analysis such as that sketched above, this 'nucleus' appears as a recognizable structural whole, embedded in a framework — which, for the rest, reproduces the nuclear pattern in such a way that the same charaterization of total effect, as one of form gradually emerging from chaos,25 is applicable to both and valid for this work without reference to the conclusions of the foregoing discussion. Both form and chaos are relative: earlier portions of the poem, which have definite 'rhyme-schemes', e.g., with respect to the weak requirements of assonance, are 'formless' in terms of the stricter requirements imposed later. The compositional use made of a distinction between rhymes and non-rhymes is, broadly speaking, similar to that noted in connection with Elizabethan drama, but apparently much more complex, in that Blok is employing not just a single relation R and its complement, but a whole hierarchy or sliding scale of relations Rx, ..., R n . Another reason for investigating the rhyme-structure of so-called 'unrhymed' verse springs from the observation that rhyme-canons may differ, not just as to their specifications of detail and relative laxity or stringency, but also with regard to basic relational properties. In particular, a given R may or may not contain word-pairs of the form (x, x) ('identical rhymes', 'tautological rhymes', or the like). If it does contain all such pairs, it is reflexive and may be an equivalence relation. The " Given the poem's dream-ancestry, it is tempting though probably fanciful to try to correlate this progressive regularization of its form more or less directly with features of hypnoidal (hypnopompic) states — say with the behavior of electroencephalograph tracings during gradual awakening, showing replacement of the slow irregular delta waves of deep sleep first by more or less random bursts of activity and finally by the waking alpha rhythm. On a quite different level of analogizing, cf. Kornej Cukovskij's characterization, in his memoirs on Blok, of the poet's development during the relevant period (around 1905) as a sort of slow and difficult awakening from the 'dreaminess' of his earlier style (K. HyKOBCKHfi, AneKcaudp EAOK, KOK nejioeeK u nosm (Beedenue e no33u>o EAOKQ) (IleTporpafl, 1924), 106). It is not really clear that one analogy is worth much more than another.

10

ROBERT ABERNATHY

assumption that reflexivity holds is often made tacitly and on occasion even explicitly — so by Kondratov (op. cit.), who uses expressly relation-theoretic language: "H TOHHbie h HeTOHHbie pwjiMM [in his classificatory system for Russian rhymes] oGjiaflaioT CBOHCTBOM pecjuieKTHBHocTH (SJIEMEHT TOACAECTBEH ce6e)...". Cf. LoS,28 who remarks (p. 52) in connection with a discussion of the historical origins of i hyme in repetitious parallelisms of Slavic folk poetry: "Naturalnie, w tych wypadkach rymujq z sob^ identyczne wyrazy...". But this 'naturally' is unwarranted: on the contrary, a feature of many rhyme-canons is precisely the outlawing of 'identical rhyme', i.e., the relegation of the pairs (x, x) to the domain of non-rhymes or at all events to a status distinct from and possibly inferior to that of exact rhymes in a system of coexisting relations. In this event, not only is R non-reflexive, it is actually irreflexive and by the same token certainly not an equivalence (though it may stand in a readily statable relationship to an equivalence relation, say R* such that R = R* — I, where I is just the set of all pairs (x, x)). R may also fail of reflexivity by virtue not of exclusion of 'identical rhymes' as such but rather of its imposition of formal requirements which these pairs cannot satisfy. This is the case, for example, of English exact rhyme when it is defined, as it often is, to require — in a fashion sharply at variance with the 'supporting consonant' clauses typical, e.g., of French or Russian rhyme-definitions — 'non-support' of the stressed vowel by preceding consonants.27 Similar considerations, with the roles of vowels and consonants and of right and left directions interchanged, apply to Old English and other early Germanic alliterations.88 Irreflexive for quite different reasons is the traditional endrhyme of the Welsh cywydd couplet: here one of the matching words must be a monosyllable, the other a polysyllable, so that 'identical rhymes' (and, given the stress-pattern of Welsh, even isotonic ones) are clearly out of the question.29 The same conclusion follows from the kindred specifications of the Irish deibhidhe rhyme (which moreover, when the order of arguments of R is interpreted as order of occurrence of the rhyme-words in the text, fails also of symmetry, whereas cywydd rhyme is at any rate symmetric).30 " Jan Los, Wiersze polskie w ich dziejowym rozwoju (Krak6w, n.d.). " Cf., e.g., Edward A. Bloom, Charles H. Philbrick, and Elmer M. Blistein, The Order of Poetry: An Introduction (New York, 1961), 119: "Full rhyme consists in its simplest form of monosyllabic words which begin with different consonants and end in identical vocalic and consonantal sounds. To put it more generally, words which rhyme begin differently and end the same, exemplifying in sound the principle of similitude in dissimilitude..." " As to the vowel 'dissimilation' in this case, which automatically rules out the alliteration of a word with itself, cf. Roman Jakobson, "On the So-called Vowel Alliteration in Germanic Verse", Zeitschrift fur Phonetik, Sprachwissenschaft und Kommunikationsforschung, 16 (1963), 85-94 (with statistical appendix by Sigrid Valfells). The weakening of this dissimilative principle in later Germanic poetry, as noted here p. 88, may then well lead to the replacement of the originally irreflexive relation by a reflexive one; and in fact we find, e.g., in the Middle English verse account of the Destruction of Troy, Prologue line 56, the identical alliteration (batell, batell). 28 Cf. Thomas Parry, A History of Welsh Literature (Oxford, 1955), 124f. (in appendix on versification by H. Idris Bell). 80 For example, in deibhidhe rhyme the pair (rinn, airdrinn) is a rhyme, but (airdrinn, rinn) is a

RHYMES, NON-RHYMES, AND ANTIRHYME

11

Handbooks of versification devote a good deal of discussion to 'identical rhyme' (without, as a rule, separating this from the logically distinct question of homonymous or equivocal rhymes, i.e., that of the relevance of semantic features to determining whether an 'identity' holds in the first place). Nevertheless, one often runs across evidence of uncertainty about the treatment of this case. A curious example of the sort can be seen in conflicting descriptions given of the sestina, which — being a verse-form in which the occurrence of pairs (x, x) in certain positions is not merely admissible but obligatory — constitutes a species of test case for reflexivity: the sestina is sometimes said to 'rhyme' and sometimes not. 31 The latter account of it is surely preferable, since in this form (including its rhyming variants, as practised, e.g., notably by Swinburne in English) there is no equation of identical repetition to rhyme — just as in the common case of refrains in rhyming stanzas, the two devices play altogether different functional roles, and the sestina should be regarded as a special form of antirhymed verse. Irreflexivity is, indeed, so commonly characteristic of the rhyme-canons applied in practice that the point should be made that 'identical rhymes' CAN be rhymes, though clear cases are harder to find than might be casually supposed. 32 Zirmunskij cites33 some convincing examples from Russian verse in which pairs (x, x) have to be taken a s r h y m e s , e.g., (MCHH, MCHH) f r o m L e r m o n t o v ' s "TPOCTHHK", a n d Giergielewicz,

distinguishing clearly among repetitions, identical rhymes, and homonymous rhymes, notes as an example of the second kind (zar?czynach, zar?czynach) from Mickiewicz's Pan Tadeusz, very properly insisting on the point "ze na tie catosci poematu dwuwiersz ma postac zblizonq. do innych ogniw utworu, sk^d wynika, ze 'powtorzeniom' przypadla w udziale funkcja rymotworcza". 34 Frivolous employment of the device occurs, e.g., in the well-known limerick about the young lady from Riga, with a

non-rhyme, it being understood that the order in which these pairs are given is that in which their members appear in the text. Cf. Eleanor Knott, An Introduction to Irish Syllabic Poetry of the Period 1200-1600, 2nd ed. (Dublin, 1957), 18-20. 81 E.g., for W. K. Wimsatt, Jr. (The Verbal Icon (Louisville, 1954), 156) the very existence of this form is proof that "Even identical words may rhyme. In the sestina, for example, the same set of rhyme words is repeated in six different stanzas." But according to the unsigned article "Sestina" in the Encyclopedia of Poetry and Poetics (ref. fn. 1), "It is composed of 6 stanzas of 6 lines each, followed by an envoy of 3 lines, all of which are usually unrhymed." On the history of this peculiar form, which at times sheds a certain oblique light upon the problems of identical and equivocal rhyme, cf. Wilmon Brewer, Sonnets and Sestinas (Boston, 1937). Abstractly it amounts simply to an exercise in permutations; the possible superimposition of rhyme, in the ordinary sense, on this pattern is more or less irrelevant. A seventeenth-century German writer, Quirinus Kuhlmann, is said even to have produced a sestina in monorhyme (so according to Alfred Liede, Dichtung als Spiel: Studien zur Unsinnspoesie an den Grenzen der Sprache, II (Berlin, 1963), 141). 32 Cf. Henry Lanz, The Physical Basis of Rime: An Essay on the Aesthetics of Sound (Stanford University, 1931), 100-103, for some discussion of the problem with English examples. Lanz' preoccupation with explanations in terms of psychophysical absolutes precludes his coming properly to grips with the purely conventional aspects of this problem, however. 33 Op. cit., 87-91 (ref. fn. 19). 34 Op. cit., 21 (ref. fn. 7).

12

ROBERT ABERNATHY

rhyme (tiger, tiger). A striking case of free use of identical rhyme can be seen in Moritz Graf Strachwitz' Das Herz von Douglas, eighteenth stanza: Zehntausend Lanzen funkelten rechts, Zehntausend schimmerten links; "Allah il Allah"! scholl es rechts, "II Allah"! scholl es links. Since the whole composition is in abab quatrains, it is clear that the identical line-end matchings here must be evaluated as rhymes. (There is also an identical rhyme in the a lines of the sixth stanza.) In such instances, the motivation, in terms of striving for special effect, is frequently clear, but nothing follows from this about rhymepatterns — the rhetorical motivations may be operative whether pairs (x, x) are rhymes or non-rhymes. Of more interest here again is what seems to be the neglected corollary: in antirhymed verse we should find, in case the underlying rhyme-prescription is irreflexive, pairs of the form (x, x) admitted in positions where rhymes are excluded. This expectation is in fact borne out by the previously adduced example of Browning's The ring and the book (Book VII), which, as has been seen, is constructed on the principle that adjacent lines do not rhyme: in lines 912f., however, occurs the 'rhyming' couplet: So we are made, such difference in minds, Such difference too in eyes that see the minds! And similar matchings appear in 335f.: (pain, pain), 625f.: (forgive, forgive), and 1582f.: (bear, bear). It is obvious that, for the purposes of this 'blank' — i.e., antirhymed — verse, these pairs are non-rhymes, and in the light of this and the observations made earlier, it is possible to determine quite precisely the kind of rhyme with respect to which this work is 'unrhymed': viz. English exact rhyme defined to be irreflexive.35 The blank verse of the same author's Fra Lippo Lippi exhibits the same pattern (discounting the poem's parenthetical rhyming interludes): in lines 187f. we have (soul, soul), in 240f. (know, know); and many other examples of this sort could be found. 36 In this connection it is also interesting to note that, in Lowell's criticism of Milton cited above, the specimens given of Milton's supposed transgressions against the rules of English blank verse include cases of 'identical rhyme' — which Lowell is thus assuming to be rhymes in such verse, so that he is operating with a definition sharply 86 All twelve parts of The ring and the book are on essentially the same pattern as that selected for illustration here, but in some of the others there are exceptions requiring special explanation (cf. notably the use of rhyme at the beginning of Book VIII — apparently as one indication of the Procurator's tendency to let his language run away with his meaning). *• Cf., e.g., from an exceptionally rhyme-conscious American poet, Poe's To Helen (66 lines of blank iambic pentameter — not to be confused with the better-known rhyming poem of the same title), which contains two 'identical rhymes' between adjacent line-ends.

RHYMES, NON-RHYMES, AND ANTIRHYME

13

at variance with that evident in Browning's practice. (As for Milton, either definition is presumably irrelevant to his work.) Some even more striking observations can be made on some other antirhymed verse-forms. In Swinburne's Sapphics there would be, under an equivalence definition of English exact rhyme, 13 such line-final rhymes in a mere 80 lines of verse (counting all possible matchings of line-ends and not just adjacent ones). But all the cases involved are of the form (x, x) — e.g., the name "Aphrodite" occurs four times in line-final position, which means that it alone accounts for six of the matchings in question. When these cases are relegated to the non-rhyme category, the poem is seen to be completely antirhymed, in the sense that no two lines anywhere in it rhyme. A larger-scale composition answering to the same description is Browning's One word more, where in 201 lines examination of all possible line-end pairings turns up no less than 52 'rhymes' — enough to provide for a poem of the same length in abcb quatrains (ballad stanza), so that it might be said that all that keeps this from being a 'rhyming' form is a matter of arrangement. However, all but three of the putative rhymes are (x, x) pairs, and the remaining three are of the inclusive kind (e.g., (immortal, mortal) in lines 76 and 158) which are likewise ruled out by definitions of English irreflexive rhyme along the lines remarked on above. 37 It would thus appear that the poet has maintained a studied avoidance of line-end rhyme over a span of 200 lines — a span much greater than that involved in any recognized pattern of positive constraints. In tentatively attaching significance to such findings, due weight must of course be given to the fact that non-rhymes far outnumber rhymes in general, and for the meters of the last two examples (Sapphics and trochaic pentameter) are more abundantly available than for the iambic pentameter of the previous examples showing only short-range negative constraints; but on the other hand one can easily fail to appreciate the significance of the fact that, for any finite vocabulary V, there is a fixed upper limit — fixed namely by the number of elements in the partition of V by R — on how long a text can become and avoid rhyming altogether. 38 Against the background of the theoretical considerations brought out above, it could be predicted that there will be quite real, if elusive, differences between antirhymed verse like that of these English examples and, e.g., the (relatively) antirhymed portions of Blok's "HonHaa OnajiKa" (which, as has been seen, runs the gamut from deliberate avoidance of exact rhyme in the beginning to its equally deliberate employment in the end). As Zirmunskij points out, 39 Blok makes rather free use, in the 87 Indeed, cases of this kind (tantamount to the 'riihrender Reim' highly valued in German verse of certain periods) are in discussions of English versification sometimes lumped indiscriminately as 'identical rhyme' with the cases more properly so-called: cf., e.g., de Selincourt, op. cit., 12f. (ref. fn. 5). 38 Note also that, when rhyme and meter are related to one another in usual ways, the imposition of a fixed metric scheme sharply increases the difficulty of avoiding rhyme, by allowing a word to be paired only with members of a subset of V which excludes most of the words with which it does not rhyme but includes all those with which it does. " Op. cit., 89f. (ref. fn. 19).

14

ROBERT ABERNATHY

typical rhyming forms of which most of his work consists, of 'tautological' (identical) rhymes. It would seem to follow that these pairs (x, x) will be unavailable to him as non-rhymes, and will not occur in adjacent line-ends, at least, in his antirhymed verse. This is so in fact: there is a noteworthy absence of such matchings, both in the work already considered at length 40 and in other of Blok's ventures into antirhyme. Cf. in particular the four poems in fairly regular iambic pentameter grouped under the title "BojibHwe mmcjih", aggregating about 350 lines — not, to be sure, a great deal of material, but enough that, on the comparative evidence, a like amount of English blank verse would be apt to betray the tendency to exploitation of 'identical rhymes' as non-rhymes (cf., e.g., Fra Lippo Lippi, which in 379 lines of blank verse has two such instances!). 41 Thus there is reason to believe not only that Blok's rhyme-canon differs for positive structural purposes, in a logically fundamental way (by virtue of reflexivity vs. irreflexivity of the principal relations involved), from ones commonly reflected in English poetry, but also that when Blok does n o t rhyme, that from which he abstains is not at all the same thing as that from which English poets frequently abstain. The argumentum e silentio, which necessarily figures in some of the preceding discussions, is one from which it is intrinsically difficult to obtain hard-and-fast conclusions. However, this paper will have achieved its aim if it has made the point that negative evidence cannot in principle be ignored, and that various familiar notions about verse structure in terms of rhyme have, in this direction, corollaries which tend either to be overlooked because a proper logical perspective is lacking, or neglected because of the bulk of material needed for their empirical investigation. Perhaps it is not too much to hope that exploration of these questions may shed a clearer light on subtleties of poetic form which, in critical literature, are all too frequently dismissed with vague allusions to ineffable somethings dimly seen. UNIVERSITY OF COLORADO 40

Nor are there any such figuring as rhymes in the limited exactly rhyming portions of this work. In lines 300-302 occurs (npjweT, npimeT), but this is at most an equivocal (homonymous) rhyme. 41 The small proportion of Blok's writings not in definite rhymed forms imposes an inescapable limit on the force of inferences which can be drawn, but there may be enough evidence to justify making the account of his antirhyme technique a little more precise: namely, in these poems the avoidance of rhyme seems to extend (as in the case of Browning's The ring and the book) only to adjacent lines. In ">Kn3in> M o e r o npHHTejia", part 6, lines 17-19, the identical rhyme (ctoji, ctoji) occurs across one intervening line, and in "B ceBepHOM Mope" lines 60-62 similarly the non-identical rhyme (rjHWHT, roBopirr) and in lines 46-48 the equivocal (Bexn, Bexii) (nom.pl. in one instance and gen.sg. in the other). This latter composition also includes an incidental couplet (lines 27 f.), which however seems to be specially motivated (quotation!). One further word of caution is indicated: observations of the kind cited fall short of authorizing sweeping generalizations to the effect, say, that English rhyme is irreflexive, Russian rhyme reflexive, or even that Blok's practice of the latter is monolithic in this respect. On the contrary, a few instances can be found in which this author seems to use identity matchings as functional non-rhymes, though the case is difficult to prove. One seemingly clearcut example appeared in some lines of "B flioHax", as originally published, which, interestingly enough, were deleted in the definitive edition (cf. reference in fn. 22, pp. 439 f.).

B.

AFLPHAHOBA-IIEPETIÍ

3AMETKH O JIEKCMKE CJIOBA

O

TIOJIKY

HTOPEBE

C

MHoroHHCjieHHbix pa6oT, nocBameHHbix H3yneHHK) «3biica Cjioea o noAKy Mzopeee 3a nocjieflHHeflecaraJieTH»,UCHHBIMH Ha6jHO«eHHHMH OTMeneHti HccjiejioBaHHa P. O. ^ÍKo6coHa, nojiyneHHbie HM H3 conocTaBjieHHa C AaHHblMH DHCbMeHHblX IiaMHTHHKOB H JKHBblX CJiaBHHCKHX H3MK0B. 3T0 HanpaBJieHHe B HayKe o KpynHeñiueM naMaraHKe apeBHepyccKoñ JiHTepaTypbi eme aaJieKo He Hcnepnajio Becb MaTepHaJi, npoHHO CBH3biBaiomHH H3MK CAOBO — ero JieKCHKy, 4>pa3eojiorHK>, cHHTaKCHc — co cTapiiiHM nepHOflOM HCTOPHH pyccKOH JiHTepaTypw H a3biKa. HacToamwe 3aMeTKH npeaiiaraioT eme HecKOJibKO JieKCHiecKHX napajuiejieñ PEFLH

K OTFLEJIBHBIM CJIOBAM Cnoea o nojiKy Mzopeee.

1. Pa3bicxaHHH yKpaHHCKHx apxeojioroB h aHTponojioroB y6eflHTejibHO noKa3ajiH, JICMKH n e p e e e j i H J i H C b B X I I I Bene H3 CeBepcKoií 3 e M J i H , c j i e a o -

HTO K a p n a T C K H e

B a T e J I b H O HBJIHÍOTCH I I O T O M K a M H yjIHTCH H T H B e p U e B .

B

HX H3bIKe

COXpaHJUOTCH

IiaMflTHHKaMH,flpeBHOCTbK O T O p b l X HHKOraa He n o f l B e p r a j i a c b c o M H e H H i o , a Taicace — H C/ioeoM o noAKy Mzopeee. B H H M a T e J i b H o e H3yHeHHeflaHHbix3ana,HHO-yKpaHHCKHX pyicoimceñ (J)eoaaJibHoro nepHoaa B CBeTe j i e K C H H e c K o r o M a T e p n a j i a , H3BjieieHHoro H3 n p o H 3 B e A e H H H 4>ojibicjiopa JICMKOB — apxaH3MbI,

3aCBHAeTeJIbCTBOBaHHbie

1

3aflana 6y^yiyero.

C o o 6 m a i o H e K O T o p b i e (JRNICTM H3 3 a n a f l H O - y K p a n H C K H X

rpaMOT

P030BHM.2 P. O. 5IKO6COH, KOMMeHTHpyH CJIOBO "jiaaa" ("Yace HaMb cBonxb MHjibixb Jiaflb HH MbICJIHK) CMblCJIHTH ...", "HeMy MblHeiHH XHHOBbCKbIS CTptjIKbl ... Ha MOe» Jiaflbi BOH?"), cnpaBefljiHBo yica3biBaeT, HTO 3TO CJIOBO, OTcyTCTByiomee B apyrax ApeBHepyccKHx nHcbMeHHbix naMHTHHKax — oómecJiaBHHCKoe, cjieaoBaTeJibHo ApeBHee: BO Bcex BOCTOHHO- H loacHocjiaBaHCKHx a3biKax, a Taioice B cTapoM TCHICKOM «3biKe OHO cymecTByeT c OAHHM 3HaneHneM — Myac, aceHa, HeBecTa, OAHH H3 cynpyMeHeM "JlaflOMHp" TOB. 3 T O T BMBOA noflTBepxcaaeTca 3anaflHo-yKpaHHCKHM H (aaT.n. "JlaAOMHpoBH"), 3a4)HKcapoBaHHbiM acajioBaHHoñ rpaMOToñ 1377 r., BbiaaHXIV-nepBoñ

noJioBHHH X V

HCCJieflOBaHHe:

B., H3flaHHbix B . A .

CM. CB. ROPFLHHCBKHÑ, CAOBO O noAKy ízopeee u yKpaincbKa uapoóna Bbi6paHH¡ npoSjieMU (BÍHHÍner, 1963). ! BOJI. P030B, yKpaihCbKi zpaMomu, I : XIV e. i neputa noAoeuna XV e. ( Y KHÍBÍ, 1928). 1

noeiin:

16

B. AflPHAHOBA-nEPETU

HOH KHH3eM BjiaAHCjiaBOM OnoJibCKHM CBoeMy BepHOMy cjiyre "HMeHeMb peicyin JLA^OMHPT BOJIOIUHHÍ". TpaMOTa OTflaeT eMy 3cmjih no "TepHaBe peice H rozuii nojie", HMeHya STO He3acejieHHoe "nojie" — "nycTtma" (cp. B CAoee: "yace nycTbma CHJiy npHKpbijia"). Onpeflejiaa pa3Mep flaHH c Tex, KeM JlaaoMHp 3acejniT CBOIO "nycTbiHio", rpaMOTa Ha3biBaeT "Kaa«oro KMCTH" (cp. "a MOH TH KypaHH CBÍAOMH KtMeTH").3 JÍHTepecHa» cyflbGa HMCHH "flaacbGor", He pa3 ynoMHHaeMoro B C/ioee, pacKpwBaeTca acajiOBaHHOH rpaMOTOH 1394 r. Kopojia ürejuia CBoeMy cjiyre, KOToporo OH HMeHyeT "/Jamuro flaac6oroBHH 3aflepeBeuKaa 3CMJIHHHHI> Harnea 3eMJiH poycKoe".4 3Ta rpaMOTa 3aicpenjiaeT 3a BepHbiM cjiyroa 3eMJiH "nofli rajiuraeMi." H "nofli> 3yaeHeB0MT>". HTaK oTeu 3Toro "3eMjiaHHHa flaHHJibi", KpoMe "icpemeHoro" HMCHH, HOCHJI eme BTopoe — ",ZJaac6or", T.e. B Hanajie XIV B. noMHHJia STO H3biHecKoe HMH, HO yace CHHT3JIH B03M0acHbiM Ha3biBaTb HM nejioBeKa. 2. B TeKCTe CAOBO O noAKy Mzopeee yn0Tpe6jieH0 TpafluaTb fleBHTb rjfiarojioB c npHCTaBKoií "no". IÍ3 HHX TOjibKO Tpa He noATBepacAaioTCH apyrnMa naMaTHaicaMH: "noTpena™", "noTpyHaTaca" a "nonoama". H3 3TBX Tpex rjiaronoB Jinuib "noTpyHaTaca" HeH3BecTeH B flpyrax oflHoicopeHHbix rjiarojibHbix 4>opMax (HO ecTb cymecTBHTejibHoe "Tpycic"); rjiarojibi "TpenaTH" a "noaTa" 6e3 npacTaBKa "no" BCTpeiaiOTCH B CTapuiHX naMATHHKax. K cjioBy "nonoama" ("cBaTbi nonoama, a caMH nojieronia 3a 3eMJiio PycKyio") 6jiH3Koe no orreHKy 3HaneHHH 06pa30BaHHe aaeT HnaTbeBcicaa JieTonncb nofl 1195 r . : " H Ty nonamaca y Hero BCH MepHaa Kjio6ymi n o^apaB HX ^apMH MHoraMa a oTnycTH HX". H. H. Cpe3HeBCKHñ nepeBO^HT rjiaroji "nonaTBca" — HanHTbca, ymiTbca (flo nbaHa) (MamepuaAbi, II, CTJI6. 1192). Kaic BHABM, B CAoee "nonoama" HMeeT TO ace 3HaneHHe — HanoaJia ao nbaHa. 0 6 e e35K. 1554 r.

(CM. Cpe3HeBCKHñ, MamepuaAbt, II, 1195).

3AMETKH O JIEKCHKE "CJIOBA O nOJIKY HrOPEBE" y

17

C e p a n H O H a B j i a f l H M H p c K o r o w r a e M H a r a e 6 o j i b i u e He 3 a c B H f l e T e j i b C T B O B a H H b i i í

r j i a r o j i " n o n p o n o B e a a e T " ( c r p . 6).

3.

He

HMeeT T o p o r o

opM

cymecTBHTejibHbix laHHHMH ' V h

"b",

Cpe3HeB-

npHBe/jeH npHMep H3 c T a p m e r o cnncica

c oKOHHaHHeM

"ne",

n a p a J i J i e j i b H b i x T e M ace c j i O B a M c

Koraa

cymecTBHTejibHoe

nojiynaeT

npacraBicy

" y " : "yKopM -

H . H . C p e 3 H e B C K o r o , I I I , CTJI6. 1 1 5 7 , 1 1 8 7 - 1 1 8 8 ,

"i"

yicopMHe",

Mame-

" f l b H b - y f l b H H e " , " c b M b p T b - y M b p T H e " , "yxcac - y a c a c H e " (CM. n p H M e p b i B

puajiax

OKOH-

H " b " . O c o ô e H H o HacTbi (JìopMbi Ha " n e " n p a 6 o j i e e O6MHHHX H a

1 2 2 2 , H T . « . ) , HO e c T b H

"oötflb - oôtflHe". fljiH

cJiOBa " y e f l H e " c X I BeKa H 3 B e c T H a n a p a ^ u i e j i b H a a opMa " i f l b "

(BapnaHT

" » A b " ) , K O T o p a a , K p o M e 3 H a n e H H H " n n m a " , H M e e T H a p y r o e — " T p a n e 3 a " , COOTBCTC T B y i o m e e C M b i c j i y " y e f l H e " B n p H B e f l e H H O H H H T a T e H3 CAOBO

O

noAKy Hzopeee,

y n o T p e ô j i a e T C H c T e M ace n p e z y i o r o M " H a " : "3i>BaiHe H a flflb" ( I l a m e p u K X I B.), " c b x o f l a m e c f l H a t a b , a p y r a p y r a H c n a e T e " CM. H . H . C p e 3 H e B C K H H ,

MamepuaAbi,

(Anocmo/i no cnucny

H

CmaücKUü 1220 r.

—

I I I , CTJI6. 1 6 2 0 H 1 6 4 2 ) . B 3 H a n e H H H " T p a n e 3 a "

H c T e M ace n p e f l j i o r o M " H a " y n o T p e ö J i a e T c a H 4—A-A.

— ¡ja y — 1 >»jU¿ 4 líyo

-ICLJJ

c

¡Jjj

J .

J^T"

«aJL« .

J Jj*\

jtJii^ry

I JL

Oj-bj

J* ' •

'

yJ- ^jlpjl

J |>L~=-

ojj

^

\K

3BOJIIOUHH CTHXOCJIOJKEHHH B nEPCUHCKOfi H 0 3 3 H H

27

B 3TOM COOSMEHHH, He HMea BO3MOJKHOCTH oxapaKTepnaoBaTB Bee ocofieHHOCTH 3 B y K 0 B 0 f i K0Mn03HIJHH HOBOH H033HH, OrpaHHIHMCH BLIHBJieHHeM BCeX p a 3 H O O 0 p a 3 H H X BHflOB CTHXOCJIOJKeHHH

apy3 C n0Ka30M HX o6me&

aBOJHOIJHH.

OSOSMHB npaKTHKy nosTHiecKoro TBOpiecTBa CBOHX cooTeiecTBeHHHKOB apaScKHii jiHTepaTypoBea Xajina H6H AxMafl apaxHflH (718-786 r. H. 8.) B KOHqe VIII B. COSRAJI cTpofiHyio Teopnio craxocjiojKeHHH - 'HJIM aji-'apyn, o6HapyjKHB Bee cymecTBOBaBnrae HO H nocjie HCJiaMCKoft HO33HH apaSoB pa3MepH. KaK H3BecTHO, 9TO 6HJIO 15 KJiacciwecKHX pa3MepoB (TaBHJi, Ma^iiA, 6acnT, Ba$np, KaMHJi, xa3aji;jK, paMaji, paAH?a3, MyHcapnx, xa$H*IAETCH B TOM, ITO P A 3 P U B MEWAY

MeTpmecKoft cncTeMoii H peajiBHHM PHTMOM 6HJI 3HaiHTejiBH0 coKpameH, 3a cieT i e r o KOJIHHCCTBO pa3MepoB yBejininjiocB OT 16 30 126. KaraiM ate 06PA30M nponaomjio STO?

IIOHCHHM npiiMepaMH. Bo3BMeM cjieayiomee CTHxoTBopeHHe aji-MyTaHa66a:

t\Jj\ U-JJI J f^UaJI

¿

[ F & A ¿L»A *

y j

iijbijl ^

¿.I j

i

S F J U L JLS JJI

j

[»^j

KaK BHFLHO H3 HepBOH CTOHH, CTHXOTBOpeHHe HaUHCaHO pa3Mep0M KaMHJI

¿IPUI« ¿IPI

\Y« TK ' ^VVV ¿>jjt> ^R^1

P. M. AJIHEB

28

Orhslko npH HTGHHH cpa3y öpocaeTCH b rjia3a, ito b oßonx ßefrrax nosT flonycKaeT TaK HaaHBaeMiiS nnocTac, 3iixa$, t. e. nonivieHBi ctohbi c ee BapiiaHTOM. IIpHTOM Bapaai^HH CTOn HaCTOJIBKO CHJIbHa, HTO B Herapex CTpOKaX 113 12 clon JiHinB 5 CTon othochtch k HiicTOMy KaMiiJiy ( - u - u u ) , a ocTajitHue K npyrHM pa3MepaM (3 aromi k pafl5Ka3y jl.fr...,» a apyriie ¿ U j ¿ll*J k HewTpaJiBHHM CTonaM). 9ia Bapnan;iifl CTon hbjihgtch caMOft xapaKTepHoñ nepToñ apaScKoro apy3a, hto jjaeT no3Ty jiHKBHflHpoBaTb mohotohhoctb puTMa MeTpnnecKoä CHCTeMBi, npH^aTt piiTMiraecKoe MHoroo6pa3He craxoTBopeHHip h, tsikiim o6pa30M, nocTHHb Öojibineü BHpa3HTejibhocth. BosbMeM npyroä npmtep H3 A6y-HyBaca:

jijiI

isiîj • j j j ü l

j l l (¿JU Jî

íSj^¿ ¡J \'jJSj * Jtfs ¡J "bUi b B 9THX ßeÜTax, HanncaHHHX aByxcTonroiM PaMajieM jihiiii. neTBipe ctoiibi paMana ( ¿y ^IpU a Bee ocTantHBie ctoiih 3Hxa$Bi, t. e. noflMeHeHH cjieJiyiOmHMH CTOnaMH

OMpIî HTO npHBOflHT K 3HaHHT6JIBHOMy H3MeHeHHK)

pHTMa MeTpnqecKOii cxeMii. Bot 9Ty ocoßeimocTb apy3 noTepajx b nepcH^CKOfi H033HH H HHCJIO Bápnan,liii 6bIJIOflOBe^eHOflOMHHHMyMa, a 3a TO KaSKflOfi 3HXa$ npeBparajicfl b caMocTOHTejibHBiñ pa3Mep. TaK H3 BapnaD,Hii cTonix jj^Uli 06pa30Bajmch cjieflyiomHe pa3MepBi: 1. PaMajiH MaxßyH

i ¿ï^Ui t

¿y*^

nan HanpHMep,

11

(¿Aijjjr* (j>y & Jj1 i 1 tsiVj

^ *í ^

¿r-3.1' V

1

2. Painami Max3y$ : 0!>Uli C^ApU CjMPU OMPU (Sjj Jk) i i m x u,e3yp, HTOSBI KOHIHTB CTpoKy H e TaM, rae TpeSyeT 9TO pa3Mep, a TaM, r^e KOHiaeTCH peajibHan peib B BH^e BOnpOCOB H OTBeTOB. B O T n e p B H e RBe C T P O $ H CTHXOTBOpeHHH:

? ijJJ^tj

jS-

! ¿5Ú*JIPLI

— —

b ¿J>. — jT _ ? íOJ^íT ¿ ^ j j j j f ¿^Í3JI

^

4j

. . . ! ¿¿L)

í ol J l ' aL

_ -

30

P. M. AJIHEB ? ! -

J j

j*}

—

¿.J***?

jjlpjl

—

I—Jjj JÍ V ei ( j L » ^ j l

— —

Ji

isjj^ji

^J»

-

—

i All

skhboh p e r a , Bce ottchkh bhjiotb no MeiKflOMeTHö. BMecTe c TeM 9to n e npo3a, a HacTOfliipie cthxh co CBoeit otmgtjiiiboö pHTMHiecKoíi JiHHiieií, KOTopoe n p o x o f l H T i e p e 3 Bce Bocean. CTpo$ npoHSBeseHiin. HtoGli hcho y j i o B i i T t stot pHTM cthxh HyHtHO HHTaTB He TJia3aMH hjih Bnojirojioca, a BCJiyx, c o o ß p a s y a c b c HopMaMH pçaJibHOû pe™. T y T hcho o m y r a M HacKOJitKO TecHO cjiHBaioTCH K a K BHflHO B

3TIIX c T p o $ a x

nepejjaHH Bee HtoaHcu

ee HHTOHai^HOHHoro ßoraTCTBa

HonojiHHH a p ^ r îKHBoro r o B o p a pySKHBaeTCH,

apyra,

noBtiman

aiwoijHaHajiBHoe B03fleñcTBne, HHTOHaijHH

h phtm MeTpimecKOH chctcmh.

HTO

CTpOKH,

CHHTaKCH^eCKH

TaK, npn

htchhh

3aK0HieHHBie

c p a 3 y oÖHa-

peqeBhie

0Tpe3KH

BHyTpH KaSK^OÖ CTpOKH HO TOHaJIbHOCTH np0H3H0meHHH ReJIHTCH Ha HeCKOJIBKO

rpynn. B nepBofi cTpo$e: Irpynna: Ç (¿XÍ*Aj\j j¿>- àj — ¿SÓJuIpL.

—

-jI —

Kan bh^ho, 3T0 3aBH3Ka He Höcht CKOJiLKo-HHÔyRL aKcnpeccHBHoro xapaKT e p a H np0H3H0CHTCH pOBHHM Cpe^HHM TOHOM. B CJieflyiOmHX CTpOKaX flHaJIOra, c o a e p a t a m n x HeoHtH^aHHoe c o o ö m e H H e h CHJiBHa« CTeneHB yawBjieHHH, CMemaHHoro c raeBOM h coMHeHHeM BUSBaHHHe sthm cooömemieM, toh pe3K0 n o B H i n a e T C H h hoxorht 30 n p e j j e j i a : .. . . . . ?

^^Jl IZjjp- — 4! « Ä t e j l . . . I

-

3B0JH0UHH CTHXOCJIO/KEHHfl B IIEPCHUCKOÎÎ II033HH

31

B TpeTbeñ r p y n n e nHTOHai^HH npiioßpeTaeT HHOÖ xapaKTep, OHa HeMHoro naaaeT. H e XOTHT M n p i i T i c n C yjKacHHM H3BGCTIIGM, XOTHT, ^TOÖBI OHa ÖHJia HenpaBflOñ. Í BJUYT

j . . .

ß

—

! J I J I I I J A ; a¡j

_

r p y n n a cTpoK coflepíKHT pa3BH3Ky. Y S E A U B M U C B , HTO COOÖMEHNE npaBflonoaoÔHO, 0311H113 aBTopoB a n a f o r a MeTO,n lineerai cHiuieT TpaaimHOHHbie UpOKJIHTHH: RIOCJIE^HHH

AI» E J I I V J j > U J I ^ ( ¿ J J I Í

jjJLjb

ji^ J.-J jl

ÜIUJ^J iwJjli

¡ji-i

-

J

Becb 3TOT 3aKJIH)1HTejIbHblñ aKKOpA npOH3HOCHTCH MGpiJO, nOCJie^OBaTejIBHO h KaK 6H HanoMiiHaioT aHaGMy. T a K nocTpoGHbi Bce 8 CTpo$, 113 KOTopnx COCTOHT npOH3BefleHIie. BBIHBHB HHTOHaiiHOHHyio C T p y K T y p y , MBi cpa3y oÖHapyjKHBaeM H M e T p i i l e c K y i o CTpyKTypy cTHxoTBopeHHH. Ka/Kflan HHTOHaiiHOHHaH eflHHHija, ecjin CKaHfliipoBaTb Hx, COCTABHT 3aK0HieHHyi0 CTpoKy B pa3Mepe paMajin-MycaMMâHii M A K C Y $ (HJIH M A X 3 Y $ ) C TOHHOH CXGMOÖ : ¿¿y** -jl

Ç ¿SwUí

¿¿y^

Kan BHflHO, Jiiiiub b nepBOH CTpoKe ecTB pHTMHHecKan HHepi^iiH paMajiH MaxöyH, a ocTajibHue CTpoKH He noflxoRHT hh noa KaKyio cxeMy. OflHaKO, ecJiH. crpynnHpoBaTt CTpoKH B TOHajiBHO oflHopoflHHe rHe3flBi, TO nojiy^HM : 1. jeHa 6np 6ainra xaßap BapMU? Bh3hm roHuiy KepHM Ú^Uí 2. A . . . A . . . Earn, ha, ha... o Ha rajptiß ne KepyM

ó%¿ ¿f^** ht Kan BHflHO, 8TO HacTOHm,iiií paMajiH - MaxöyH. OflHaKO HOflOÔHaH rpynnapoBKa no TOHajitHocTii npoH3HomeHHH conpmKeHa -C GoJIblllHMH TpyaHOCTHMH H HeB03M0/KH0 ee BHIiyKJIO COÖJHOßaTB Ha npOTHHîeHHH i^enoH HBecH. IIoaTOMy XyceñH JJjKaBHfl, OTKasaBiniicB ot HHTOHau,noHHoro crpynnHpoBaHHH CTiixa, BBeji Tan Ha3HBaeMue JieüMH, Gojibmiie nay3H MeJKßy CTonaMH. HanpHMep, ero 3HaMeHHTaa meca "Illeiix CaH'aH", onyöjiHKOBaHHaa. b 1914 r., HanacaHa craxaMii b pa3Mepe xa$H$H-Max6yH:

SBOJIIOIJHfl CTHXOCJIOJKEHHH B IIEPCIWCKOH Ü033HH

33

OflHaKO 3Ta KaHOHHiecKaH ece "IHeñx Cbh' aH" mbi BCTpeiaeM pa3Hbie BapiiaHTBi xa3aflîKa, paMajin, My3ap'n H T. JJ. 9 ï a CMeHa pHTMa B CBH3H C H3MeHeHHeM KapTHH II HBJieHHH flaBajia noaiy bosmohîhoctb nofliepKHyn» 3HaMHTejiBHOCTb BbipaHîeHHOH mbicjih c HCKjiioHHTejiBHOö ciiJioË, rjiyöate Jiymie pacKpuBaTB xapaKTepti aeâcTByiom n x Jiim. IIOHTH OflHOBpeMeHHO C aBepßaÖ^iKaHCKHMH nOBTaMH K pe^OpMe CTHXOCJIOHteHHH npHCTyniIJIH H HpaHCKIie XyHOHÎHHKH COIOBa. IIpHHHTO, HTO pe^opMy nepcHflCKoro apy3a BnepBtie npoii3Be.ii ochoboiiojiojkhhk "HOBoro c r a x a " HuMa lOflîKHflîK. CaM H i m a lOaatiiflJK tojkc cnrrraji TaK. O^HaKO cjie^yeT OTMeTHTB, HTO nepBtie cepbe3HHe n o m m a i b 3tom HanpaBjiemiii Gbijih npeflnpiiHHTH BnepBHe y Miip3ane 9mKii. 9dikh HeoflHOKpaTHO OTKpiiTO sanBJiHJi,

HTO

JLKI çljjl aAs? . , . ¿ b

csi!/^**.

jz+jû

ts^i^i • • • . . . o\j>\ CjLJ^Í

OflHa H3 9THX pe$opM, KOTopue npefljiaraji hobt, KOCHyjiacb cTiixocjioHîeHiiH. BnepBHe b HCTopiwi nepcH^CKOÄ h033hh Sidkh Hapymnji 03H006pa3iie pHTMHHecKOH bojihbi KJiaccHiecKHX pa3MepoB. Bot otphbok h3 ero "HöBpy3HaMe", HanncaHHoro b rojçH ero npeÔHBaHHH b sMarpai^HH b JïcTaMÔyjie (1914-1918).

3B0JI10IÏHH CTHX0CJI05KEHHH B HEPCHÄCKOß n033HH

35

j j j i l eJUs ^yLfcOsJ

J* j\j

si

J

1

b

ùit>

J s .

jl¿T J y * a i y j

Ilo $opMe aio CTHxoTBop6HH6 HanoMHHaeT MyaTaaa^. OflHaKO h H3MeHeHHe KOJiHiecTBâ CTon (y MycTaaajj H3ocHjma6n3M h paBHOCTonne hcbhSjihmh) h pa3Mep (MycTasaa immeTCH b pa3Mep xa3aflHtH MaK$y$) cBHaeTemcTByioT o TOM, HTO B RaHHOM CTHXOTBOpeHHH HapymeHH OflHO H3 OCHOBHHX HOpM T p a f l H I^HOHHOrO CTHXOCJIOJKeHIIH H HanHCaHO Ha HHOfi MGTpHICCKOÄ OCHOBÖ. Beflb flJIH KJiaccHHecKoro apy3a H30CHJiJia6n3M h paBHOCTonne He3HßjieMH. OanaKO host, hto6h Sojiee peaiiHCTHiecKH h BupasHTe jibho H3o6pa3im> CMeny HacTpoeHHft h xapaKTep onncHBaeMHx npe^MeTOB, He npiiflepjKiiBancb aacTUBiniix KaHOHOB apy3a, BiiaonaMeHHeT ero, nofliimneT npncnocaÖJiHBaeT k cbohm xy^oatecTBeHHHM 3asaiaM. Kan 6hjio noicasaHO Burae, b craxoTBopeHHn HanacaHHOM npaBHJiBHHM xa3aAJKeM c yce^eimeM nocJie^Hero Rojiroro cJiora, b nepBOM CTHxe neTHpe ctoith 15 cjioroB, bo BTopoM - flBe ctohh h 7 cJioroB, b TpeTteM, n a n b nepBOM, b leTBepTOM KaK bo BTopoM, a b hhtom h raecTOM c r a x a x no Tpn ctohh n i l CJioroB h b nocjieaHHx CTHxax no 7 CJioroB. üofloÖHHe nsMeHeHHH pHTMHiecKoii CTpyKTypM pa3MepoB y 9idkh BCTpeiaiOTCH nacTO. OcoßeHHO 9to mh HaßjHOAaeM b ero noaMe "H^eaji". OcTaBjiaa MyauKajibHyio HHepipno KJiaccniecKoro pasMepa Heii3MeHHoä, host, coo6pa3yHci> c xapaKTepoM nsoöpaHtaeMoro npeflMeTa, MeHHeT ^Hana30H pHTMHiecKoii bojihh, t o yflJimraeT, t o coKpamaeT ee. 8 t o HOBaTopcKiiii npneM 9iiikh b aáJibHeñmeM npnßeji k C03J^aHHK>, eCJIH TaK MOÎKHO B H p a S H T b C H , BOJlbHOZO apy3a,

KOTOpHÄ BnOCJieflCTBHH

noJiyiHJi nmpoKoe npHMeHeHae. CaM 9hikh hcho C03HaBaji, hto xoth ero hoboBejjeHHH Be 6y^yT npiIHHTIJ CTOP OHHHKaMH TP a^HI^HOHHHX $OpM CTHXa1), HyjKHO np030jDKaTB noHCKH hobhx $opM CTHxa. Oh b to Hie BpeMH KaTeropneCKH BHCTynaji npoTHB Tex, kto, no ero mhchhio, non bhjjom ooHOBJieHHH noaralecKOÄ $opMH, xoTeoi HaneTB Ha nepcHflCKyio ho33hio nonoraeHHoe nnaTbe H033HH a p y r n x HapoflOB : yL»

j «JLíp ¿JI ¿Z* i-yUI ^LAISJ

òljlTJiil»»»jl Y — J-® i CjJJS" iTÍ-J!;

36

P. M. AJIHEB

-

ti^S"AÍ

a^U- ^ p ^ l j j l

eijj^j

«)t> jjiLfc» j i

4S~C..J1

8 t o t npH3HB n o a T a , c y a n n o h g k o t o p b i m 3&hhhm, 6 h j i h b h 3 B 9 h l i HOBaTopCKHMH 9KCnepHMeHTâMH, KOTOpHe IIOHTI1 OflHOBpeMeHHO C 9lIIKH, npOBOflHJI H f f l t f a lOjKHflíK. 1 9 2 2 r . o h o n y ß j i H K O B a j i c b o i o 3HaMeHHTyio n o s M y " A $ c a H e " .

KaK

8T0 ycTaHOBJieHO TOHHO, nop; b j i h h h h g m $ p a H i í y 3 C K o ñ j i i i T e p a T y p H , c KOTopoft x o p o i n o 6biJi 3H&KOM no3T, o h H a ^ a j i co3aaBaTB h o b h g -

HoBOBBefleHHH HHMa He BHXOAHJIH 3a r i p e t e JIBI yKa3aHHoro paciJieHeHHH CTon pa3Mepa Meatfly CTHxaMii H B n o c j i e f l y i o m n x npoH3BefleHiiHx: ((Jlj^í

ìi\y\st-

((

o J u j i

1vai ))

H T.fl. IlepBOe CTHXOTBOpeHHe, B KOTOpOM ÜHMa CymeCTBeHHO OTOmejI OT HOpM KJiaccHiecKoro apyaa, HanHcaHO HM B copoKOBtix r o f l a x . B 3TOM npon3BefleHHH HHMa

flOBeji

flo

JioraiecKoro

3aBepmeHHH

OHBITH CBOHX

npeflmecTBeHHHKOB,

CTpeMHBiHHXCH nofliHHHTb a p y 3 HOBHM n o a r a i e c K H M 3aflanaM-nepeflaTB npH N O M O M H B H Y T P E H H E Ñ 3 B Y K 0 B 0 I I KOMIIO3HIJHH IIHTOHAN^IIOHHO-PIITMHIECKOE

60-

raTCTBo coBpeMeHHoiî JKH3HH, a He pnTMH3HpOBaTB eë. K a K yBHflHM HHœe, HHMa ne OTKa3aJicH ÒT rJiaBHoro npHHijHna KJiaccaieCKoro c r a x a -

BHyTpeHHeñ 3ByK0B0ö cHMMßTpiiii, pHTMa, o6pa3yeMoro npH

NOMOMN nocJieflOBaTejiBHoro lepeflOBarniH flOJirnx H KOPOTKHX cjioroB, a Jimnb CHjiiHO H a p y m n j i o6H3aTeoiBHoe AJIH KaHOHHiecKoro

apy3a

0flH006pa3HHx

COH3MepHMBIX, MOHOTOHHO IIOBTOpHIOmHXCH pHTMHieCKHX BOJIH. 9niKH

flo

HHMa,

TâKîKe II3MeHHJI flJIHHy MeTpHieCKOñ BOJIHH. OflHaKO OH COXpaHHJX

MejioRHKy, My3HKajibHyio HHEPI{HK) pa3Mepa. O H

flocraraji

3TO xeM, ITO BBI-

ßPACHBAJI He H3MEHHH H a ï a j i b H y i o H KOHennyio n a c r a c r a x a , Jinmb oflHy H flBe

38

P. M. AJIIIEB

CTonH c TaKHM paciCTOM, hto6h HBMeHHJicH Jiiirni. o6i>eM, a He xapaKTep pHTMHiecKHX BOJiH. HanpHMep, BOJiHa Tana

no xapaKTepy cBoero BByiaHHH HfleHTHiHa c bojihoS

Ohh OTJiHiaioTCH apyr ot npyra jurait oöteMOM. Huna ate noineoi flansme. Hhcjio cjioroB h CTon oh yBemniiBaji h coKpamaji t&k chjxbho, i t o KOJiiiieCTBeHHoe H3MeHeHHe nepexo^mio b KanecTBeHHoe h cthxh jmnn> oTjjajieHHO HanoMHHann o tom, i t o ohh pa3MepeHti. B o t nepBoe CTHXOTBopeHiie, HaimcaHHoe coraacHO yKaaaHHHM HopMaM : ¿iUjJül

Ji

pHTMHiecKaa

BO I I ,

SeKJiaMaU,HOHHyiO

b CTHxax

BHpa3HTejIbHOCTI>

B

HHTOHaii,HH HeKOTOpHX

nocjieayjomnx np0H3B6^GHHflx H H M A eme 6ojn>me yBejinnnji pa3pHB B OSIEME CTPOK. HanpHMep, B CTHXOTBOPeHHH A ^ J

BCTpenaeM TaKHe CTPOKH:

4.T. «Jjfcj

j*jA

j A j L tjJ ¿¿ly^

^

U s j l j i ^y* ¡j^rf

U c ^ y b j i OVJ ¿J] I jAJl . . . j ^ U - ^ a ^¿j* j l

J'

ol^ ol^J

&

KaK BH;i;HO H3 xapaKTepa CTOII, CTHXOTBOpeHHe HanncaHo pa3MepoM paMaji. KaK H3BecTH0 caMult nojmtiii h npaBiijibHtifl: paMaji He flonycKaeT 6ojiee 16 CJioroB - qerapex CTOH B ohhoS CTpoKe. 3flecb we KOJIKTOCTBO cjioroB CTpoK fl0X0flHT ho 29 cjioroB (6ojiee ceMH CTOH!), a KopoTKHe CTHXH coaepacaT Jiamb 9 cjioroB. TaKOH paspHB Meatfly HHCJIOM cjioroB B CTpoKax, XOTH h OHH pacnojioHteHH CHMMeTpHHHO HO THIiy ^y^pli eCTeCTBeHHO 3aTpyflHHeT BOCnpiIHTHe «aHHoro TeKCTa KaK craxoTBopeHiie. OT aioro 3M0u,ii0HajibH0e B03REIICTBHE npoH3BeAeHHH H AOXOfllHBOCTb n03THHeCK0fi MHCJIH, eCTeCTBeHHO, BHaiHTeJIbHO CHH3HaHCb. PaHbrae Bcex STO NO^YBCTBOBa c r a x a oco6oe BHHMaHHe. 3 f l e c b BectMa y M e c r a o öy^eT cocjxan>CH Ha BHB03, KOToptiñ n p a r a e j i Hnivia nocjxe HOJiriix 3KcnepHMeHTOB H KOToporo saGtuiH HeKOTopHG Mojioane noaTH, cHHTaromae ce6a

e r o nocjieaoBaTejiHMH.

OH roBopuJi: ^jîLaj

JjIj j p l i J^jj i».jt»»» ^ a j ^ a i x j t C i j j

t r i J**" U* ¡J\*ljl

• "^^í4

K

[¡y**" &

Jjl^ij* j j i j l j l j ^ Ü

4S" ¡ J j & i j ï j J » j i

y à

¿S" pjlJU* U .

¿JL-MS" ( j j a jJ

jA j l U i à j A y j j j Öjj

. ¿jljjlj C w í

ö l j f - j «cAj) j L î l

i)

(,/^tiJ

A» C*AL5> ö j j

¿y* C i j y p ( j j i l j i 4

òlw'l

coHtajieHHio, H m a He p a c K p u B a e T K a n n e npaBHJia HJIH co3^aHHH pHT-

MH^ECKORO

cTpoH c r a x a OH c w r a n HEOÔXOAHMHMH. noflnepKHBa« o6fl3aTejib-

HOCTb 3ByK0B0ñ CHMMeTpHH flJIH CTHXa, JIHfflb B ORHOM MecTe, HO BecbMa CKyno noaT yKa3HBaeT Ha O6III;HH npHHijHn ee co3flaHHa:

ljuffiy

jjbi

¿JL :

Jj&wyd» ti>Lolj,ij AliliS"

y n o p f l f l O i e H H H

6oJiee MejiKHe efliiHnu,Bi. 8TO y n o p H f l o n e H H e s o c r a r a e T c n n p a noMomn npneuoB xyflOHtecTBeHHO pacuieHeHHoro ciiHTaKciica. T A K A nepBan H BTopan c T p o K H n e p B o r o n e p w o f l a p H T M i n e c K H nocTpoeHH He TOJIbKO pH(|>MOÖ, HO H HOJ1HHM CHHTaKCHieCKHM HapaJIJieJIH3MOM: 2 n p O C T H X npefloioHteHHH COCTOHIRHX H3 C K a 3 y e M o r o H n o a j i e j K a m e r o , K O T o p t i e p a c n o o i o HteHU

CHMMeTpHHHO.

TpeTbH

CTpOKa,

TeMaTHieCKH

OTTaJIKHBaflCb

OT

HHTOHâi^HOHHO K a K 6 H B o c x o f l H T B B e p x , m o c r a r a n npe;ne.!ia o n y c K a e T C H

HHX BHH3

B n o c j i e f l , H H x C T p o K a x H s a B e p m a e T n e p H ^ . B i o p o ü n e p n o ß - npeflJiOHceHne C BOCXOFLHMHMH H HHCXOFLHIUHMH HaCTHMH HO CBOeMy CTpOeHHK) COBHAJJAET C nepBHM

h

KaK

6h

o6pa3yeT

pHTMHiecKHft

napajijieji. TaM

Hte, r ^ e

aBTop

o ß o c o ß j i H e T KaKoö-HHÖYFLB NEPHOFL, CTpoHT e r o B HHOM KOMHO3HIJHOHHOM n j i a H e , OH y c i i J i H B a e T B H y T p a H e r o a p y r i i e p i i T M O o 6 p a 3 y i o i u , H e 9JieivieHTbr, K â K , H a n p n Mep, pH^My,

pacnojiOHteHHe

flOJirax

H KpaTKnx

H np.:

FDW4

JUT

b

FRI

BHYTPH p e n e B t i x

oTpesKOB

44

P. M. AJIHEB

B 9tom nepaoae KpoMe oSHJibHoñ ph$mobkh, m h HaS-moflaeM yaHBHTejibHo H6TKoe K0Mü03imn0HH0e eflHHCTBO, AOCTHrHyroe CHHTaKciiHecKHM napajuiejiH3mom, c0np0B0HmaeM0e napajuiejiHSMOM puTMHiecKiiM. Kam^bie flBe c t p o k h coa^aioT oflHO npocTeñinee cjiojKHO-no^TOHeHHoe npe3Ji0JKemie. B HeieTHbix CTpoKax 1, 3, 5 rjiaBHHe npe^JiOHíemin 2, 4, 6 noflHHHeHHHe npeflSOHteHHH. IIparoM Bce HJieHH npe^JioiKeHHH, GOCTOHmae H3 CHHTaKCHiecKii napajuiejitHHx cjiob, pacnojioHteHH Ha p h t m h i c c k h napajuiejibHbix MecTax. B t o ate BpeMH Kaffiflan l e r a a n CTpoKa rpaMMaTiraecKH aaMKHyia c Heieraoft. CjiHBancb BMecTe l + 2 ; 3 + 4 , 5 + 6 , KaK 6 h o6pa3yioT nepBHrayio eammuy piiTMiiiecKoro pn^a. BMecTe c TeM conepaíamaflca b 9tom nepaoae BocxoflnmaH jihhhh 8 - 5 CTpoKa B086yHíaaeT peaKijHio HcntiTaHHyio b npeRHsymnx nepuoflax h KaK 6 h B03BpamaeT Hasa^. 9 t o nyBCTBO saKpenjineTCH 6-m cthxom, KOTopoe yate MyeTCH c nocjieflHHMH craxaMH npeawaymnx nepiiOAOB. TaKHM 06pa30M npn IIOMOmH K0Mn03HU,H0HH0r0 yMGHHH Ha CJiaSoM MeTpHíeCKOM $OHe C03flaeTCH BHCmaH rapMOHHH, K KOTOpOÍi CTpeMHTCH D03T. Bce y ^ a i H H e b piiTMiiiecKOM OTHOHieHHH CTHxoTBopemiH oSjiajjaioT 3toü BHyTpeHHeñ rapMOHHeft, K0T0paH flocTHrayTa He npn noMomn CTpororo co6jiiofleHHH MeTpHiecKHX TpeSoBaHHñ, a nocpeflCTBOM cjihh&hhh OTaejitHHx ajieMeHt o b MeTpHKH c HeTKofi CKpynyjie3Hoií K0Mn03Hu;H0HH0ñ OTflejiKOü. HeT HafloSHOCTH rOBOpiITb,- 1T0 BHineOIIHCaHHBIH BHfl KOMn03HIJHOHHOñ OTflejIKH He noBTopaeTCH b flpyrnx CTiixoTBopeHiiHx. IIoaT Kaní^biñ pas nojibayeTCH hobhmh ^opMaMH 3Toro SoraToro apceHajia pHTMiiiecKOit Bbipa3iiTejibHOCTH. OaHOBpeMeHHO C HHMa B oSjiaCTH C03flaHHH H O B H X npHHI^HÜOB pHTMíraeCKOÜ opraHH3ai^HH hpoh3boahjih o h h t h h spyrne noaTH. OflHaKO 9KcnepHMeHTii 3THX "pe^opMaTopoB" He yBeHnajiHCb ycnexoM, h6o Bce ohh CTpeMHJiHCb aBTOMaTHHecKH nepeHecra eBponeñcKiie opMH cTHxa b nepcHflCKyio jiHTepaTypy, Hacera Ha nepcHflCKyio no33Hio, no BtipaHíemiio 9mKH, noHOraeHHoe HJiaTbe npyrnx HapoflOB. H3 Bcex pe$opM cthxocjiojkchhh jirnnb CHCTeMa H H M a noJiyiHJia Sojibiuoe pacnpocTpaHemie h jierjia b 0CH0By Hosoro HanpaBJieHHH b nepcHfleKOü ho93hh. Tenepb Gojibuihhctbo mojio^hx H09T0B nnmyT BOüibHHM apy30M HiiMa. Cpe^H HanSojiee TajiaHTjmBbix nocjieaoBaTejieñ H H M a moíkho yKa3aTb H a Haaepa Hanep-nyp, Ommhs, TaBajuio.nn, Cañe, BaMfla«, AüeHfle, CnHByin Kecpañn, XacaH X y H a p M a H f l H h MHornx flpyrax, ibH npoH3BeaeHHH noJib3yioTCH 3acjiyHteHH0ñ nonyjiHpHOCTbro cpe^a nrapOKHX KpyroB npocBemeHuux wraTejieü. OflHaKO BMecTe c TeM, cJienyeT OTMeTHTb, 1TO KpOMe yKasaHHHX H09T0B, eCTb H I^ejlblft pnfl a p y r a x H03T0B, CHHTaiomHX ceSfl nocJieflOBaTejiHMH HHMa. C y « H no h x npoii3Bej^eHHHM o h h ycBOHJiH caMtíe HeyaaiHue c t o p o h h 3KcnepHMeHTa HiiMa, o t k o t o p h x oh OTKaaajiCH B nocjieflCTBHH. 9 t H JHOñH, JIHDIHB CTHX O^HOrO H3 CaMHX KpyHHHX CpeflCTB Bbipa3HTeJIbHOCTH 9M0qH0HaJIbH0ñ OKpameHHOCTH, BHHBJIHIOmeftCH Hepe3 BHyTpeHHIOIO rapMOHHH), pHTMHiecKOÉt CTpyKTypH, ROBejiH c t h x flo cTeneHH caMoft o 6 h h h o h np03H c

3B0JII0IIHH CTHXOCnOJKEHHH B HEPCHJICKOH IT033HH

45

3JieMeHTa.Mii phtophkh. 9 t o , ecTecTBeHHo, ßpocaeT TeHb Ha yaana iiojjjihhhhx HOBâTopoB, flaeT jiiiiiiHiiH pa3 n « m y npothbhiikaM hobhx Sojiee coBepmeHHtix noaTHiecKHx $opM. 06o6man no3THiecKyio npaKTHKy b coBpeMeHHoä nepcaacKOft no33HH, Ha mom B3rjiHR, MOJKHO c^ejiaTb cjie^yiomae bhboah. IIoHCKH HOBBIX CHCTeM ßbKIH 06yCJI0BJieHH HOBOÖ HCTOpmeCKOH fleÖCTBHTejIbHOCTblO, HOBHMH XyHOHieCTBeHHHMH 3aflaHaMH. Il0HBJI6HHe CBO6oftHOr0 apy3a He

HBJIHeTCH npO^yKTOM HOCyJKHX Htie B0BM05KH0CTH flJIH yBejIH^eHHH CMHCJIOBOÄ Harpy8KH H pHTMHHeCKOË BHpaSIITeJIbHOCTH CTHXa. MocKBa

W. S I D N E Y A L L E N

CORRELATIONS OF TONE A N D STRESS IN ANCIENT GREEK "Hitherto, the attempts to clear up the role of the accent in Greek verse suffered from a false assumption that the accent, if it participates in the verse structure, must gravitate toward the downbeat, and also from a too summary, overly simplified treatment of the accent itself." Roman Jakobson, Selected Writings, I, 270.1

T

H E above-quoted words of Roman Jakobson have already helped to stimulate a limited essay in restatement so far as the accent itself is concerned;2 it was there observed that "the Greek accent is traditionally conceived as consisting essentially of the culminative high tone, the remainder of the word having automatically low tones and so being structurally irrelevant"; and it was argued that one might rather envisage the Greek accent as ideally comprising "not simply the high tone but a more extended 'contonation' including both the high tone and the following fall; ... a syllable marked with the circumflex did in fact contain this combination within itself". Thus a word such as avGpamoq would have the pitch-pattern ~ \ _, with a falling glide on the second syllable, returning the pitch from high to low; in a word such as 8c5pov the circumflex indicates that the high pitch occurs on the first mora of the long vowel, and the falling glide therefore comes on the second mora, i.e., within the same syllable, giving a pitch-pattern _ ; whereas in 66pou the high pitch is on the second mora, and so the fall occurs on the next syllable, i.e. \ . By employing the 'mora' concept, Jakobson had succeeded in giving a more concise statement than was hitherto possible regarding the limitations placed (in Attic) upon the recession of the accent from the end of the word, namely (ibid., 263): "The span between the accented and the final mora cannot exceed one syllable"; and by introducing the idea of a 'contonation' which included the post-tonic falling glide, it became possible further to simplify the rule by stating it in the form: "Not more than one (vocalic) mora may follow the contonation." 3 1

From "Z zagadnien prozodji starogreckiej", in Z zagadnien poetyki: Prace ofiarowane K. Wdycickiemu (Wilno, 1937), 73 ff. 8 "A problem of Greek accentuation", in In Memory of J. R. Firth (London, 1966), 8 if. 3 N.B. The final diphthongs at and ot are generally to be treated for this purpose as ay, oy, i.e. VC, in the same way as e.g. final ov.

TONE AND STRESS IN ANCIENT GREEK

47

The post-tonic fall may thus be considered a relevant feature of the accent. But the 'ideal' pattern of the contonation would of course be precluded if the high pitch occurred on the final mora of a word, since there would then be no place for the post-tonic fall; and it was suggested that this was the reason for the replacement, except in pausa, of the final acute by an orthographic 'grave' — though opinions may differ as to what precisely the grave accent-mark signified. Only recently has a striking parallel to this principle come to my notice, in an account by Trubetzkoy of the accent in Serbo-Croat, 4 where he observes that the free tonal accent has a rising pitch-pattern, and the beginning of the following syllable is held at the same level as the end of the accented syllable: moreover, "Diese Mitbeteiligung der nächsten Silbe ist für die phonetische Realisation des freien Akzentes im Serbokroatischen unbedingt notwendig, und daher ist die Freiheit des Akzentes dadurch beschränkt, dass er nicht auf der letzten Silbe eines Wortes ruhen kann." In Greek the 'ideal' pattern is perhaps fully realized only in the regularly recessive accentual system of Aeolic. More recently, by means of a statistical study of preferences in Greek verse regarding the location of words of different quantitative patterns, I have attempted to establish the principles governing the incidence of STRESS in plurisyllabic words in ancient Greek. 5 It was assumed that in the language, as well as in verse, context must be taken into account in assessing the quantity of final syllables — i.e. that word-final VC is light before initial vowels but heavy before initial consonants; so that the quantitative patterns of individual words might vary with their environment. Account was also taken of the principle of 'indifference' applicable to the final syllables of most types of verse line, which was assumed to have a parallel in the pre-pausal position of normal speech. This investigation led to the following basic hypotheses: (a) words were primarily stressed on their last heavy syllable; (b) a secondary stress fell on preceding heavy syllables if separated from the primary stress by at least one (quantitative) mora; (c) a (probably weaker) stress fell on an initial light syllable if followed by two light syllables. After allowing for the principle of 'indifference', this hypothesis led to an average of 90-95% agreement between verse-ictus and linguistic stress, with virtually 100% agreement in the coda of the line. More specifically, it provided a single and simple explanation for various 'laws' regarding the positions in which heavy word-finals may and may not occur (including Porson's, Naeke's, and Maas's), all of which reduced to the avoidance of juncture where this would produce conflict between stress and ictus. Since relatively few Greek words end in more than two light syllables, it followed that a primary stress generally fell on one of the three final syllables of a word; and 4 6

Grundzüge der Phonologie, 191. "Prosody and prosodies in Greek", to appear in TPS, 1966.

48

W. SIDNEY ALLEN

in Greek (unlike Old Indian and the ancestral Indo-European) the high tone also was restricted to these syllables. Moreover, both the recession of the high tone and, partially, the position of the stress6 were regulated by the structure of the final syllable. But in Attic (unlike Aeolic) the tone was not regularly recessive; and whereas the STRESS was regulated by the QUANTITY of the final SYLLABLE as a whole, the TONE was regulated by the LENGTH of its VOCALIC nucleus only (hence the references above in the one case to 'quantitative' and in the other to 'vocalic' morae); the monosyllabic or disyllabic nature of the contonation might also depend on vowel-length.7 Moreover, whereas the quantity of the final syllable, and so the stress-pattern, might vary with environment, the tonal pattern was (with certain limited exceptions) invariable. The conditions of the Attic tonal accent thus provide a prima facie expectation of some degree of intersection with the phenomena of stress, but also of some degree of divergence. It is here proposed to test these expectations in some detail; but two caveats should first be entered: (1) One does not necessarily presume direct causation for any correlations that may emerge; and (2) A correlation of tone and stress does not necessarily mean a correlation of stress with HIGH tone, but rather with any element of the contonation. It is traditionally agreed that there is no correlation between the Greek tonal accent and the ictus of any classical metre; in W. Beare's words,8 "Word-accent is not a structural element in classical Greek verse. The accents which we write and print have clearly no relation to the metrical beat." And if we are basically correct in postulating a high level of agreement between verse-ictus and linguistic stress, it logically follows that we may not expect to find any correlation between linguistic stress and tonal accent. But the traditional statements regarding the tonal accent and Greek verse have taken account only of the high tone, i.e. the acute accent (and the marked 'grave' as an allotone of the acute) or the acute component of the circumflex; they do not therefore necessarily conflict with our expectation of finding some degree of correlation between the stress and SOME element of the contonation. Stress and tone, as we have seen, are governed by different though intersecting factors, namely QUANTITY of syllables and LENGTH of vowels (or diphthongs); and it is essential to keep the distinction absolutely clear. QUANTITY is either HEAVY or LIGHT, and length is either LONG or SHORT; the distinction may be maintained graphically by using the familiar SUPERSCRIPT macron and breve for vocalic LENGTH, but a SUBSCRIPT macron and (inverted) breve for syllabic QUANTITY (though these may for convenience be placed below the vowel of the syllable). ' '

Since if the final were heavy, it carried the stress; but if not, not. As e.g. in Xfj£ai — \ _ beside Xé^ai — The distinction between 'monosyllabic' and 'disyllabic*

contonations is reminiscent of the Norwegian accents often so called, on which see especially E. Haugen & M. Joos, "Tone and intonation in East Norwegian", AclaPhilol. Scand., 22 (1952), 41 ff. • Latin Verse and European Song, 89.

49

TONE AND STRESS IN ANCIENT GREEK

The material to be investigated involves four at least partially independent variables — tone, stress, length, and quantity; and the polydimensional nature of their mutual interrelations raises complicated problems of diagramming and exposition. The last two factors, however, intersect in Greek to the extent that a light syllable always implies a short vowel, and a long vowel always implies a heavy syllable; and a more readily apprehended picture may be presented by first examining a hypothetical, simplified model in which it is presumed irrelevant to distinguish length from quantity — i.e., in which the position and nature of the contonation are governed, like the stress, by syllabic quantity without regard to vocalic length. Apart from the symbolizations already proposed, we shall indicate final syllables by the numeral 1, penultimate by 2, and antepenultimate by 3; 9 stress by a subscript point; high tone by the acute accent (as orthographically), falling tone by the grave accent (orthographically unmarked, post-acute), high + fall by the apex (orthographically circumflex), and the final (orthographic) 'grave' by x . Optional syllables will be enclosed in brackets. In terms of this model and symbolization, the following possible combinations of stress and contonation patterns are permitted by the rules of Attic accentuation in plurisyllabic words in other than pre-pausal position; where quantity or stress is not indicated, it is irrelevant. (a)

[3] 2 1

(b)

(f)

3 ^ 1

(g)

[3] 2 1

(c)

V

[3] 2 1

{d)

T

[3]i> T

(e)

[3] 2 1 T

^

3 2 1

These combinations represent positive and negative correlations of stress with the various tonal patterns as shown in Table I. TABLE I

/ X

ft« A

+ /

—

d, g %

b

a

d, g

f

c, e

0

- v i

The pre-pausal position has been disregarded in this classification for the reason that in some cases, as a result of the principle of 'indifference', the stress-pattern is dependent upon more extended phrase-rhythms ;10 the interpretation of the principle is in '

As in J. P. Postgate, Short guide to the accentuation of ancient Greek, § 126 fT. Discussed more fully in TPS, 1966. Briefly the principle is that a pre-pausal light syllable may be stressed if the penultimate is also light, and if preceded by a basically rising rhythm; and that a pre-pausal heavy syllable may be unstressed if the penultimate is also heavy, and if preceded by a basically falling rhythm. 10

50

W. SIDNEY ALLEN

any case not always certain, and we shall continue to disregard this position in individual doubtful cases. In subsequent discussion also it will be necessary to leave out of account correlations involving the 'grave' ( x ) accent, in view of our uncertainty regarding the nature of the tonal pattern represented by it; since there are equal positive and negative correlations (Z>+ : a—), this omission should not seriously affect other conclusions. From Table I above it will be seen that the high tone has only one positive correlation with stress, and two negative correlations; but that the falling tone has two positive correlations against one negative, and that the high + fall has only positive correlations. The evidence of Old Indian, and of the Delphic hymns, seems clearly to suggest that from a purely phonetic point of view there was little if any difference between the high + fall (circumflex) and the fall (post-acute), both perhaps being phonetically representable in some such form as whenever, in subsequent discussion, we wish to class these two together, we shall refer to them by the title svarita ('intoned'), applied to both categories by the Old Indian grammarians. In Table I the svarita will be seen to have four positive correlations and one negative. Thus the provisional picture so far emerging is the perhaps rather surprising one of a predominantly NEGATIVE correlation of high tone with stress, and a POSITIVE correlation of svarita with stress, particularly when, as in the case of the high + fall, it comprises the whole contonation. 12 Apart from the general rules of accentuation in so far as they are governed by vowel-length in the final syllable, this situation is further encouraged by the so-called 'final trochee rule' of Attic, whereby, if the vowel of the final syllable is short and of the penultimate long, and if the latter is accented, the accent must be circumflex — thereby precluding (still in terms of our simplified model) a combination of the type [3] 2 1 in which there would be a positive correlation of stress with high tone and a negative correlation with falling tone. 13 Moreover, the single positive correlation of stress with high tone (in the type (/)) is liable to some restriction when the antepenultimate syllable is heavy, i.e., in the dactylic ending — 3 2r> 1. For in nominal forms, in so far as the position of the accent represents a historical recession from the final, this is the pattern to which 'Wheeler's Law' is applicable, viz. "Daktylisch ausgehende Oxytona werden zu Paroxytona" Cf. Allen, Phonetics in Ancient India, 87 ff. The only previous suggestion known to me of such a situation dates from 1872, in J. H. H. Schmidt's Griechische Metrik; cf. p. 195: "Somit dürfte die Thatsache ganz offen vor Augen liegen, dass der griechische Accent völlig unabhängig von der Quantität ist, und dass mindestens in eben so vielen Fallen die Neigung erkennbar ist, denselben im Gegensatz zu der letzteren zu verwenden, als ihn damit zu vereinigen", and p. 204: "Von dem Acut unterscheidet sich der Circumflex noch dadurch wesentlich, dass er nothwendig mit langer Silbenquantität verbunden ist; er entspricht also in der That den deutschen Accent-Icten mehr wie die anderen griechischen Accente." 13 Thus, e.g., Attic nalSet; beside Doric naiSg?, and Attic contraction ¿axflrcei; (cf. Horn. £CT(I6TE Gujarati rän, räjaputräh > Guj. räut. It is perhaps at least partially relevant to this difference of development that the Old Indian tonality, which preserved the freedom of position inherited from Indo-European, and also permitted the contonation to extend across word-boundaries, was for these reasons incapable of contracting any patterns of linkage with wordstress such as were encouraged by the positional restrictions of Greek. " Cf. Jakobson, "Die Betonung und ihre Rolle in der Wort- und Syntagmaphonologie", TCLP, 4 (1931), 164ff. (166: "Als Sprachen mit dynamischer Betonung werden nämlich solche Sprachen bezeichnet, in denen der Betonungsumfang phonologisch immer der Dauer des silbigen Phonems gleicht"). " Cf. Jakobson, TCLP, 4, 182: "Die monotonische Tonstufenkorrelation kann nicht mit der Quantitätskorrelation der Vokale im selben phonologischen Plan eines Sprachsystems koexistieren." A corollary of this was the monophthongization of the diphthongs, which had taken place in the course of the preceding centuries. M With a possible exception in Marathi: cf. R. L. Turner, JRAS, 1916, 203 ff. " Cf., e.g., Wackernagel, Altindische Grammatik, I, § 254 (with Debrunner, Nachträge, p. 171); A. Thumb, Handbuch des Sanskrit, I, § 55.

62

W. SIDNEY ALLEN

This investigation still leaves many problems unsolved; but it is hoped that, in conjunction with other studies referred to, it may have helped to explore some rather obscure accentual territory, and incidentally to remedy somewhat for Greek the shortcoming that "No searching attention has yet been paid to the pertinent question, what variational functions, if any, are performed by the contrast between the different pitch forms of word accent in the verse of those European languages where this difference finds phonemic utilization".88 TRINITY COLLEGE, CAMBRIDGE

" Roman Jakobson and Alf Sommerteli, "On the role of word pitch in Norwegian verse", Lingua, XI (1962) ( = Studia Gratulatoria dedicated to A. W. de Groot), 205.

HHKOJIA0

JIHTEPATyPA K

M C T O P H H

HflEÏÏ

H

AHFLPEEB

H K O H O r i H C b :

B

M O C K O B C K O Ï Ï

P Y C H

i

C

O O T H O I I I E H H . H

H

B3AHMOFLEÑCTBHE

naMHTHHKOB nHCbMeHHocra (a

HHorfla TaKxce (jíojibiuiopa) H pejrarH03Horo HCKyccTBa B apeBHepyccKoñ KyjibType eme He oôcjiefloBaHo

CKOJIBKO

HHÔyzu» HcnepnbiBaiome

H J I H XOTJI

6bi B o6o6maiomHX onepKax, HecMOTpa Ha KpynHbie ycnexH B H3yneHHH apeBHe-

pyccKoro HCKyccTBa H apeBHepyccKHx jiHTepaiypHbix npoH3BeaeHHH, ocoöeHHo 3a nocjieflHHe iixecTb .necjmcoB jieT.1 IIcoTOMy MoaceT 6biTb onpaB^aHHoñ uejib

STOH

CTaTbH OÔpaTHTb BHHMaHHe Ha HeKOTOpbie HepTbl TaKOrO B3aHMO,HeHCTBHfl B X V I BEKE, HjuiiocTpHpyiomHe K TOMy »te £BN>KeHHe Hfleií B M O C K O B C K O H PycH.

1 BO3MO2KHO, Ï T O Taxoe HBJICHHC o6i>«CHaeTCa 3HaHO ycHJiHBineñcfl B 3TOM BeKe cneunajm3auneô (lacro KpafiHe y3Koö) HccjieflOBaTeneä. Ooth npHMep, tphhcah 6p0CHBnmäCH B raa3a: A. M. KHPUHHHHKOB onyÔJnucoBaJi B XypuaAe Mumcmepcmea Hapodnozo npoceeufeuuH (CamcTüeTepöypr), 1 8 7 8 ,fleica6pi>,2 6 8 - 3 2 7 ; 1 8 7 9 , ÄHBapb, 3 6 - 6 4 ; 1 8 7 9 , OeBpajib, 1 7 5 - 2 7 8 , oieia HHTepecHoe HccjieflOBaHHe: "CBHTOÔ reoprnñ h Eropañ Xpaôpwfi", rae, Mencny npoiHM, ynoMHHaeT H o T.Ha3. rioroÄHHCKoä HKOHe CB. reopnw c narmeM H O erapojiafloaccKHx (jjpecicax CB. reopnw, Ha KOTOPHX ecn. jno6ontiTHan «eTajn. onpefleneraoro BapiiaHTa jiereHfl (uapeBHa Beaer 3Men, o6fl3aB ero IIOHCOM ; a Ha IIoroflHHCKOÄ HKOHe 3a por). OflHaico, B CIIHCOK TpyaoB no HCTOPHH pyccKofi HKOHTI, cocTaBJieHHbiñ peaaKTopaMH MOHyMeHTajn>Horo Tpyaa H . I I . KoHAaicoBa Pycacan HKOM, I I I ; TeKCT, iacn> nepBaa (Ilpara, Seminarium Kondakovianum, 1 9 3 1 ) , 3Ta paöoTa KHPIIHIHHKOBa — OHeBHflHO, no 4>opMain>Hi.iM npnHHHaM — BKjnoieHa He 6i>ijia (CM. CTp. XV). B MOHorpa^im : Josef Myslivec, "Svaty Jin ve v^chodokfestanském umëni", Byzantinoslavica, V ( 1 9 3 3 - 3 4 ) , KHpmreHHKOB ynoMHHaeTC», c HHM Becerea nojieMHKa, HO HeT npsMoâ CCBIJIKH Ha ero paôoTy — co3flaeTca BneiaxjieHHe, HTO Myslivec Ha caMOMfleneormpaeTC« Ha paôoTy A. PticTemco, Jlezenda o ce. Teopzuu u dpaKOue e eusanmuücKOü u cAaenuopyccKoü Aumepamypax (Offecca, 1 9 0 9 ) , B KOTOPOÔ Bejica cnop c KHPUHHHHKOBIJM. B npeBocxoÄHoä H oiein» ueHHoñ KHHre B . H . JIa3apeBa, &pecKu Cmapoü Jladozu (MocKBa, 1 9 6 0 ) , 3 4 , ynoMraainw o KHprnwHincoBe yace BOBce HeT. Cneunanmaiow KaK 6w oTOflBHraeT paßoTbi o6meKyjn>TypHoro 3HaieHM. EH6jiHorpa(J>HHecKnx — onetn. pa3jnriHi>ix no oGteMy — oÖ3opoB, oxBaTbmaJonmx pe3yjn»TaTW myHemm HKOHOHHCH B XIX H XX Beicax, ecn, HeMajio. OFLHH H3 cymecTBeHHtix HamicaH B . H . JIa3apeBbiM B ero KHHre Hcnyccmeo Hoseopoda (MocKBa-JleHHHrpafl, 1947), 16-22. CM. TaKîKe pKfl HHTepecHbix noflpoÖHoereö B rrpeBocxoflHOM H3AaHHH Kamcutoz ôpeenepyccKoû oKueonucu (MocKBa, rocyaapcTBeHHafl TpeTbJtKOBCKaa rajiepea, 1 9 6 3 ) , 7 - 3 8 — onepK HanncaH B . H . AHTOHOBOA. 0 6 HCCJIEFLOBAHHH /ipeBHepyccKott jmTepaTypw HeKOTopoe noHarae «aeT 0630p HOBUX nyôjmKAQHÂ, CHCTEMATHHECKH BEJIYIQHÂCJI B Revue des Études Slaves.

64

HHKOJIAB AHflPEEB II

Ilpe»fle Bcero HafljieacHT npHHHTb bo BHHMaHHe hcckojibko npeflBapHTejibHbix 3aMeHaHHii, He Tpe6yiomHx cnemiajibHbix AOKa3aTejibCTB. Bo nepBtix, cjie^yeT iiomhhti>, hto b nepaofl bociiphhthh Pycbio 'BH3aHTHHCKoro HacjieflCTBa' HKOHonncb, KaK h ocraJibHbie bh^m xpHCTHaHCKoro acKyccTBa, yace onnpajiacb Ha orpoMHyio jiHTepaTypy, BcepaBHO 6mjih jih s t o EBaHrejine h flpyrae CBameHHbie khhth, xchthh cbstmx, nocTaHOBjieHHa ijepKOBHbix coGopoB hjih HHbie npoH3BefleHHH nHCbMeHHOCTH, oôbiHHo KaHOHHiecKH npoBepeHHbie. 3 t o iio3bojihjio I^CpKOBHblM aCHBOnHCIiaM C03flaBaTb npOH3BefleHHa H306pa3HTejIbHbIX HCKyCCTB B AHflaKTHnecKHX iiejiax 'HayneHHa' MajiorpaMorabix hjih bobcc HerpaMOTHbix.2 KoHeHHO, npn s t o m nojiynajiacb rpoMa^Haa aBTOHOMHaa HHepuna HcxyccTBa, oTMenaeMaa ero HCTopHKaMH, ho b cjioacHBnieiicH h CTporo-HepapxHiecKH-onpeaejieHHoii CHCTeMe pejiHrH03H0H XHBonHCH c ee aKTaM 0606meHHa: "IlycTb Bee xynoHcecTBeHHbie «bjichhh 3flecb npom,e, ho ohh pa3BHBaioTca 8 TeMa 'BH3aHTH»CKoro HacjieflCTBa' raynajiacb MHoroKpaTHO. HeKOTOpwe HHTepecHwe cooôpaxennn cm. A. Mazon, "Byzance et la Russie", Revue d'histoire de la philosophie et d'histoire générale de la civilization, 19 (1937), 261-277; A. Meyendorff and N. H. Baynes, "The Byzantine Inheritance in Russia", Byzantium, 1962, 369-391 H 6n6jiH0rpaiyjeHHa."6

HKOHOnHCH

"H3BeCTHbIH

flocraxceHHe

KoHflaicoB nepejIOM

Eßponw

ycMaTpHBaJi

B

c

KOHua

HKOHOrpa^HieCKOH

B

nepnoa

XIV

TaK

BeKa

TpaflHIXHH

B

Ha3bi-

pyccKoö

B

CTOpOHy

H y B C T B a H 3 K C n p e C C H H " , H e K O T O p y i O "pyCCH(|)HKaUHK)" HKOHHblX T H n O B . O H OTMeiaJI, HTO B X V I

BCKC B HKOHOIIHCH, B e e

OKa3bIBaeTCH yneHbiü öbuiH

COOTBeTCTBHe

BbMBHraji

Teopnio

"HenocpeacTBeHHbie

TO c r p e n e c K o r o

CJIO)KHbIX K 0 M n 0 3 H I i H H X ,

"HTajIbHHCKOMy HTO,

noMHMO

BblCOKOMy

o6mero

BJIHSIHHH H H 0 3 e M H b i x "

BOCTOKA, c X V

ÖOrOCJIOBCKHX

B03p0ÄfleHHK)";

ncTopHHecKoro o6pa3UOB,

TeMaX,

npH

3TOM

napaJUiejiH3Ma,

" e c j i H He H3

BeKa noflHHHHBineroca xyaoacecTBeHHOMy

HTajiHH, BJIH«HHK>

UTaJIHH."7 3Ta

oömaa

x a p a K T e p H C T H K a n y T e i i p a 3 B H r a a p y c c K o ö HKOHOHHCH ö b i j i a c j e j i a H a

H . I L K o H f l a K O B b i M flo T O T O , HTO OH M O T 0 3 H a K 0 M H T b c a c B e j i H K O J i e n H b i M H p e 3 y j i b TaTaMH, CKHX. 8

xoTa

HecMOTpa

nocTpoeHHÖ

yneHoro, OAHH

6bi, nepBbix

3HaT0Ka

ajia

B TocyflapcTBeHHbix

npoflojiaceHbi

ero,

KaK-6bi,

HHbix

XI03HUHH

'HKOHHOTO x y ^ o a c e c T B a ' —

HKOHonHCb

npeame

UepKBH

K0MN03HIMAX

•

("Bca

AeBH3)

H,

oömnx BbiBOflbi

fljia

aHajiH3a,

B cymHocra,

jiyninero

»CH3Hb

HeM

y

6ojibniHHCTBa

He

H

Bcero HTO

6biJio

no3flHeöiiiHX

c

HccjieflOBaTejieii

H e o 6 b i H a H H y i o Ä H H a M H K y p a 3 B H T H a STOH o ö J i a c T H a p e B H e -

pyccKofi KyjibTypbi. H . II. KoH^aKOB

HHorfla —

yneHbiH

H3 e r o

HccjieaoBaHHaMH,9

c o6o6meHHaMH

cneniHTb

CHHTe3a",

flaJibHenuiHMH

Macrep-

HMeHHO p y c c K O H C T a H K O B O H »CHBOnHCH, T . e . , HKOHbl, n O f l H e p K H B a i O T —

HecKOJibKO

3aAaHaM

pecraBpauHOHHbix

H a S T O T (j)aKT H H e c M O T p a H a T O , HTO H e i c o T o p b i e 6HJIH

He

HejiioÖHBmero

fleHb

pa6oT

flenaji

pejiHrH03H0e

B03HHKH0BeHHe nycTbiM

CBOH a H a j i H 3 b i , He y n y c K a a H3 BH/ia, HTO acxyccTBo, HKOHHblX

cjiy»cHBmee

TeM,

3 B Y K O M FLJIA J i r o ^ e i i

HKOHHaa T e x H H K a , 6 j i e c K H n r p a KOJiopHTa —

.zumaKTHHecKHM

CMbICJIOBOe

TOH s n o x n .

flBHJKeHHe

BnpTyo3Haa

B KOHCHHOM c n e T e —

B —

TOJibKO

COH., 8-9. HHTepecHO, HTO noao6m>ie BbiBOflbi, coBceM c HRUX n03nunfl

H . I I . KoHAaxoB, ymi.

H HHaie TpaKTyeMbie, H a n m a i o T B03HHKaTb B coßpeMeHHofi Hayice: K . O n a s c h , " R e n a i s s a n c e u n d Vorreformation Deutschen

in

der

Byzantinislawischen

Demokratischen

ÄHÖpeH PyÖAeea

u Enutfianua

Republik,

üpeMydpozo

H . I I . KoanaKOB, YKAI.

COH., 9.

8

H . I I . KoHßaicoB, YKA3.

COH., I X .

"

B . H . JIa3apeB, Hcmopun

'

Orthodoxie",

I (Berlin,

eusaHmuücKoü

1 9 5 7 ) ; H.

Aus

der

byzantinistischen

C . JInxaieB, KyAbmypa

Arbeit Pycu

(MocKBa-JleHHHrpafl, AKafleMHH C C C P , 1962).

sicueonucu,

I, 358.

der

epeMeuu

66

HHKOJIAH AHflPEEB flocTiraeHHii

e p e f l C T B o , flJiH Il03T0My

BblBOflbl

HBJieHHH,

HHOrfla

pejiHrH03Hbix uejieñ

KOH^aKOBa

JIHUieHM

CBOHCTBCHHilX

hjih

COBpeMeHHbIM

iiojihthkh.

u e j i e ñ uepKOBHOH

KpaHHOCTeñ

HayHHOH

nOCTpoeHHHM,

b c B o e ñ hctophhhocth. 10 K o H f l a i c o B H H K o r a a H e H C T O p H i e C K H X « B J i e H H H — n o 3T0H n p H H H H e H M e H H O

TepaioT

MOflepHH3aUHH KOTOpbie

Torfla

3 a 6 b i B a e T o cneuH(j)HHHOCTH ero

B b l B O f l b l B OTHOIUeHHH

pyCCKOH HKOHOnHCH C y m e C T B e H H b l flJIH T e M b I í t a H H O H C T a T b H . B

TpeTbHX,

M05KH0

CHHTaTb

BepHbIM

H

flOCTETOHHO

(nocKOJibKy T e M a B3anMOfleñcTBH$i jiHTepaTypbi

HaTJIHflHblM

h hkohoixhch

Ha P y c n

nOJIOJKeHHe

bcc

ace,

xoth

He H c n e p n b i B a i o m e , b KaKoii-To M e p e o c B e m e H a HccjieflOBaHHHMH), h t o n p H co3flaHHH

hkoh

c 6ojiee

hjih

MeHee

hhcto

pyccKHM coaepacaHHeM, onpeaejieHHbiM

cneimHKOH

p y C C K O r O M e C T O p a 3 B H T H H , p y C C K H e M a C T e p a C ÔOJIbHIHM B H H M a H H e M OTHOCHTCH K npeflaHHHM, oôjieneHHbiM 0CH0By

h

b tckcth, hx

KOTOpbie H a n p a B j i a i O T

HMeHHO T e x c T b i ,

b

KOTOpbie a a i o T

hm

KOHKpeTHyio,

xyfloacecTBeHHyio TexHHKy.

n e p B y i o o n e p e ^ b ecTecTBeHHo,

3khthhhoi"o

oaoaeMyio

Hhhmh

cjiOBaMH,

xapaKTepa,

onpefle-

JIHIOT C03,HaHHe H H K O H O n H C H O r o T H n a CBHTbIX, ' n p O C H H B U I H X B PyCCKOH 3 e M J i e ' , H (jjaKTHiecKHe TeMbi

Moncho

xchthkhmx

KJieiiM.

y K a 3 a T b , KaK H a B e c b M a x a p a K T e p H b i ñ n p H M e p

HjunocTpHpoBaHHe

apeBHepyccKHMH

MacTepaMH

b stom othoihchkh,

JiHTepaTypHbix

npoH3BeaeHHH

' c T p a c T O T e p n u a x ' , KHH3bHX-6paTbHX E o p i i c e h T j i e ó e , n e p B b i x pyccKHX j i e H u e B 3Jiy

chjioio', h

Ha BbipaôoTKy pyccKHMH xcHBonHciiaMH

n e p B b i x CBHTblX, K a H 0 H H 3 0 B a H H b I X pyCCKOIO I i e p K O B b K )

(xoth

Ha o

'HenpoTHB-

HKOHHoro

sthx

rana

OHH H H e ÔbIJIH

nep-

bwmh — b x p o H O J i o r H H e c K O M o T H o u i e H H H — H 3 c B f l T b i x H a P y c n ) . 1 1 3 T a T e M a — 0 6 H 3 o 6 p a a c e H H a x B o p n c a h T j i e ô a — n o j i y n n j i a H e K O T o p b i ñ otkjihk b c n e i m a j i b H O H J i H T e p a T y p e , xoth b n e p B y i o o n e p e A b B H H M a H H e H C C J i e f l O B a T e j i e f t 6 b i j i o n o c B a m e H O MHHHaTiopaM 3aHHBIIiaflCH

J i H i i e B b i x c n H C K O B CKCIMHUH. HMeHHO

flpeBHepyCCKOH

OflHaKo, HeaaBHo

CTaHKOBOH

OTpaaceHbi jiHTepaTypHbie npoH3BefleHHH o E o p n c e xopouio

flOKyMeHTHpoBaHHOH

noHBHJiacb

XCHBOnHCbK),

h

nOCKOJIbKy

Fjieôe.12 K coacajieHHio,

p a ô o T e a B T o p BOBce He o ô c y a y i a e T B o n p o c ,

paôoTa, B

Hen

b stoh moxcho

j i h cHHTaTb H K O H o n n c H b i e THnbi cBHTbix ô p a T b e B b KaKoii-jiHÔo M e p e n o p T p e T H b i M H . O p H H H M a n b o B H H M a H H e cj)aKT o n p e a e j i e H H o f l y c T o i Í H H B o c T H T H n a B o p n c a h 10

N . Andreyev, "Nikon and Avvakum on Icon-painting", Revue des Études Slaves, 38 ( =

Pierre

Pascal)

Tjieôa

Mélanges

( P a r i s , 1961), 37-44.

11 T . I I . 4>e«0T0B, Cenmbie JJpeeneü Pycu (X-XVII cm.) ( P a r i s , Y M C A P r e s s , 1931), 19. PyccKan iíepKOBb n o 3 x e n p H 3 H a n a KynbT B a p a r o B — MyieHHKOB 3 a Bepy, 4>eflopa h H o a i r a a , n o r n 6 n m x b H3bIHeCKHË nepHOfl nOJIHTHKH BjiaflHMHpa CBHTOCJiaBOBHTO, H 3aTeM KHarHHH O j I t r H H KHH3H BjiaflHMHpa I, npnpaBHHB h x b pyccKoft uepKOBHOñ cHCTeMe — k paBHoanocTOJibHWM n p o c B e r a TejwM P y c u . 12

3 . C . CMHpHOBa, " O T p a » e H H e JiHTepaTypHbix npoH3BeaeHHô o B o p a c e h T j i e ô e b

pyccKoñ CTaHKOBOü HCHBonHCH", Tpyàbi omdem

àpeeuepyccKoii Aumepamypu,

flpeBHe-

15 (1958), 312-327.

HeoôxoflHMO npMBCTCTBOBaTb noHBJieHHe 3toS p a ô o T w , HOTOMy hto, KOHerao, n p o c r o He0 6 x 0 f l H M 0 npOH3BeCTH paCCMOTpeHHH 3TO0 TeMbI COOTHOIUeHHH JiHTepaTypHbix H HKOHOITHCHblX naMHTHHKOB «peBHeñ PycH b CBeTe r e n e p e m H e r o 3HaHHH o pa3BHTHH uepKOBHoit cTamcoBoS 3KHB0IIHCH. H a f l O HafleHTbC», HTO 3 . C . CMHpHOBa nOflpOÔHO 3aÔMeTC« 3T0ÍÍ TeMOft H Ha MaTepHane Mockobckoë P y c H : cm. e e HHTepecHyio CTaTtro " 0 6 oahom jiHTepaTypHOM cio»ceTe b aareonHCH

KOHna XVI B.", Tpyàw omdeAa dpeenepyccKOÙ Aumepamypu, 16 (1960), 365-373.

67

JIHTEPATyPA H HKOHOÜHCb

Ha COXpaHHBIIIHXCH HKOHaX H TO OÖCTOHTejIbCTBO, HTO cjiyac6a H M ÖbIJia COCTaBJleHa MeacAy 1021 h 1039 roflaMH, M O X H O ayMaTb, HTO nepBbie H K O H H C S T H M H C B H T M M H , yÖHeHHBIMH B 1015 ro/iy, AOJDKHbl 6bIJIH OTpa»CaTb BOCIIOMHHaHHS O BHeiUHOCTH CBHTMX

öpaTbeB.13 TaicaH cKjioHHocTb

K

nopTpeTHoeTH ecTb 6e3ycüOBHaa, nepTa

HKOHOÜHCH B O O Ö m e H pyCCKOH, B HaCTHOCTH.14 OflHaKO, pHflOM M03KH0 npHBCCTH H H H O H npHMep, KOTfla HKOHOnHCUbl He MOrjIH, noBHflHMOMy, cjieflOBaTb 3TOH noroHe 3a noprpeTHbiM np0T0THn0M, co3flaBaTb TpaOTUHK) nopTpeTa. Ilpn co3flaHHH H3o6pa>KeHH» BjiaziHMHpa CBHTOTO HKOHOnncHbie MacTepa aojixcHbi

ÖHJIH

aaTb 'HfleajibHbiä o6pa3' 3Toro

Heo6biHaHHO

rHÖKOrO nOJIHTHKa H OflHOTO H3 OCHOBOnOJIOÄHHKOB XpHCTHaHCKOH KyjIbTypbl Ha PyCH.15 ÜBHO, HTO HKOHbl, H3o6paJKaK>mHe ero KaK CBHTOrO, MOrjIH nOHBHTbCH TOJibKo nocjie ero KaHOHH3aunn, T.e., noBHflHMOMy, TOJibKO BO BTopofi nojioBHHe XIII BeKa, yace nocjie MOHrojibCKoro HamecTBH« H pa3pymeHHH Tex naMSTHHKOB, rae

MOTJIH

6biTb H3o6pa»ceHHH BejiHKoro

KH»3JI,

KaK aoHaTopa. 16

Il03flHHH KaHOHH3aUHH KHH35I BjiaflHMHpa o6l>HCHHeTC5I OTCyTCTBHeM eflHHOH TOHKH 3pCHHH Ha eTO JIHHHOCTb, B HaCTHOCTH, TpeHeCKHMH C O M H e H H H M H OTHOCHTeJIbHO ero 'CBHTOCTH* —

HanpOTHB, pyCCKHe aBTOpbl OÖblHHO HaXOflHT H CHJIbHbie CJIOBa H

yßeacfleHHbie apryMeHTw B nojib3y "KaraHa BjiaflHMHpa".17 ECJIH

r. II. O e A O T O B npaB,

OTHOCH K

XIII BeKy o(j)HijHajibHyK> KaH0HH3aimi0

BO

'BcepyccKOM Macurra6e' (H6O He HCKJHoneHa B03M0)KH0CTb, HTO Morjio cymecTBOBaTb öojiee paHHee MecTHoe noHHTaHHe, 18

XOTH

STO

Majio BepoaTHo),18

IIOHHTHO,

HTO

r. FI. OeflOTOB, yKü3. cot., 20. 3. C. CMHpHOBa, yKü3. cot., 313, npHHHMaeT npeflnojKUKeHHe, cjiynae 6biJiH HaimcaHbi B 1020 HJIH 1026 r. 14 H. n . KonnaKOB, yKa3. cou., 11-19. KoHeiHO, 3Ta nopTpeTHocrb ycnoBHaa, CTHjTrooBaHHaH, O6YCJIOBJIEHHAH H TCXHHKOÖ, HO BancHo OTMETHTB, HTO OHa cymecTByeT H HTO TPAFLHUWSI EE B OTHOineHKH OTflejitHtix cBHTbix Hpe3BbraaflHO ycToöHHBa. 15 H. H. OKyHeBa, "H3o6pa«eHHH CBHTOTO BjiaflHMHpa", BAaduMupcKuü CöopnuK e na/uamb 950-jtemu/t Kpeufenun Pycu (Eejirpafl, [1938]), 197-200. HeK0T0pbie nepTbi nopTperaocTH, KaK yxa3biBaeT H. H. OKyHeBa, M O D M 6bi 6biTb HaiiaeHti Ha MOHerax h BHCJIOÜ CBHHUOBOÖ nenaTH BejiHKoro KHH3H. H o HKOHOimcHbi noinjiH HHbiM nyTeM, KaK H MHHHaTiopHCTbi CKQ3QHUH o EopHce H rjie6e, H co3flajiH THII 6naroo6pa3Horo ce/joro Myapeua. 16 r. n. $eflOTOB, "KaHOHH3amm CBHToro BnaflHMHpa", maM otce, 191. 17 M . / J . NPHCEJNCOB, OnepKU no ifepKoenoü u nonummecKoü ucmopuu Kueecxoü Pycu X-Xll ee, (CaHKT-IIeTep6ypr, 1913), 104, 106-108, 303 H CJI. HeKOTopwe ero aaea pa3BHTbi: H. Koch. "Byzanz, Ochrid und Kiew", Kyrios, 4 (1938). A. B. KapTameB, Onepxu no ucmopuupyccKOÜ ifepiceu, I (IlapiDK, Y M C A Press, 1959), 105-138, 157-170. B. A. P030B, "JIHHHOCTb BjiaflKMupa CßHToro B pyccKoä JiHTepaType", BmduMupcKuü CöopnuK, 155-173. Paa HHTepecHbix cooöpanceHHfi no TeMe, TaK hjih HHaie, CB«3aHHbix cflaHHoönpoöJieMaTHKoö, Haxo/WTCH, HaHHHaH c VIIraaBbi,B KHnre M . B. JleBneHKO, Onepm no ucmopuu pyccKO-eu3cmmuücKUx omHouienuü (MocKBa, 1956) — ueHHbie 6ü6jiHorpaj)HHecKHe H OTiacTH npo6jieMHbieflonojmeHH«K STOÜ paöoTeflenaeTA. B. OjiopoBCKHä, "K H3yHeHHM HCTOPHH pyccKO-BH3aHTHflcKHX OTHOineHHfi", Byzantinoslavica, 20 (1959), 63-74. L. Müller, "Zum Problem des hierarchischen Status und der jurisdiktionellen Abhängigkeit der russischen Kirche vor 1039", Osteuropa und der deutsche Osten, Beiträge aus Forschungsarbeiten und Vorträgen der Hochschulen des Landes Nordrhein- Westfalen, 3 (1959), OTpmiaeT o6ocHOBaHHOCTb nmoTe3bi IIpnceJiKOBa H ero CTOPOHHHKOB. K coJKajiemno, eMy ocTanca, no-BHflHMOMy, HEH3BECTEH BjiaduMupcKuü CöopnuK (Eejirpafl, 1938), H, B HACTHOCTH, BBRMEHA3BAHHAFL craTb» T. II. OEFLOTOBA O KaHOHH3aUHH BjiaflHMHpa. 18 H. HaKOJibCKHfi, Mam. ÖÄH noepeMeumao cnucna dpeeuepyccKux nucameAeü (CFI6., 1906), 232. HTO IIKOHI>I BO BCHKOM

68

HHKOJIAH AHflPEEB

HKOHonHciíbi eflBa JIH yace MOHIH CHHTaTbca c YCTHHMH ONHCAHHHMH BHCIIIHOCTH BjiaflHMHpa I HJIH npHHHMaTb BO BHHMaHHe cxeMaTH3HpoBaHHbie H3o6paaceHHH ero Ha MOHeTax. HKOHonHciiti noiiuiH no HHOH JIHHHH, a HMCHHO CMMCJIOBOH, JioraiecKOiiepKOBHOH, coBepmeHHo ecTecTBeHHoií B TOH 06cTaH0BKe: OHH co3flajiH Ha HKOHax THn npecxapejioro Myapeua, ce,noBJiacoro CTapua, xoTa HCTOPHKH C npaBOM coMHeBaioTca, Mor JIH 6BI SHTB TaKHM KHH3B BjiaflHMHp, yMepiirafi, NOBH^HMOMY, B B03pacTe 5 3 - 5 5 J i e T . 1 9 OflHaico, AJIH iiicoHoniicna B X I I I BeKe KHH3B BJIAAHMHP paci^eHHBajica yace — rjiaBHbiM 06pa30M — no JIHHHH anocTOJibCKOH, KaK oxapaicTepH30Baji ero nepBbiñ naHerapHCT, npecBHTep ABopuoBoñ IJEPKBH B EepecTOBe y BejiHKoro KHH35i .flpocuaBa M y a p o r o , MjiapHOH, 6yAymHH nepBbiñ MHTponojiHT H3 pyccKHx B RHCBC: " . . . xBaJiHTt »ce noxBajibHbiMH rjiacbi PHMbcKaa CTpaHa IleTpa H IlaBjia, HMHace BtpoBaTa BÍ> Icyca XpncTa Cbma Eoacia; Acia H Eijjecb, H üaTMb — IoaHHa BorocjioBa; HHÍÚH — 4>0My, ErnneT — Mapica; BCH cTpaHbi, H r p a ^ w H jiioflie HbTyTb H cjiaBHTb KoeroacAO HX yorrejui, nace HayHHina npaBocJiaBHoñ B t p i . IIoxBajiHMb ace H MM ... BnajiHMepa. ,.." 2 0 BjiaflHMHp noAo6en anocTOJiaM, 'paBHOanOCTOJIbHblH KHH3b\ HcyflHBHTejlbHO n03T0My npOHHTaTb B CBOflHOM TOJIKOBOM noíuiHHHHice XVIII BeKa yace BOBce HUHaJIbHbIM pyKOBOflCTBOM flJIH MacTepoB, a Jinuib nocoóneM. Hx cocTaBHTejm npHAepacHBajiHCb n p n 3TOM onpe/iejieHHoñ cxeMbi: (1) B03pacT H3o6paacaeMoro CBHToro, (2) BOJIOCM na ro/roBe H Gopofle, (3) oaeac,zia. n o B03pacTy Bce H3o6paacaeMbie cBHTbie pa3flejiaioTca Ha Mojioflbix, cpeflHHx jieT ('cpe/ioBeic'), erapHKOB ('cTapen;'). B paccMOTpeHHbix Bbiine YMCCTHO NOANEPKHYTB, HTO

19

H . H . IleTpoB, "/IpeBHHH H3o6paHceHHa CB. BjiaflHMHpa", Tpydw KueecKoü

ffyxoenoü

AmdeMUit,

1915, Hiojii-aBrycT, 354-356. Cp. Taicace H. H. OicyHeBa, yRaí. con., 198. ]\o iero 6wjia 3a6i>iTa BHenmocTi. KH. BnaflHMHpa noKa3biBaeT 4>aKT, MTO BMCCTO Hero HHoraa H3o6paacaroT Bacmum BeJiuxoro. TÜM jice. 20 H . K. ryfl3HH, XpectnoMamuH no ópeaneü pyccKoü Aumepamype XI-XVII

sexos (MocKBa, 1952),

31-32. Cp. HHTepecHwe saMcnaHHH o "Críeme o 3aKOHe H EjiaronaTH": C. JlHxaieB, PyccKue Aemonucu u ux KyAbmypuo-ucmopmecKoe ínanenue (MocKBa-JleHHHrpafl, 0000), 51-58, a Tanate H. y. ByflOBHHii, OóufecmeeHHO-no/iumutecKaH MHCM ffpeeneü Pycu (MocKBa, 1960), 43, 65-72. 21 H. H. OKyHeBa, ym3. coi., 199-200. CM. HKOHBI CB. BjiaflHMHpa: H. N . JlaxaneB, MamepuaAbi djw ucmopuu pyccnazo UKOHonucauuH, I, JVS 207 (CaHKT-IleTep6ypr, 1900). H. II. JIüxaieB, JIuifeeoe oteumue ... ce. Eopuca u raeSa, Ta6jinua I. O npopHcax HKOH CB. BnaflHMHpa B HK0H0nHCHi>ix noflJIHHHHKOB — JlHxaieB, Jluifeeoe oteumue ..., 40; H. H. neTpoB, yxas. cot., 358-360. KaTanor apesHepyccKofl acHBonHCH, II (MocKBa, 1963). Onnc. 410, Ta6ji. 17, 60-61.

69

J I H T E P A T y P A H HKOHOITHCB

npHMepax E o p a c — flaxce

no/uiHHHHKH

cpeaoBeK, D I E S — o6HapyacHBaioT,

Kmeit, BJIA/iHMHp —

KaK

noanepKHBaeT

HX

c T a p e u . I l p n STOM nccjieflOBaTejib, He-

npecTaHHoe pa3BHTHe npHMeHHTejibHo K BonpocaM BpeMeHH. IIOOTHHHHKH 6HJIH C03flaHHeM caMHX HKOHonHcueB fljia yaepxcaHHa B n a M a r a

npe^aHMÍi, npaKTHKH

HKOHonHcaHHS, " o KaKHX JIH6O 3aMMCJiax CBH3aTb npH nocpe/icTBe noajiHHHHKa CBo6o^y HCTHHHoro TBopnecTBa 3flecb He M o r j i o 6biTb H p e ™ " — CBeTCKHe BJiacTH HHKorjia Ha HHX Ha PycH He TaKHM 06pa30M,

OKa3HBaeTca —

AYXOBHBIE HJIH

oirapajiiicb. 2 2

H3 C0B0KYNH0CTH Bcex yKa3aHHbix nepT

—

Heo6xo£HMbiM cflejiaTb BMBOJI, HTO p y c a c a a craHKOBaa xcHBonncb, HecMOTpa Ha CBOK) eCTeCTBeHHyiO TpajIHUHOHHOCTb B MaCTepCTBe H CBH3aHHOCTb TeKCTaMH H npe^aHHeM B coflepacaHHH cymecTByiomHx K0Mn03HiiHH, GbiJia o 6 j i a c T b i o HeH3MeHHoro ABH»ceHHH, pa3BHTH», HanpaKeHHoro a a e i m o r o

TBopnecTBa, H6O, KOHCMHO,

caMa MaccoBocTb pyccKoií CTAHKOBOH 5KHBonncn, 6oraTCTBo ee TeMaTHKH H pa3Hoo6pa3HocTb pyccKHx HKOHonHCHbix UIKOJI noKa3biBaioT, HTO HKOHbi 6MJIH OTpaaceHHeM

ToraaniHero

pyccKoro

co3HaHHa,

'yM03peHHa

B

KpacKax',

no

jieTyneMy

BbipaaceHHK) KHa3a E . H . T p y 6 e m c o r o . JXJiaFLAHHOIÍCTaTbH OCO6CHHO BAACHO n o a nepKHyTb,

HTO H B 3TOT —

co6biraaM

—

eme

nepHOfl pa3BHTHa

oópameHHbiií pyccKoñ

K HCTopHHecKHM npe^AHHAM H

CTaHKOBOH

acHBonHCH,

npH C03flaHHH

HOBblX pyCCKHX HKOHHblX TeM, HKOHOIIHCUbl OnHpaiOTCa Ha HfleH, Ha TeKCTbl, TaK HJIH HHane cBa3aHHbie c npo6jieMaTHKoñ, c co3HaHHeM p y c c K o r o o6mecTBa TOH snoxH. HKOHONHCB FLBHKETCA B Ty s n o x y 3a nHCbMeHHocTbio, oTpaacaa JiHTepaTypHbie NPOH3BE,NEHHA HJIH c j i e ^ y a HX OCHOBHMM H ^ E A M . 2 3

III

X V I BeK — fljia Hauieñ TeMbi —

0Ka3biBaeTca OCO6CHHO Bbipa3HTejibHbiM. Eivty, KaK

H Bceñ MOCKOBCKOH PycH, a o j i r o 0TKa3bIBaJIH B flHHaMHHHOCTH, B nO/rbeMe HaeñHOH XH3HH, B coócTBeHHbix nepTax co3HaHHa. HauiyMeBiiiHe onpeaejieHHa H . A . E e p a a e B a KaK-6bI nOaBO^HJIH HTOr HejIOH s n o x e OTHOUieHHa K B03HHKH0BeHHI0 'HMnepHH' 22

H . B . RIOKPOBCKHÁ, IJepKoewH

apxeoAozun

e cen3u

c ucmopueü

xpucmuaucxoao

ucKyccmea

( n e T p o r p a a , 1916), 2 1 1 - 2 1 7 . as

E . H . T p y ó e u K o ñ , yM03peuue e Kpacmx ( I f e r p o r p a a , 1916). 3aMeHaTenbH0, ITO aaace H B nepaoflbi «BHoro ocjia6jiemw BHHMAHHH KFLPAMATH3NP0BAHH0MYpaccKa3y H HFLEOJIORMECKOMY OÓMCHeHHIO B pyCCKOfi HKOHOI1HCH npHXOflHTCJt, flJIH HOJIHOrO HCTOpmeCKOrO nOHHMaHH« ee, BCNOMHHATB MHp j i H T e p a T y p b i H ugefi. B X I V Bexe jiHpHiHOCTb, x a p a K T e p H a a a n a Bceñ E B p o n w B 0 T H 0 m e H H H o 6 p a 3 a E o i K H e ñ M a T e p n c Mjia/ieHueM, cooTBeTCTByeTFLBJIEMWMT o r f l a n m e f t m r r e p a T y p w CHaHTe, n e T p a p x a H n p o i H e ) , npeJioMJieHHbiM, i c o H e m o , B YCJIOBHOCTHX qepKOBHoro n c K y c c T B a ( H . N . K o n a a K O B , ymí. con., I V , 207). I I p H HMCHH O e o ^ a H a Tpeica HCTOPHK HCKyccTBa coBepmaeT 3 K c x y p c B H C T o p m o HfleñHoS 6 o p b 6 b i najiaMHTOB c BapjiaaMHTaMH, 3anyMi.iBaeTca H a n n o ó e a o S HCHxaCTOB B B»3aHTHH. B . H . JIa3apeB, eo$aH rpex u eeo IUKOJIÜ ( M o c K B a , 1961), 14-34. I I p H HMCHH A H f l p e a P y 6 n e B a HCTOPHK n c K y c c T B a CTPCMHTCH o c o 3 H a T b T y yMCTBeHHyro aTMoccj)epy, KOTOpa» BO MHoroM onpefleJiHJia pyóneBCKoe MHpoomymeHHe, H «aace onnpaeTCH B 3TOM Ha flaHHbie 0 6 naeax H w i a C o p c K o r o , XOTH OH »CHJI NO3«HEE Bejiratoro MacTepa HKOHOÜHCH (B. H . JIa3apeB, ÁHdpeü

PyÓAee

u eeo uiKOAa

[ M o c K B a , 1966], 10-13).

70

HHKOJIAH A H ß P E E B

rpo3Horo,

MHoroHaixHOHajibHoro

rocyaapcTBa,

ueHTpaJirooBaHHoii

aepacaBbi:

"MOCKOBCKHH n e p n o f l 6 H J I c a M b i M n j i o x H M n e p H O f l O M B p y c o c o f i HCTOPHH, c a M b i M AYIUHBIM, H a n 6 o j i e e a 3 n a T C K o - T a T a p c K H M n o C B o e M y n i n y . . . . X a p a K T e p H O HCHC3HOB6HHC CBHTbIX KHH3ÊH ÜOCJie n e p e H e c e H H f l r p e X O B H O H BJiaCTH H a

BejIHKHX KH33ÊH

MOCKOBCKHX. . . . HOCH«}) BOJIOUKOH — p O K O B a H ( { w r y p a H e TOJIbKO B HCTOPHH n p a B o CJIABHH, HO H B HCTOPHH p y c c K o r o

uapcTBa.

BMECTE c

HoaHHOM

rpo3HMM

HyjKHO C H H T a T b r j i a B H b i M o ö o c H O B a T e j i e M p y c c K o r o c a M O f l e p a c a B H H . . . . "

24

ero

EepaaeB,

K c o a c a j i e H H K ) , o ô j i a a a j i M a j i b i M n y B C T B O M HCTOPHHHOCTH. H a c a M O M a e j i e MOCKOBCKAA P y c b n o J i H a H ^ e H H o r o ABHÄCHH» H —

B NACTHOCTH, e e X V I BCK HPE3BBRAIIHO

o a c H B J i e H H b i i i H n o j i H b i H B a x c H e f i n i e r o c o a e p a c a H H H 3 T a n B p a 3 B H T H H p y c c K o ö MHCJIH, 3 T a n H f l e H H o f t 6 o p b 6 b i e ^ B a JIH H e BO B c e x o ô j i a c i a x T o r f l a u i H e f i K y j i b T y p b i . N P A B ^ a , 3TO H f l e Ü H O e OXCHBJieHHe HaCTO BOBCe H e OTrOJIOCKH B 0 3 p 0 > K A e H H 5 I HJIH P e ( j ) O p M a i I H H , KaK

XOTejTOCb

6bl

BHfleTb

HeKOTOpbIM

HCTOpHKaM.25

3TO

HfleHHOe

OaCHBJieHHe

CB»3aHO, B n e p B y i o o n e p e ^ b , c n o 6 e , a a M H T o r o K p b i j i a uepKBH, K O T o p o e

Ha3biBaioT

' c T H X c a T e j i H M H ' . B e p o H T H o , e c j i H 6 b i H e 6 w j i o HX n o 6 e f l b i B 1 4 9 0 r o s y H a s 'JKHJIOBCTBYIOMHMH' H B 1 5 0 3 r o a y

Haa

'HecTHacaTejiaMH',

HCTopna

HAeiiHoro

pa3BHTHfl

MOCKOBCKOH P y c H 6 b w a 6 b i c y m e c T B e H H o HHOH. T e n e p b O H a , 0 ; m a K 0 , O K p a m e H a B noôeaoHOCHbie

TOHa HocH(J)jiHHCKoro n o H H M a H H H nepKOBHOH ÄH3HH. 3 T a

no6eaa

' c T H a c a T e j i e ö ' , 6 e 3 COMHÊHHJI, 6 b i J i a B b i r o f l H o ö fljia p y c c K o r o n c K y c c T B a , j u i n u e p K O B H o r o C T p O H T e j I b C T B a , itJTH p a 3 B H T H H p a 3 H b I X BHflOB U e p K O B H O r O HCKyCCTBa : n o ô e ^ o HOCHAA u e p K O B b

xoTejia

BHAHMOI-O

npocnaBjieHHH

CBoero

coio3a

c

BepxoBHon

B J i a C T b I O , HTO H O T p a H C a j I O C b — C r p O M a / I H O H HHTeHCHBHOCTbK) B HCKyCCTBe X V I B e K a MOCKOBCKOH P y c H . 2 8 C o K ) 3 B e p X O B H O H BJiaCTH H HOCH e p y HAeHHMX HCKaHHH, CKJIOHHOCTb K p a 3 f l y M b K ) , K n O H C K â M H 4 > o p M y j i H p o B K a M HOBbix noHHTHH.27

ECJIH B X V

B e K e 3HAHHTEJIBHOE

HanpaaceHHe

" "

H m c o n a ö E e p f l a e B , Pyccxaa udea (IlapHHC, Y M C A P r e s s , 1946), 7, 9, 11. P i e r r e K o v a l e v s k y , Manuel d'Histoire Russe ( P a r i s , P a y o t , 1948), 9 0 - 1 3 6 .

"

SI. C. Jlypbe, HdeoAOBuiecKan 6opb6a e pyccKoü nyÔAuifucmme KOHifa XV — Havana XVI eexa

(MocKBa-JIeHHHrpafl, 1960). 87 KOHCÎHO, nyÖJTHKaiDffl AicafleMHH H a y x C C C P n o HCTOPHH n e p n o f l a 6 p o c a i o T i a c T o H0B0e o c B e m e r a e Ha M a T e p n a n : H . A . Ka3aKOBa H Si. C . J l y p b e , AHtnucßeodaAbHbie epemmecxue dewncemx

Ha Pycu XIV Havana XVI eem (MocKBa-JIeHHHrpafl, 1955); üocAamri Hocutßa Bo/toifKozo (craTbH H . I L EpeMHHa H K. C . J l y p b e ) ( M o c K B a - J I e m m r p a f l , 1 9 5 9 ) ; Ä. C . J l y p b e , HdeojiozmecKan 6opb6a e pyccKoü nyÔAuifucmme KOHifa XV— nanam XVI eexa (MocKBa-JIeHHHrpa.il, 1 9 6 0 ) ; A . A . 3HMHH, M. C. Ilepeceemoe u eeo coepeMemuKu (MocKBa, 1 9 5 8 ) ; A . H . KjinöaHOB, PetßopMaifuomue

deuMcemiH e Poccuu e XIV—

nepeoü nojtoeune XVI ee. (MocKBa, 1960). HecoMHeimo, 3TH paßoTH

HCKjnoHHTentHO HHTepecHbi, XOTH, caMO c o 6 o f t pa3yMeeTcn, cymecTByeT p a n nyHKTOB p a c x o a m e r a a Meatfly HX aBTopaMH, a Taicace B03M0aan>i H H o r a a coMHeHHH H C TOUCH 3peHHfl TOTatejw, n o f f i a c

JIHTEPATyPA H HKOHOIIHCb co3aaBano oxomaime

71

1492 r o a a , Koraa npeacKasbiBajiact 'KOHHHHa Mnpa', TO,

n o c j i e oTcpOHHBaHHH 3TOH flaTbi 'BceMHpHoro npnuiecTBHH XpucTOBa' e m e

Ha

KaKyK)-TO nacTb

Ha

THCHHH

JieT

(B

nepBbiä njiaH BbicrynHJiH HHbie

"IfojioxceHHH nacxajiHH" MHTponojiHTa MOTHBH,

6ojiee 3eMHbie,

NEM

3OCHMM),

3cxaT0Ji0rna.

B

caMOM

"M3JioaceHHH nacxannH Ha ocMyK) Tbica^y JieT . . . B Heft ace naeM BceMHpHoro npHiuecTBHH XpncTOBa" BbiABHrajiacb — BnepBbie Ha p y c c K o ö noHBe — Hjjea o "HOBOM

r p a a e KoHCTaHTHHa — MocKBe", KaK

BCEJIEHCKOM

ueHTpe. 2 8 Y

MHTPO-

noJiHTa 3OCHMH 3Ta Hflea Bbicica3aHa anojioreTHHecKii: "HOBWH KoHCTaHTHH" — HßAH

III, 'noBejieBiiiHii' cocTaBHTb H3JioxceHHe. H o o n e m . c x o p o

(B 1 4 9 9 - 1 5 0 1

r.)

3Ta TeMa no^BepraeTca cßoeo6pa3HOMy nepeocMwcjieHHio B nocjiaHHH «PHJIO^CH c r a p u a E j i H a 3 a p o B a MOHacTbipa B "XPHCTOJIK>6HBOH CBOÖO/IHOH 3eMJie IICKOB",

nocjiaHHOMy

HM

HßaHy I I I

B C B J B H CO

BTopofi — oneHb SHaiHTenbHoii — a o

8 0 %

—

KOH(J)HCKamieH BejiHKHM KHH3eM uepKOBHbix 3eMejib B H o B r o p o , a e . 2 9 3 T O nOCJiaHHe HJio(J)eH BOBce He ecTb ariojiorHH MOCKBBI, HO ecTb, B cymHOCTH, TpaKTaT B 3amnTy IiepKOBHHX HMymeCTB, TpaKTaT npOTHB OTpHUaTeJIbHblX aBJieHHH B 3KH3HH UepKBH HJIH B pyCCKOM UepKOBHOM 6bITe. E r o uejlb HailOMHHTb BejIHKOMy KHH3K) 0 6 e r o 06jmHH0CTH 3aiuHmaTb npaBocjiaBHyio uepKOBb,

"nane coJiHua

B

BcejieHHofi

CBCTHTCS"

—

STO

—

H6O

"MocKBa

—

TPCTHH

PHM",

icpenKHH, y6e»(fleHHbm roJioc

eme a o n e p n o a a noöeflbi HOCHCJUIHH, HaMeTHBUiefica B 1 5 0 3 r o s y , X O T H 6opb6a co CTopoHHHKaMH HHJia CopcKoro NPOFLOJIXCHTCA BCK> nepByio N O J I O B H H Y

CTJRACATEJW,

X V I Beica. 30 ÜMeHHO NOFL 3THM — CTSDKaTejIbCKHM — 3HaKOM H npOHfleT 5KH3Hb pyCCKOH C T A H K O B O H XCHBONHCH

X V I Beica, H 6 O "BoroMnpocBemeHHas 3eMJis pyccKaa" — n o a

pyKOBOflCTBOM HOCHtjuiflH 6yfleT CTpeMHTbCH H K yCHJieHHOMy LtepKOBHOMy CTPOHHE BNOJME YÖENCNEHHORO HOBATOPCICOÄ OCTPOTOÄ NOAXOFLA, NPN KOTOPOM HHORAA CCTCCTBCHHBI EBTOPCKHC YBJIETEHH« H — B

KaKoft

TO MEPE — NPEYBEJIHHEHHA B OTFLEJIBHBIX OUEMCAX HJM B U B O W .

HCIÜII0HHTEJII.H0 IJEHHA, NPA 3TOM, NY6NHKAUHA «PEBHEPYCCKHX TCKCTOB. " C . JLYPTE, HdeoAozmecKast 6opb6a . . . (MOCKBA, 1 9 6 0 ) , 3 7 9 - 3 8 0 . " Nikolay Andreyev, "Filofey and his Epistle to Ivan Vasil'yevich", Slavonic and East European Review, 38 (December 1959), 1-31, noKa3biBaeT, HTO o6menpinMTaH, Bcjien 3a B. H. ManHHHHbiM — Cmapeif EAuasapoea MOHacmupn Owiotßeü u ezo nocAanux (KneB, 1901) — flaTHpoBKa nocjiaHHH 30-MH roAaMH XVI Bexa onm6oHHa: Ha caMOM jjejie OHO 6 H J I O HanpaBneHO He HßAHY IV, a HßaHy III, OKOJIO 1500 rofla, B CBSBH Kax-pa3 c ero noBTopHMMH KOHHCKauPMMH aepKOBHtix 3eMejn> B HoBropoflCKoä o6jiac™. H. H. MacjieHHHKOBa, "K HCTOPHH co3flaHHfl TeopHH 'MocKBa — TpeTHä P H M ' " , (IIo noBo^y cTaTtH H . E. AanpeeBa "®HJio(J)EA H ero nocJiaHHe K HBaHy BacanbeBHHy"), Tpydu omdejia dpeenepyccKou Aumepamypbi, 18 (1962), 569-581, He coraaniaeTCH c AOBOflaMH H. E. AanpeeBa, H6O, ecJiH OHH BepHti, "STO HeH36e»mo noBJiemio 6w 3a co6oö He TOJÜ.KO nepeocMwcjieHHe MHornx naMHTHHKOB oömecTBeraoö MBICJIH, HO H Bcero npouecca cmianbiBamw aneojiorHH ueHTpajiH30BaHHoro rocyflapcTBa. B rajcoM cuynae TeopHH 'MocKBa — Tperafi PHM' Bonuia 6bi B KOMiineKC HfleojiormecKHX HBjieHHfi KOHiia XV—Haiajia XVI BeKa". CoBepmeHHo BepHO. Tax 3TO H 6HJTO. npeflejiH 3a6ny»«eHHä H . H . MacneHHHKOBOfi onpeaeiunoTCH B apyrofl pa6oTe. H o 3flecb Mbi npHHHMaeM, KOHCIHO, CBOH co6cTBeHHbie BWBOflbi o BpeMera 4>opMyjwpoBKH ®Hjio4>eeM Teopra "MocKBa — Tperaä PHM". ,0 T . I I . «SEFLOTOB, "TPAREFLHH PYCCXOFI CBSTOCTH", Cenmue ffpeeueü Pycu, 1 8 9 - 1 9 2 . HEXOTOPBIE AAHHBIE HA 3TY TEMY MOJKHO HAFITH B ONEM. OSCTOAIEJIBHOIL KHHRE: H . A . KA3AKOBA, Baccuan IJampuKeea u ezo COHUMHUH (MOCKBA-JLEHHHRPAFL, AXAFLEMHA HAYX C C C P , 1 9 6 0 ) . A . B . KAPTAMEB, yM3. cot., 4 8 9 H EN.

72

HHKOJÏAH AHflPEEB

TejIbCTBy, HTO BM30BCT ^BH)K£HH6 H B (J)peCKOBOM HCKyCCTBe H B HKOHHOM

flejie.

HMCHHO STO ô o j i b m o e flBHXceHHe B p e j m r H 0 3 H 0 M wcKyccTBe 3acTaBHT HocHiJuisHCKoe pyKOBOflCTBO MOCKOBCKOH UepKBH nbITaTbCH HaiipaBJMTb pyCCKyiO HCHBOIIHCb

—

TaKOB c M b i c j i MeponpHHTHH M H T p o n o j i H T a M a K a p H H B n o j i o B H H e X V I Beica, K o r ^ a OH,

npn

noMomH

aBTopHTeTa

uepKOBHbix

coöopoB,

c({)opMyjiHpyeT

HecKOJibKo

O6IUHX ' n p a B H j i ' o p y c c K o i í p e j i H r a 0 3 H 0 H XCHBOIIHCH. I I p H STOM oÔHapyacHTca

ero

BeCbMa HeBbICOKHH TeopeTHHCCKHH ypOBeHb B 3THX B o n p o c a x . 3 1 M t o »ce 3a TeMbi jieacaT B n o n e BHHMaHiia HKOHonnciieB B 3 T y s n o x y ? HOBCHLUHH HccjieAOBaTejib n o f l n e p K H B a e T , HTO OÄHOH h3 BaacHbix x a p a K T e p H b i x n e p T p y c c K o ñ xcHBonHCH X V I Beica STO e c T b o ô o r a m e H H e

HK0H0rpa(|>HH: "npHTHH H HH0CKa3aHH9,

H JiereH/ibi, CHMBOJIH H a j i J i e r o p H H , CUCHM B e T x o r o

anoKpH(J)bi

3aBeTa

H

KajiHncHca, HCTOPHH c o T B o p e H H a M n p a H C T p a n i H b i i i cyzi c HeBHAaHHbiM a o

AnoTOÎÎ

n o p b i p a 3 M a x o M oTOÖpaacaiOTCH B HacTeHHWx p o c i r a c a x h HKOHax 3 T o r o BpeMeHH. Bo3HHKaeT MHOKeCTBO ÄHTHHHblX HKOH, HTO HeCOMHeHHO, HaXOflHTCH, B CBH3H C pacuBeTOM HaxoflHT,

arHorpatjîHHecKOH

HTO H HOBbie

noflaioTCH B 3Ty

snoxy

jiHTepaTypbi."

H CTapbie MacrepaMH

cioxceTbi

3TOT

HOBeñiiiHH

pyccKoro

"vpe3BbiHaHHo

Hccjie.no BaTejib

pejiHrno3Horo

pa3BepHyro

HCKyccTBa

H noapoÔHo,

c

6 o j i b i u o H H a r j i a / i H o c T b K ) " . H c c J i e f l O B a T e j i b n o f l n e p K H B a e T , HTO 3 T y " H J i j i i o c T p a T H B HOCTb, CTpeMjieHHe K T i u a T e j i b H o i i H ^ e T a j i b H O H n e p e f l a n e H 3 o 6 p a a c a e M o r o cioxceTa H e o 6 x o f l H M O n o c T a B H T b B CBH3b c n p H c y m e ñ

JiHTepaTypHOMy CTHJIIO TOH

nopw

n o B e c T B O B a T e j i b H o c T b i o . " H c c j i e f l O B a T e j i b r o B o p H T , HTO STO OTpaacaeTca " B c r p e M jieHHH

nepeaaTb

CHMBOJIHKy"

»3HKOM

^CHBOIIHCH

H B OÔHJIbHOM

BBefleHHH

cjioacHyio

GorocjiOBCKo-flomaTHHecKyio

HCaHpOBblX

CUeH,

ÔbITOBblX

3IIH30fl0B

B

CTaHKOByio »cHBonHCb. 3 2 H a . n o n o a n e p K H y T b , HTO STH HaGjHOfleHHJi, .naBaeMbie, KaK o ô o ô m e H H H , Hpe3BbiHaiÎHo HHTepecHbi H, no-BHflHMOMy, c o B e p m e H H o BepHbi. ,H,jra 3 T 0 Í Í CTaTbH BaXCHO, HTO OHH TOHCeCTBeHHbl C HaÔJIIOfleHHaMH H . n . KOHflaKOBa, KaK OTHacTH yace Bbime yKa3aHHbiMH, TaK H CO c j i e / i y i o m H M : " . . . T M I M H3 rpenecKHX CT3HOBHTCH pyCCKHMH,

HK0H0rpaHHeCKaH

CXeMa

OaCHBJiaeTCH

OÔCTaHOBOHHblMH

r p y n n a M H , o)KHBaeT, T e p a e T CBOK) MepTBeHHOCTb, craHOBHTcn HCHBOIO K a p r a H o f t . " 3 3 r p o M a a H b i e TOJiHKH TaKOMy HanpaBjieHHK) HKOHOIIHCH a a j i H , KOHCHHO, H ' e p e r a necKHe' flBHaceHH» MHCJIH, He p a 3 HaHHHaBuiHe O T p n a a T b c a M y i o H f l e i o noKJioHeHHH HKOHaM, a, CTaJio SbiTb, H p e j r a r H 0 3 H y i o acHBoimcb — T e M a HKOHonoHHTaHHH He p a 3 oôcyxcflaeTCH B p y c c K o i í JiHTepaType. 3 4 M o a c e T 6 b m > , HeT n p e y B e j i H H e H H » B y r e e p a c aeHHH:

"HeT

HH

OAHOTO

pyccKoro

6orocjioBCKO-noJieMHHCcKoro

K O T o p o M He 3 a i U H m a j i o c b 6bi HKOHonoHHTaHne CTOJib floporoe p y c c K O M y

TpaKTaTa,

B

HejioBeKy."35

81 H . E. AHApeeB, "MnTponojiirr MaicapaS, KaK Aemenb pejiHTH03Horo HCKyccTBa", Seminarium Kondakovianum, 7 (1935), 227-244. " 3 . C. CMHpHOBa, yxo3. cou., 365. " H . IT. KoaaaKOB, ym3. cou., 9. B CBoeM Tpy^e KoH^aKOB npHBOflHT orpoMHoe KOjnnecTBO MaTepnaJia, HJiniocTpHpyiomero 3TH ero o6o6niaiomne yTBepacfleHiw. 34 A . B. KapTameB, yK03. cou., 451 H CJI. H . E. AaapeeB, "HHOK 3HHOBHÔ OTCHCKHÂ O6 HKOHOnoiHTaHHH H HKOHOnHcaHHH", Seminarium Kondakovianum, 8 (1936), 262-263. " JX- IÎBeTaeB, npomecmancmeo u npomecmanmu e Poccuu do 3noxu npeo6pa3oeamü (MocKBa, 1890), 520. Cp. t a o c e J\. I|BeTaeB, JIumepamypmH 6opb6a C npomecmancmeo M « MOCKOSCKOM

JIHTEPATyPA H HKOHOIIHCb

73

E C T E C T B E H H B I M 0TBeT0M Ha coMHeHH« H J I H OTpuuaHHH B 3 T O M Bonpoce OKa3bIBaJIOCb öojibiuee B H H M A H H E K ' H K O H H O M Y XYAOACECTBY'. Pa3BHrae M C C T H H X noraTaHHÖ CBaTbix yBejiHieHHe KOjinnecTBa MOHacTbipeü, nocTeneHHoe OHapoflHHBaHHe npaBOcjiaBHH, yTepflBiuero C B O H nepBOHanajibHO apHCTOKpaTHqecKO-ropo^cKoü xapaKTep, THnHHHbiii flJiH fl0M0Hr0JibCK0H snoxn pyccKoö uepiCBH, HaKOHeu, orpoMHefiinaa pa6oTa MHTponojiHTa Maicapna H ero 'aicafleMHH' no o K H O

3AHHMATBCA

npo6jieMoö BO3HHKHOBCHHS TaKHX HKOH, KOTopaa, K TOMy xce no-BHflHMOMy, eme OACHFLAET HCKYCCTBOBEFLIECKORO HCCJIEFLOBAHHH. 5 3 H a M N P H X O A H T C A O T M E R A T B TOJibKO

4>aKT HeoÖblHaHHOH CJIOHCHOCTH TaKHX K 0 M H 0 3 H H H H , T.e., NPEFLNOJIARAETC« HeKOTOpblH pacußeT 6orocjioBCKHX HHTepecoB, KaKoe-To ABHACEHHE MMCJIH. OnaTb TAKH, B cHJiy TopacecTBa HocHtJjjiaHCKoro Kpbma uepKBH, noHTH HeT AaHHbix o ' j n o ö o M y a p a x ' , K0T0pbie AOJIACHBI 6HJIH 3AHHMATBCA TaKoii npo6jieMaTHKofi. JIio6onbiTHo,

HTO,

eCJIH O H H X CJIBLUIHO, OÖblMHO C n03HUHH KpHTHKH H O T P H I I A H H H . TaK B 1 5 1 8 H J I H 1 5 1 9 r o a y 6biJio npe/uioaceHO KaKoe-To CKa3mue

o MHCTHKO-flnaaicTHHecKOM cioaceTe

MaKCHMy TpeKy, xoTopoe N O C J I E ^ H H H K P H T H K O B A J I , xoTa H yKJioHnrica O T peuiHTejibHoro

CYXCÄEHHH. 6 4 Ü 3

n p o u e c c a FLBHKA B H C K O B A T O R O B 1 5 5 3 - 5 4 ROAAX B H / I H O ,

HTO,

BHFLHMO, 3TH CK3.3aHHa H TOJIKOBaHHH flajIH yace TJiy60KHe KOpHH H OXBaTHJIH He TOJibKO HOBropOflCKHe H ÜCKOBCKHe paHOHbl, BCerfla HeCKOJlbKO CKJIOHHbie K BOJIbHOflyMCTBy, HO yTBepflHJIHCb H paCUIHpHJlHCb B CaMOM CTOJIbHOM Tpafle. 55 3HaHHT, cjieayeT 3aKjiioHHTb, HTO flojiacHa 6bijia cymecTBOBaTb KaKaa-To nncbMeHHocTb 06 3TOM, KaKHe-TO 4>HJI0C0(j)CKHe TpaKTaTbl HJIH, X0Ta-6bI, KaKHe-TO C(J)OpMyjIHpOBaHHbie TOHKH 3peHH« Ha 3TOT npejiMeT. OTcyTCTBHe HX Bejio flaace K npeAnojioxceHHaM o HHCTO xynoacecTBeHHbix nyTax paöoTbi HKOHHBIX MacTepoB: T.e., KaK-6bi 06 OTCTynjieHHH HX C n y r a HflefiHoro OÖOCHOBaHHH JIHTepaTypOH. K CHaCTbK), HOCH(j)JiaHCKa5I KpHTHKa COXpaHHJia fljia Hac CBOH KpHTHieCKHe apryMCHTbl npOTHB TaKHX COHHHeHHH, 56 a CJlyHafi BbIHBHJI H OFLHO

H3 HpOH3BefleHHii: CKa3anie

06b uKonrb: " C y m e e . B b o 6 p a 3 t flaBHfloßi, M a T o t f i eBaHrejiHCTb rjiarojierb

o H e M t " "KHHra poacTBa I n c y c a X p n c T a C b m a BaHaMH c p t r a x y H 3Baxy " o c a H H a C b m y

flaBHfloBa"

/I,aBHÄOBy" ( M e ,

(1,1),

H OTPOUN

2 1 , 9 ) , H naa.

CT>

cjienua:

" C b m y flaBHAOBb noMHJiyfi Hacb" ( M e , 20,29-30) H xaHaoHbiHH o a m e p n MOJiacb ToxcAe pene ( M e , 1 5 , 2 2 ) , noHeace FLABH^oBo HMH B e j i i e H Hecrao Game BO M3paHJiH, H6O jtlaBHflb npT>Bbifi BI> u a p t x b yro^H E o r y . H a r j i a ß t HMaTb MHTpy, a Ha paMy 0M0(})0pb, noHeace OTb A B O I O KOJiiHy Bomioraca X p n c T o c b — n o Iio,nt OTb upbCKaro, no JleBin ace O T I CBaTHTejibCKaro, B b poflocjioßiH B o r o p o f l H U H n H i i i e r b . O T b BeHua ero ceMb p o r b , noHeace ce^MH BtKOMb TBopeub, a 6yaymeMy B t x y oTeub. /JecHaa ero pyKa n p o c T e p T a Bbicnpb: Bb BTopo3aKoHin nnrnen»: "ilKO B03flBHrHy H a He6o pyKy M O K ) H KJieHyca FLECHHUEIO acrny a3b BO B T K H " (32,40), H Bb CTpauiHbift fleHb peneTb n p a B e f l H b i M b " n p i n f l i T e 6jiarocjioBeHHiH OTiia M o e r o , HacjiiziyHTe yroToBaHHoe BaMb uapcTBo He6ecHoe" ( M e , 25,34) H T O T O pa^H AECHAA CTpaHa cB%Ta

"

H. riOKpOBCKHfi, yK03. cou., 205-208. H. II. KomtaxoB, y>CA3. COH., IV, 268-298. TepMHH MHCTHKO-FLHFLAKTHHECKHE TCMI»I MCNEN KoHflaxoBbiM, KOTOPHÄ TaK H Ha3MBaeT XI maBy CBoero Tpyaa. M C . KanyrHH, 3unoeuu UHOK OmeucKuü u ezo öoeocAoecKO-noAeMunecKue u nepKoeno-ymmeAbHue M

npou3eedeuuH (CaHKT-IIeTepöypr, 1894), 254-255.

" H. E. Annpeeß, "O flene Ätaxa BHCKOBaToro", Seminarium Kondakovianum, 5 (1932), 219-241. " Rudolf M. Mainka CMF, "Zinovij von Oten\ ein Russischer Polemiker und Theologe der Mitte des 16. Jahrhunderts", Orientalia Christiana Analecta, 160 (Roma, 1961).

HHKOJIAH AHflPEEB

80

HcnojiHeHa ; o rnyeñ pyui, e ace aptacHTb Ment, hko " n o o m p i o hko mjitjHÌio Meni. MOH, npÍHMeTT> cyfll. pyKa Moa H BMflaM'b MeCTb BparOMI, H HeHaBHflHmHMb MH Bi.3aaMi)" ( B T 0 p 0 3 a K 0 H i e 32,41), Meni, ace o t b ì t b H a r p i n r a b i n , nace p e n e T i " O T t H f l t T e OTT» MeHe npOKJIHTÍH BT. OrHb BÍ4HMH, yrOTOBaHHblH fliaBOJiy H a H e J I O M l ero" ( M e , 25,45), cìh o t b ì t ì o c T p i e Bcaicaro Mena oôoiOflyocTpa. PyKaBHua acejit3Ha, hko acejiÌ3o HeyMerneHo, rHÌBi> n p o c r a e r o He yMsrieHT> ôy^eTi» rptniHbiJvrb n o cyflHOMT» AMI, a TMa rpiniHHKOMT», bt. Hioace HMyTb n o n r a . A l e MJia/ib Bb 6paHXXT>: "oôJieneTCH, p e n e , bi> 6paHH npaBAbi h B03JioacHTb uuieMi> h cyjxb H e j n m e M t p e H b npiHMeTi. h n o o c T p H T t ratBi. Ha npoTHBHbw" (IlpeM. 5,18), rpflfleTi. noôifly coTBopHTH Ha fliaBOJia. H a Kpecrk »ce c i / i H T t , hko p.a. CHHAerb KpecTOMb bt. a,nb h

TaMo apbacaßy CMepra pa3opHTi> h oTTyzjy norn6iuee oBia Ha paMO B3eMi> h ki> CBoeMy O m y npHHece, cboio öoacecTBeHHyio iuioTb oaecHyio Bora Onta nocaflH npeBbime Bcixb Hanau i. h BjiacTeñ h Bcaicaro hmchh HMeHyeMaro. A eace cepa rptxb. A eace Ha m a ß t AflaMOBe BiflpyaceHb Kpecrb, noHeace Ha KpecTt pacnaTcn XpncTocb h KonieMb npoöofleca Bb peöpa, H3bifle KpoBb h Bofla, na. Kpoßiio 3eMJiio h ott> He h co3^aHHaro AflaMa ocBHTHTb, boaoio ace — p.a. CBHTbiMb J c p e m e m e M b rptxbi ero h Haina ohhcthtì. Ho, o BjiaflbiKO XpncTe Boxee Hamb, Bb AeHb cTpauiHaro h npaBe^Haro TBoero cyfla aecHbw nacra npaBeflHbixb n p H i r a Hact h BinHbixb

TBOHXb ôjiarb

yjiyHHTH

Hacb cnofloÔH hko ÓJiarañ h HejioßtKOJiioöeiib. AMHHb."87

Ecjih B3rjiHHyTb Ha H3o6paaceHHH, co3aaHHbie no 3TOMy TeKCTy, noJiyiaercn h 3^ecb Ta ace CBH3aHHOCTb c tckctom, KaK mm BHflejiH h npeacfle. H 6oJiee Toro 6e3 3HaHHH TaKoro TeKCTa HeB03M0aceH 6biJi 6w TBopnecKHH npouecc HairacaHHH o6pa3a, a c noTepeñ TeKCTOB oh ocTajicH ra/iaTejibHMM, nocKOJibKy ceKyjiaproaipiH co3HaHHH BbiTecHHJia onbiTHoe 3HaHHe sthx chmbojiob, ajuieropHH h öorocjioBCKHX HaMeKOB.

V

BepOHTHO, B CBH3H C 3THMH OÖCTOHTeJIbCTBaMH npOHBJIHeTCH HHOrfla HeKOTOpaH TeHfleHUHH y HCKyCCTBOBeflOB K ynpomeHHK) H, OÖXOfly MHCTHKO-AHflaKTHHeCKHX

TeneHHH hkohoiihch, npn 3tom KaK-6bi M o a e p H H 3 H p y e T C H e e c o f l e p a c a H H e , cboahtch k n p o ô J i e M a M 3ctcthkh. S8 HHOrfla MeTOflbl H3yneHHH CHJIbHO MeHHK)TCH, OTKpblBaH HOBbie r0pH30HTbI.59 Ho, Bee ace ocTaeTcn HacymHOH sa^aneH KOHKpeTH3anHH H3yneHHH KaK-pa3 K0Mn03HUHH HKOH MHCTHKO-AHflaKTHHecKoro HanpaBJieHHH, HeMy n o j i o a c H J i o6ema57 T e K c r CKÜ30HUH: H . E . ÀHflpeeB, "Hhok 3hhobhë Otchckhû o6 HKOHonowraHHH h hkohonHcaHHH", Seminarium Kondakovianum, 8 (1936). Tau nee aHaiiH3 CKasanun. 68

I O . H . ÍÍMHTpHeB, "TeOpHH HCKyCCTBa H B3rjlaflM Ha HCKyCCTBO B nilCBMeHHOCTM flpeBHefi PyCH",

Tpydbi omdejia dpeenepyccKoü Aumepamypbi, 9, 97-118. fl. C. JlHxaneB, Hejioeex. e Aumepamype dpeeneü Pycu (MoCKBa-JIeHHHrpa«, 1958).

JIHTEPATyPA H HKOHOIIHCb

81

60

lomee Hanajio JI. A. MauyjicBHH. KaxeTca, Toraa CTaHeT coBepiiiemío H C H O H HfleHHO-HHTepaTypHaa ocHOBa H 3Toro TeHeHHH pyccicoíi H K O H O I I H C H XVI Beica, B K O T O P O H B C T O J I B H O B O H (J)OpMe H C BMpa3HTejIbHOH C H J I O H OTKpblBäKDTCH HeKOTOpbie HfleñHbie noHCKH KaKoñ TO nacra pyccKoro oömecTBa TOH anoxH. CAMBRIDGE UNIVERSITY

,0

JI. A. MaoyjieBHi, "XpoHOJiorHfl penteiJioB /iMHTpoBCKoro coöopa BO BjiaflHMHpe 3aJieccKOM", P0CCUÜCK020 Hucmumyma Hcmopuu Hcxyccme, I, B. II (üeTporpafl-MocKBa, 1922).

EoKezoÒHUK

H O W A R D I. A R O N S O N

THE GRAMMATICAL CATEGORIES OF THE INDICATIVE IN THE CONTEMPORARY BULGARIAN LITERARY LANGUAGE

I n j ^ H E Bulgarian verbal system, perhaps the most complicated in the Slavic linguistic world, has been the subject of many conflicting analyses. The . JL present study attempts to describe, in structural terms, the grammatical oppositions underlying the various indicative forms of the Bulgarian verb, thereby showing the basic meanings of the various forms and illustrating how they are integrated to form a total subsystem within the greater system of the verb. This will be done by using the examples cited by earlier investigators, and trying to reinterpret them in the light of the system presented here. The first task must be the determination of exactly what forms make up the Bulgarian indicative. Andrejóin (Kategorie, 7)1 lists nine tenses, which can be divided as follows (the forms are the third person singular of the verb cetà 'read'): A. Simple forms 1. present (segasno vreme) ceté 2. aorist (minalo svàrseno vreme) cète 3. imperfect (minalo nesvàrseno vreme) cetése B. Complex forms containing a form of sàm 4. perfect (minalo neopredeleno vreme) cél e 5. pluperfect (minalo predvaritelno vreme) bése cél C. Complex forms, containing a form of ste 6. future (bàdeste vreme) ste ceté 7. future perfect (bàdeste predvaritelno vreme) ste e cél 8. future preterite (bàdeste v minaloto) stése da éeté or ste cetése 9. future preterite perfect (bàdeste predvaritelno v minaloto) stése da e cél Both imperfective and perfective verbs can exist in all nine forms. There are in addition the forms of the so-called 'reported mood' (preizkazno naklonenie). Formally, we find the following types (all marked by the absence of the auxiliaries e, sa in the third person): (a) cél, (b) cetjàl; (c) bil cél; (d) stjàl da ceté, (e) stjàl da e cel. 1

References are to the bibliography on p. 98. All Bulgarian forms are cited in transliteration.

THE GRAMMATICAL CATEGORIES OF THE INDICATIVE

83

AndrejSin (Kategorie, 10; Gramm., 199) posits a whole series of nine often completely homonymous forms corresponding to the nine non-reported 'tenses' above. So, cél corresponds to the aorist ; cetjàl to the present and imperfect (two homonymous series); bil éél corresponds to both the perfect and pluperfect; stjàl da ceté to the future and future preterite (2); and stjàl da e éél to the future perfect and future preterite perfect (2). However, Maslov (Kvoprosu, 312-313) and after him Demina (Pereskaz., 363) have correctly pointed out that recourse to homonymy is not acceptable, and that one must rather procede along formal grounds, recognizing only four reported tense-forms; within the reported series there is a neutralization of oppositions found in the non-reported. There exists also a series of emphatic reported forms (formi za po-silno preizkazvane) marked by the presence of the participle bil. These are bil cel2, bil cetjàl, stjàl bil da éeté, and Stjàl bil da e éél.* We shall next try to show that of the nine 'tenses' traditionally described as indicative, the four forms containing ste or stéSe are in reality marked modals not belonging to the indicative. Golqb (83-93) has shown that the future preterite is basically not an indicative tense, but rather a conditional irrealis, a marked modal form. Janakiev has similarly pointed out that the Bulgarian 'future tense' does not necessarily refer to an action occurring after the speech event, but can refer even to the past; e.g., Dobàr covek beSe pokojnijat ni sàsed, pomagaSe ni — decata ni sie pogleda, kato kupuva edno-drugo ot pazara, i za nas ste kupi ja grozde, ja praskovi... (Janakiev, 426) 'Our late neighbor was a good man ; he used to help us — he will look after our children buying one or another thing at the bazaar, and he will buy either grapes or peaches for us...'. The future is characterized by Janakiev as a categorical presumptive. The future perfect is also clearly modal, as can be seen by comparing Toj e eel knigata 'He has read the book' with Toj Ste e éel knigata 'He must have read the book'. We shall therefore exclude the forms containing Ste, StéSe, and stjàl from our analysis, regarding them as belonging to one of the marked moods of Bulgarian, along with the imperative, the òi/i-potential, the forms of the simple conditional (e.g., dàvvam 'I would give', dàvvah 'I would have given'), and the ifa-conjuctive series and not part of the unmarked indicative. The present study will, as a consequence of the above, be restricted to those indicative 'tenses' traditionally called present, aorist, imperfect, perfect, and pluperfect. The approach followed will be based in the main upon the works of Jakobson. The first grammatical categories to be discussed are common to all the forms of the indicative (as well as to many non-indicative forms). So, all of the forms which concern us are marked for voice. The basic opposition is INTRANSITIVE (+)/TRANSITIVE (—); the former are marked by the fact that they can never take a direct object, the 2 It will be noted that we retain bil éél as two homonymous forms. Bil éél1 is Andrejèin's perfect and pluperfect reported, and bil cél% is the emphatic reported, corresponding, according to Andrejòin, to the aorist, perfect, and pluperfect. We, following Demina (Pereskaz., 368-69), view the two sets as homonymous. This is especially so in our view since the bil in the two forms can be said to have different derivational histories, bibeing the normal reported counterpart of bése, and bil, being the marker of the emphatic reported.

84

HOWARD I. ARONSON

latter by the ability to occur either with or without a direct object, cf. Toj éeíé kniga, 'He is reading a book', with Toj ceté, 'He reads', i.e., 'He can read'. Transitive verbs, as the unmarked members of the opposition, can undergo a transformation into the marked category; this occurs in the formation of passives (e.g., Knigata e detena, 'The book is read') and reflexives (e.g., Knigata se ceté, 'The book is being read'). Also common to all indicative forms is marking for aspect, where the perfective is marked as opposed to the unmarked imperfective. The former obligatory denotes that the action has reached its terminal point, while the latter is unspecified as to the completion or lack of completion of the given action. The indicative forms are all marked for the oppositions of person/number, while the complex forms are in addition marked for gender/number. The third person can be regarded, as in Russian (Jakobson, Shifters, 6), as non-person, since, unlike the first and second persons, it does not obligatorily specify the existance of a grammatical subject; cf. ceté 'he is reading' which can be expanded to Iván ceté 'John is reading' and valí 'it is raining', which is impersonal and not capable of a similar expansion. The first person is the marked member of the personal opposition first/ second person; the first person always indicates a participant in the speech event, the second person (in its so-called 'generalized' use) need not, e.g., Ucenieto ne e cjar da go izpies, ce da stanes ucen 'Teaching is not a medicine that you [ = one] drink to become learned'. Within the gender oppositions the neuter, which functions parallelly to the third person, is unmarked; its presence does not obligatorily mark the presence of a grammatical subject; cf. celo knigata < momee celo knigata 'he (the boy), it is reported, read the book' but valjálo 'it rained (reported)'. The feminine is marked with respect to the masculine, since the latter can refer to both male and female subjects, while the former cannot refer to a male subject. The opposition of number functions in the same manner whether correlated with gender or person; in both instances the marked member of the opposition is the singular, which can refer only to a singular personal subject. The unmarked non-singular (traditionally, the plural) makes no such specification; it can refer to either one or more than one personal subject. The sentence, Cetete mnogo dobre 'you read very well' can refer to one person (polite form), e.g., Drugarju, cetete mnogo dobre or to a group of people, e.g., Drugari, cetete... 'Comrade [resp. comrades], you read very well'. In the sentence Cetes mnogo dobre there is no such double possibility of interpretation.3 The same relationship holds true for the first person, where the plural forms, e.g., cetém 'we read', can assume a singular meaning as in the so-called 'editorial' or 'royal we', whereas cetá 'I read' can never refer to more than one person. The opposition singular/ non-singular found in the Bulgarian verb and personal pronouns (in Russian it also 8

On the expressive level the number-marking of the second person presents a most unusual feature. We may refer to it as a 'shifting marking'. By this is meant that one cannot determine which member of the ti/vie ('thou'/'you') opposition is expressively marked without reference to the social relationships existing between the speaker and the addressee. If the /(-form is normally used in addressing a given person, vie becomes marked in a shift to the plural form; similarly the use of ti is marked if the relationship is a formal one.

THE GRAMMATICAL CATEGORIES OF THE INDICATIVE

85

marks the predicative 'short' adjective forms, as opposed to the long forms) contrasts with the opposition plural/non-plural found in the nominal system. The major part of the present study will be devoted to the discussion of the basic meanings of the so-called tenses of the indicative, and to the question of the relationship existing between these 'tenses' and the reported (and emphatic reported) forms. As has already been mentioned, AndrejCin regarded the forms as basically temporal in meaning. He distinguishes between three basic tenses having a single relationship to the moment of the speech event (present, aorist, and future) and six 'special tenses' which have a double time-orientation, being related both to the moment of the speech event and to 'another moment which is mentioned either directly or indirectly'. (Gramm., 145). AndrejCin's views have been most recently defended by Stankov, especially with regard to the relationship between the present and the imperfect. However, as Jakobson has pointed out (Shifters, 4), tense is best defined as that grammatical category which characterizes the narrated event with reference to the speech event. This results in no more than three tenses: past, present, and future, the last of which, as we have seen, is absent in Bulgarian. The relationship here is between a marked past, which characterizes the narrated event as having occurred before the moment of the speech event, and the unmarked non-past, which does not definitely specify the time of occurrence of the narrated event with respect to the speech event. The non-past, then, in addition to its most common meaning of an action more or less simultaneous with the speech event, can be used to refer to both past actions (as in the 'historical present', see below) and future actions (e.g., 4 Vpetak zaminavam za Sofija 'I'm going to Sofia on Friday'; Maslov, Ocerk, 231). The Bulgarian forms marked as past are the aorist, imperfect, perfect, pluperfect; the reported forms of the type eel and bil celx \ and the emphatic reported of the type bil cel2. Unmarked in addition to the 'present' tense are the reported forms based on the imperfect /-participle, e.g., cetjal and bil cetjal. The systematic relationships existing between three of the forms marked for pastness, the aorist, imperfect, and perfect, have been the subject of much debate in the Bulgarian and Russian linguistic press. AndrejCin took the view that the perfect was opposed as 'resultative' to the 'narrational' (razkazvatelen), 'dynamic' aorist {Kam harak., 61). Maslov viewed the opposition between the aorist, imperfect, and perfect as aspectual: the imperfect is marked as denoting "an action in the process of its course (protekanie) or unlimited repetition as an unfolding or unlimitedly repeated process", a feature for which the aorist is unmarked; the perfect (and plu4

Janakiev (423-426) has viewed 'continuativeness' (kontinuativnost) as one of the distinctive features marking the non-past. But the non-past (IMPERFECTIVE) is used as the replacement for the aorist in the 'historical present', and here it clearly occurs without the meaning of continuativeness; e.g., Kato sljazohme na stancijata vsred grada, ostavjame si bagaza tam, izlizame na ulicata, nasresta tii suma hoteli, izbiram si stai prastame slugata da ni vzeme (antite\... (Al. Konstantinov, in Mutafciev, p. 37) 'After getting off at the station in the center of town we leave our baggage there, go out onto the street, opposite us — a mass of hotels, we select our rooms, send the servant to take our bags...' Here the non-past forms do not have the meaning of continuativeness.

86

HOWARD I. ARONSON

perfect) is opposed to these by being marked for resultativeness (aktuaVnost' posledstvij vyraiennogo dejstvija s toiki zrenija bolee pozdnego vremennogo plana). These two oppositions are viewed as "aspectual in the broad sense" (GlagoVnyj vid, 164166, 273-275). Demina also viewed the opposition between aorist and imperfect as aspectual in this sense, but with the aorist as the marked member, indicating 'wholeness of the action' (celostnosV dejstvija). The perfect and pluperfect are marked as tenses with "duality of the temporal characterization of the action" (Sistema, 29-30, 40-41, 46-48). Dejanova (Slavjanskata) rejects the idea of an aspectual opposition, seeing rather an equipollent temporal opposition between the imperfect and aorist. BorodiS has characterized the opposition between the aorist and imperfect on the one hand and the complex past tenses including the reported forms on the other according to two criteria: (a) the former denote actions witnessed by the speaker and the latter need not; and (b) the former denote actions viewed as concrete, real facts of reality which occurred at a definite moment in the past while the latter need not. The opposition between the aorist and the perfect used to describe a witnessed action is viewed as purely stylistic (Ob obScix, pp. 69, 71-78). In a later work [K voprosu), BorodiS ascribes to the perfect two distinct basic meanings: (a) modal, describing the relation of the speaker to a given act; (b) non-modal (purely perfect meaning); in Bulgarian the modal meaning is the more dominant and productive. With regard to the opposition between the imperfect and aorist, Maslov's aspectual solution seems the most plausible. We shall thus define the marked imperfect as denoting a past action which is obligatorily viewed as having occupied more than one moment in time. We shall call this feature 'durativeness'. This plurality of moments can underly both a purely durative and an iterative meaning; compare: Periodicnijat pecat se razvivase pod znaka na usilijata na fasistkata vlast da ograni£i i zabrani izlizaneto na demokraticnite i revoljucionnite vestnici i spisanija ('The periodical press was developing under the sign of the efforts of the fascist authorities to limit and forbid the publication of democratic and revolutionary newspapers and magazines'); and, Nastroenieto mu rezko se menese po njakolko pati v den, ('His mood would sharply change several times a day') (Maslov, GlagoVnyj vid, 256). The aorist, being unmarked, can occur with the same meanings, but only contextually conditioned. This aspectual opposition is independent of, but subordinated to, the overriding perfective/imperfective opposition. The combination of the two sets of aspectual opposition results in four forms, e.g., the more usual procete = + perfective, — durative;procitaSe = — perfective, + durative; and the rarerprocetese — + perfective, + durative; and procita = — perfective, — durative. The meanings of these forms must be defined by the combination of the meanings of the underlying markings.5 Note that the combination of two identical markings (i.e., + + or ) is more marked than the presence of two different markings. •

For a detailed discussion of the four aspectual combinations, see Maslov, GlagoVnyj, 252-275.

THE GRAMMATICAL CATEGORIES OF THE INDICATIVE

87

The aorist and imperfect are generally regarded as witnessed forms, i.e., as forms which underline the speaker's actual witnessing of the narrated event. This would mean that the narrator of the speech event was a witness of the narrated event. This is, in fact, the most common meaning separating the aorist and imperfect from all other forms of the verb, but the constatation of the narrator's having witnessed the event narrated by him is not always present in uses of these two tenses.6 BorodiC (Ob obscix, 70) mentions the use of the aorist and imperfect to denote "unwitnessed events contemporary to the speaker", giving the example Maza i ubiha prez turskata vojna ('They killed her husband during the Turkish war'), where it is contextually clear that the hero of the story did not witness the killing. In Kosev's History of Bulgaria, for example, the reported tenses are used regularly for events occurring before 1918. Events occurring thereafter are generally given in the non-reported forms. But, it is clear that the author did not witness each of the events described. In the following example from Jovkov (cited in Demina, Pereskaz., 343) we find an especially clear example of the use of the aorist for an unwitnessed action: — Edna zvezdapadna, — kaza Marin. — Umrja njakoj (' "A star fell", said Marin. "Somebody died"'). The first aorist, padna, indicates that the falling of the star was witnessed by Marin. But the second, umrja, clearly refers to an unwitnessed action. In this environment (for reasons to be discussed below), one would expect instead of the aorist either the perfect, umrjal e njakoj ('someone has died') or the so-called future perfect, Ste e umrjal njakoj ('someone must have died'). The nuance conveyed by this example is one of absolute certainty: the fact that someone died is just as certain to the speaker as if he had himself witnessed the death. The speaker vouches for the truth of the narrated event. We shall call this meaning, which is basic to both the aorist and imperfect, CONFIRMATIVE. The confirmative can be considered an example of what Jakobson has named the grammatical category of STATUS, which he defines as characterizing the narrated event without involving the participants in that narrated event and without reference to the speech event (Shifters, 4).7 Although we regard the vouching for the truth of the statement narrated as the basic meaning of the confirmative, it should be emphasized again that this meaning is most commonly made manifest as underlining the fact that the narrated event was witnessed by the speaker. (This distinction between the basic meaning [Jakobson: Grundbedeutung] and the main, most common meaning [Jakobson: Hauptbedeutung] is not restricted to the confirmative. Similarly, the basic meaning of the form cete is nonpast, but it most frequently denotes an action contemporary to the speech event; • Excluded from discussion here is the use of the aorist and imperfect for what Borodic (Ob obscix, 70) calls the 'proposed witnessing of an action' as in an author's narration of fictional (and therefore non-witnessed) actions. ' Jakobson gives the same definition for status as for aspect. The two categories are distinguished in that aspect is a quantifier, while status is a qualifier. What we have called the confirmative in Bulgarian is reminiscent of the assertive use of do (does, did) in English, except that the very strong implication of witnessing found in Bulgarian is absent in English sentences of the type 'He did read the book.''

88

HOWARD I. ARONSON

the future Ste ceti we have regarded as a marked modal form; nonetheless it most often occurs denoting an action occurring after the speech event.) The question of the basic meaning of the Bulgarian perfect has produced much discussion and many conflicting solutions. Some of the resulting views have already been given. The most common solution to the problem has been to regard the perfect as a resultative form, denoting the present result of a past action. This is, in fact, the general meaning assigned to the term 'perfect' in most languages having such a category. As we shall see, this is indeed a common meaning of the Bulgarian perfect, but it cannot be regarded as the basic meaning. There exists, marginally, to be sure, in contemporary Bulgarian a complex form of transitive verbs composed of imam 'have' plus the past passive participle, which agrees in gender/number with the direct object.8 Such forms, as exemplified by Imam cetena kniga 'I have read the book' have a clearly resultative meaning, and while the so-called perfect, e.g., Cel sam knigata 'I have read the book' can have a resultative meaning, it quite often occurs without implying any present result of a past action. In addition to the 'perfect' or 'resultative' meaning, a series of other meanings has been ascribed to the perfect tense. Among these is the so-called predpolagaemoe dejstvie, 'inferred action', which has been elevated by Demina to the status of a new mood (Pereskaz., 319-324). Another use is in the sequence of tenses of the 'historical present', where the perfect is used to denote simple anteriority to the action expressed by the non-past; its relation to the non-past is the same as that of the pluperfect to the aorist. Most investigators have pointed out that the perfect is used to narrate both witnessed and unwitnessed or reported actions. It is especially common in negative and interrogative sentences, and in reported speech after such introductory words as rece, kaza 'he said', mt'sli 'he thought', etc. Statistically the perfect is most common in the first and second persons, where it can be used almost interchangeably with the aorist, and is least common in the third person. The question remains, what feature (or features) are common to all of these meanings, and serve further to distinguish all occurrences of the perfect from all other forms of the verb. In our view, the perfect is to be defined as a marked past form in opposition with the non-past, but as unmarked for the opposition confirmative/non-confirmative, thus opposed to the aorist and imperfect; and as an unmarked member of the opposition reported/nonreported.9 The perfect is therefore the least marked of all Bulgarian past tenses. Confirmation of this definition can be found by analyzing the tenses used in the Bulgarian 'historical present'. The use of the non-past to denote an action which occurred prior to the moment of the speech event does not, as we have seen, contradict the basic meaning of the so-called 'present tense', since the opposition past/nonpast is privative, rather than equipollent. The use of the historical present as a stylistic device "for giving the narration a greater liveliness and relief" (MutafSiev, 8

For a discussion of forms of the type imam cetena see VI. Georgiev. Note a similar definition of Janakiev (432): "The 'perfect' in all Slavic languages contains only one relevant semon — 'preteritiveness'." Georgiev (47-48) comes to a similar solution. 9

THE GRAMMATICAL CATEGORIES OF THE INDICATIVE

89

26) is common to most European languages, but in Bulgarian it has an additional function: since Bulgarian past actions are generally presented as either constated, witnessed (aorist, imperfect) or unwitnessed AND reported, the use of the non-past, which is unmarked for these oppositions, eliminates the need of specifying the source of the information narrated. The non-past, which is unmarked for confirmation, reportedness, durativeness, and tense can thus be used to substitute for the aorist, imperfect, and past reported: compare the following; past witnessed: Patnikat sleze ot vlaka i barzo premina prez perona. Na plostada pred garata go Sakase malka leka kola, kojato bese sprjala pod goljamoto darvo na sjanka ('The traveller descended from the train and quickly crossed the platform. On the square in front of the station there was waiting for him a small auto which had stopped in the shade under the big tree'). The reported forms would be: sljazal and preminal replacing the aorists sleze and premina; cakala replacing the imperfect cakase, and bila sprjala replacing the pluperfect bese sprjala. With the historical present we would have the non-pasts sliza, preminava, and caka, and the perfect e sprjala (Mutafdiev, 29). These last forms in each instance give less information than those in the preceding two sets. In fact, the only meaning connected with e sprjala is that the action occurred before the speech event; the fact that it is anterior to the moment of orientation of the narrated event is contextual. 10 The wide range of meanings carried by the perfect is quite compatible with its minimal markedness. So, for example, what has been called predpolagaemoe dejstvie turns out to be a contextually determined expression of the unmarkedness of the perfect. In the example, Varvim iz patja i az si mislja: Vasilka ne e glupava da stoi v takova vreme, zatvorila e hana i se e pribrala v grada (AndrejSin, Gramm., 179). 'We're going down the road and I think to myself: V. isn't fool enough to stay here in such weather, she has closed [ = must have closed] the inn and gone to town', Vasilka's closing of the inn and going to town are not witnessed (or confirmed) actions, therefore the aorist cannot be used; nor did anyone report to the speaker that she had closed the inn and gone to town, i.e., the reported past cannot be used. The meaning 'inferred action' can be viewed as a logical consequence of the marking non-confirmed, non-reported. When the speaker concludes that a certain event has taken place without having viewed that event and without being told about it, 10

Stankov has attempted to show that the imperfect is the past equivalent of the present tense. In so doing he specifically disregards the 'transferred use' (prenosna upotreba) of these forms in the historical present construction. But it is precisely in this construction that the lack of parallelism between the non-past and the imperfect is most clear, and that the almost complete parallelism between the non-past and the perfect is most evident. It is also interesting to note that the tenses used in the historical present construction are grammatically the least marked, but on the expressive or stylistic level, the construction is more marked than the corresponding aorist-imperfect-pluperfect construction. Such an inverse correlation between grammatical and expressive marking can be found elsewhere. As we have shown, the non-singular forms of verbs and pronouns are grammatically less marked than the singular, but the use of the 'royal' or 'editorial' WE is stylistically more marked than the use of the singular 'I'. There is a similar relationship between the grammatical and expressive markings of the Bulgarian genders (see H. I. Aronson, "The Gender System of the Bulgarian Noun", IJSLP, VIII, 1964, p. 93).

90

HOWARD I. ARONSON

he will use the perfect. Another example : Liceto na ToSka beSe bledno, posàrnalo, ocite ì bjaha vlazni i pozacerveni. Plakala e, Ivan razbra. 'Toska's face was pale, dejected, her eyes were moist and red. She has cried [ = must have cried], Ivan understood'. That this use of the perfect is not markedly modal can be seen by contrasting it with the use of the future perfect, which is a marked modal form: Ivancol vikna toj ... Otgovor njamase. —Zaspal e. Ivancol vikna po-silno. Nikoj ne se obadi. Ah, ste e bjagàl, zavalijata (Andrejcin, Gramm., 179). '"Ivanio!" he called. ... There was no answer. "He's fallen asleep [perfect]". "Ivanòo!" he called louder. No one answered. "Oh, he must have run away [future perfect], the poor devil." ' The perfect, zaspal e is contextually modal, while ste e bjagàl is a marked modal form. The resultative meaning of the perfect can also be explained in terms of its unmarked nature. The aorist and imperfect by their very marked nature, focus the listener's attention on the act itself, underlining the speaker's confirmation (or witnessing) of that act. The perfect on the other hand does not necessarily direct the listener's attention to the act. Thus, contextually, it may be that not the act itself, but its results are the focus of attention. We see this in the following example from Pod igoto (in Andrejòin, Kàm harak., 62) : Az znam polozitelno, ce toj e izdal vsicko na turcite... 'I know positively that he has betrayed everything to the Turks'. Here the focus is not on the actual moment of the betrayal to the Turks, but rather on its result. The same could be true for the opposition between Kupil sam si novo paltò 'I have bought myself a new coat', and Kupih si novo paltò 'I bought myself a new coat'. But here a new complication enters. In certain contexts, the use of a confirmative form (especially the aorist) for an unquestionably witnessed and confirmed action gives an unwanted emphasis to the fact of confirmation or witnessing. This is true in the first, and to a lesser extent, in the second persons.11 Here the unmarked perfect is often used to replace the more marked aorist. As Andrejòin (Kàm harak., 61) has shown, Veera pristignah (aorist) and Veera sàm pristignal (perfect) 'I arrived / have arrived yesterday', can be used parallelly to describe the same event. Andrejòin views the first as more 'dynamic', the second as resultative. But the first can be interpreted as approximately 'I DID arrive yesterday', the second as simply 'I arrived yesterday'. The preference for the perfect rather than the confirmative tenses in interrogative and negative sentences (Borodiò, K voprosu, 28) can be explained similarly. In such instances the questioner phrases his inquiry in the least marked terms; so, to a question such as Cel li e tozi romani 'Has he read this novel?', one can receive the reply Toj go cète 'he read it' (witnessed, aorist), Cel go 'he read it' (reported), or Toj go e eel 'He has read it' (unmarked, perfect). The use of the perfect in questions as opposed to the reported past can be seen 11

As will be seen below, within the past, the main opposition in the first and second persons is confirmative/non-confirmative, while in the third person the opposition can be viewed as ternary: confirmative/reported/unmarked. Again, as is usual, the more marked categories (first and second person) have fewer oppositions than the unmarked category (third person).

THE GRAMMATICAL CATEGORIES OF THE INDICATIVE

91

in the following example: Toj poiskal kafe i Sarandovica otiSla v hana. Togava uciteljat zakljucva vratata sled neja i nasocva revolvera srestu dâSterjata. Dali prozorecât e bil otvoren, ili pâk v uplahata i v otcajanieto si tja go e otvorila, ne se znae, no tja vece se prehvârljala prez nego, kogato uciteljat strelja... 'He asked for some coffee and Sarandovica went into the inn. Then the teacher locks the door after her and aims the revolver at her daughter. Whether the window was open or whether in her fright and despair she opened it is not known, but she had already crossed over to it when the teacher fires' (From Jovkov, in Mutafòiev, 128). It is clear that the narrator had no report on the state of the window (ne se znae, 'it isn't known'). Nor was the state of the window witnessed. Thus neither the confirmative nor the reported forms could have been used. The only remaining possibility is the perfect. If the perfect is used to replace the witnessed, confirmative forms in the first and second persons to avoid an overinsistance on the constating of the action, so it is also used to replace the reported past after such words as réce, kâza, misli, etc. In these constructions the reported nature of the narrated event is made explicit lexically, and so the unmarked past can be used: Za nego kazvat, ëe na Dermenka e utrepal komandira na lovnata rota, porucik Apostolov 'They say of him that at Dermenka he killed the commander of the ... company, Lieutenant Apostolov' (Borodiô, K voprosu, 20). To summarize : the unmarked nature of the perfect can be demonstrated by showing how the specific meanings of all other marked past forms are neutralized in it: 12 MARKED

AORIST: Toj cète knigata. He read the book.

PERFECT

Toj e eel knigata.

ls Forms of the type éetjàl sâm, éetjâl e, i.e., composed of the imperfect /-participle plus the present tense of sâm 'be' in all persons are not mentioned by Andrejiin, Kateg., and in the Gramm., 242, he specifically states that "this participle is used only in the corresponding reported tenses". Maslov 0Ocerk, 223, note 2) points out that this construction, found "in very rare instances ... cannot be considered correct; the norm of the literary language permits the imperfect participle only in the composition of the forms of the reported mood". A similar prohibition is found in Andrejiin, Kostov, and Nikolov, Bâlgarski ezik (S., 1962), 278. Nonetheless, such forms are found even in the writings of Andrejiin: Polucili sa razprostranenie ... i prilagatelni imena s nastavkata -im, -em (kojato njakoga e sluzela za obrazuvane na segasni stradatelni pricastija...)... (Bâlgarski ezik, XIII 1963, 348) 'Adjectives with the suffix -im, -em (which at one time used to serve for the formation of the present passive participle) also have acquired distribution.' Demina (Pereskaz., 321) gives other examples, regarding them as members of a special modal paradigm, that of predpolagaemoe dejstvie. As we have shown, this is but one contextual meaning of the perfect, and though it is easily seen in Demina's example ' Vidja li sviróicitel Toj trjabva da e imal deca i na tjah gi e nosel'. 'Did you see the whistles [in the pocket of a dead man]? He must have had children and was carrying them to them'; it is equally clear that in the example from Andrejiin the use of -im, -em to form the present passive participle is not a deduction (umozakljucenie) but an already known fact. We cannot, therefore, agree with Demina, but rather regard the cetjól e form as opposed to cél e in the same way that cetése is opposed to cète, i.e., as an opposition durative/non-durative. But since Andrejiin specifically excludes the construction from the literary language, we shall disregard forms such as cetjól e in our analysis of the total system of the Bulgarian indicative.

92

HOWARD I. ARONSON

PLUPERFECT:

Toj bese ve£e procel knigata, kogato Marija vleze. He had already read the book when Mary entered.

Toj veöe e procel knigata, kogato Marija vliza.

REPORTED PAST:

Toj eel knigata. He read the book.

Kazvat, öe toj e eel knigata.

FUTURE PERFECT:

Izgleida, Se toj ste e eel knigata. It seems that he must have read the book.

Izglezda, öe toj e eel knigata.

[RESULTATIVE, 'PERFECT']:

[Toj ima cetena kniga. [He has read the book.]

Toj e eel knigata.]

The pluperfect like the aorist, imperfect and perfect is marked as past with respect to the non-past. Like the perfect, it is unmarked for confirmativeness, thus opposing it to the aorist and imperfect. It is opposed to the reported tenses by being unmarked for reportedness. The pluperfect is distinguished from the perfect by what Jakobson {Shifters, 4) has called TAXIS, that grammatical category which characterizes the narrated event in relation to another narrated event, and without reference to the speech event. In Bulgarian, the relevant opposition is between anterior/non-anterior. The pluperfect is marked as anterior, i.e., as obligatorily denoting a past event which preceded another past event representing the moment of orientation of the narration. The pluperfect is connected most closely with the perfect, from which it is distinguished only by the latter's being unmarked for anteriority. The perfect (as the aorist and imperfect; BorodiC, Za funkeiite, 432-434) can occur with the meaning of anteriority only contextually; compare the two tenses in the following examples: Togava baba Ana mu razkaza kakvo bese se sluSilo 'Then granny Ana told him what had happened'; and Sega razvalnuvano zapoena da razkaza, kakvo e vidjal; 'Now he began agitatedly to relate what he saw (had seen)' (BorodiC, Ob obScix, pp. 75-76). The tenses described heretofore all shared in common the feature of being unmarked members of the opposition reported/non-reported. It now remains for us to discuss those forms marked as reported, eel, eetjal, bil celY\ bil cel2 and bil eetjal. According to some investigators, e.g., Maslov, Andrej5in, Teodorov-Balan, these reported forms belong to a marked modal category and do not fall into the indicative. Demina, (Pereskaz., 356-361) rejects this view, preferring to see these reported forms as belonging to a 'secondary indicative mood' (vtoricnoe iz"javiteVnoe naklonenie). We shall follow Jakobson (Shifters, 4) in regarding these forms as marked for the EVIDENTIAL category, defined as characterizing the relation between the narrated event, the narrated speech event, and the speech event. In Bulgarian, the speaker under-

THE GRAMMATICAL CATEGORIES OF THE INDICATIVE

93

lines, by the use of the reported forms, that he has learned of the narrated event through someone else's words (oral or written); what Jakobson has called the quotative evidential. The forms eel, cetjal, and eel bilt need not inherently cast any doubt on the veracity of the reported narrated event; the use of the emphatic reported forms, on the other hand, does. These latter forms, bil cel2 and bil cetjal, are the marked members of a status opposition which we shall call dubitative/non-dubitative. This opposition can be viewed as approximating a 'mirror image' of the confirmative/nonconfirmative opposition within the non-reported forms.13 Before proceding to the discussion of the interrelationships existing between the various reported forms it is necessary to question the existence of homonymous forms of the perfect and the past reported in the first and second persons. These tenses, as traditionally given, are: Singular (masc.) Past Reported Perfect 1 2 3

cél sâm cél si cél e

cél sâm cél si cél

Plural Perfect céli sme céli ste céli sa

Past Reported céli sme céli ste céli

If the basic meaning of the perfect is viewed as resultative, or as a tense with a double orientation in time, etc., then there is little to prevent one speaking of a clear opposition in grammatical meaning concomitant with formal homonymy. We would have two series opposed not directly to each other, but along different axes. But as we have tried to show, the perfect and the past reported are directly opposed to each other, the latter being marked as reported, the former — unmarked. It is not reasonable to assume that eel sam can on the one hand obligatorily denote a reported narrated event and on the other hand can be unspecified for reportedness. There is a neutralization of the opposition reported/non-reported in the first and second persons,14 and the only opposition is confirmed/non-confirmed. Only in the third person do both oppositions appear. Thus the least marked person is the most marked for oppositions of status. These relationships are summarized by the following chart: 18

There is a temptation to unite the oppositions confirmative/non-confirmative and dubitative/ non-dubitative into one opposition, marked status/unmarked status, the meaning of which would be dependent upon whether the given form is reported (marked status = dubitative) or non-reported (marked status = confirmative). 14 Under certain circumstances, mainly stylistic, the opposition between reported and non-reported is neutralized also in the third person, when non-reported events are narrated in the perfect without the auxiliary e or sa. For examples see Demina, Pereskaz., 324-325, especially note 40. To this category belongs the so-called 'admirative' (see Weigand; Andrejcin, Kateg., 68; Demina, Pereskaz., 325-328). The problem of when the auxilliary can be omitted in non-reported forms requires further investigation, but it is clear that the opposition between the perfect and past reported is not very strong.

HOWARD I. ARONSON

94

Non-confirmative Confirmative

Non-reported

Reported

1

cetox

~

cetjah

eel

2

cete

~

cetese

eel si

3

cete

~

ceteSe

¿el e

sam eel

W i t h i n the non-emphatic reported forms w e c a n distinguish first between the past reported, eel, and the durative reported, cetjal.

These are o p p o s e d a l o n g t w o axes:

eel, w h i c h corresponds t o the non-reported aorist, is the marked member o f the o p p o s i t i o n past/non-past, and the unmarked member o f the o p p o s i t i o n durative/nondurative; cetjal,

corresponding both t o the imperfect and non-past is durative and

non-past ( D e m i n a , Pereskaz.,

364-365). 1 5 Whether the durative reported corresponds

to the imperfect or the non-past is contextually determined. T h e pluperfect reported, bil ¿¿¡i is marked as past with respect t o cetjal and as anterior with respect t o 16

cel.19

There are examples of ietjdl sam, ¿etjdl used in an admirative, non-reported function; see Weigand; Andrejcin, Gramm., 239; Maslov, Ocerk, 251-253; K voprosu, 312; and Demina, Pereskaz., 150-152. From the examples cited in these sources one may conclude that the forms in question usually have a 'present' rather than imperfect meaning. Unambiguous examples are given only with the verbs imam 'have' and sam 'be' (and one example with znaca 'mean': Eto kakvo znacelo da ddrzis vlastta v racete si. Dori obstinskata vlast 'That's what holding power in your hands means. Even municipal power'). As with the omission of the auxiliary in the perfect, the admirative use of the imperfect /-participle requires further study. 16 The bil eel, form is generally regarded as corresponding not only to the non-reported pluperfect, but also to the non-reported perfect. However, the majority of the examples given by various investigators strike us as ambiguous. Many of the forms interpreted as a reported 'perfect' can also be viewed as the emphatic reported past: (1) '"Kazide! Kazi! Na kogo s&m izjal nesto, a?" "Nikomu nisto ne si bil izjal\ A dvata napoleona ot moja hak?"' (Jovkov, in Andrejcin, Kateg., 59). '"Come on, tell me! Who have I cheated out of anything?" "You haven't cheated anyone out of anything! And what about the two Napoleons from my salary?"' (2) ' "A pak, deto kazvat, 6e vSliicata se bila ovalcila v nivite, ne vjarvam. Tam e tja, tarn — v gorata."' (Jovkov, in Demina, Pereskaz., 367). '"And then, when they say that the wolf whelped in the fields, I don't believe it. She's there, there — in the forest."' (3) '"... streljali sme nie za zenki, a za svoboda ne jost streljali..."' Raznese se drezgav smjah. Goleman zagleda tamnite lica. Otde izlizaha tija knizni dumi u tjah? Te za svoboda oste ne bili streljali...' (StraSimirov, in Demina, Pereskaz., 367) '"We have shot for women, but for liberty we haven't shot yet..." A hoarse laughter resounded. Where had these bookish words of theirs come from? They had not yet shot for liberty...' Other examples can be interpreted as pluperfect: (1) 'Ne bili prosti kleiki, hodza im bil eel, magija im bilo praveno' (Jovkov, in Andrejcin, Gramm., 212). 'They weren't ordinary sticks, the hodja had cast a spell on them, they were enchanted'. (2) '"... de se izgubi? Da ne si se boril s meckata, sis stravnicata? Pak se bila javila v Lipov-djal."' (Jovkov, in Demina). '"Where did you disappear to? You didn't fight the bear, the meat-eater? It had appeared again in Lipov-djal."' (3) '"Ti znaeS li? Dunav svalja savsem maskata na Ognjanova: izobrazava go cjal! Toj bil izbjagnal iz Dijarbekir i go tirsjat pod listo...'" (Vazov, in Demina). ' " D o you know? The [news-

THE GRAMMATICAL CATEGORIES OF THE INDICATIVE

95

The emphatic reported tenses, marked for the opposition dubitative/non-dubitative, are opposed along the same two axes as eel and (etjal: bil cel2 is past non-durative, and bil cetjal is durative non-past. It should be noted that the cetjal forms, when used to denote the 'present' may often have a strong nuance of disbelief, doubt. Examples include: "Az dori ne ja poznavam." "Ne ja poznaval! Cjal svjat ja poznava, toj ne ja poznaval" (Maslov, K voprosu, 314). ' " I don't even know her." "You don't know her! The whole world knows her, and he says he doesn't know her."' " N j a m a ja, Stoilko," varna se izplasena Stoilica... "Njamalo ja. Na orehi li ste igraemT opuli se djado Stoilko (Demina, Pereskaz., p. 364). '"She isn't here, Stoilko", his frightened wife replied. "Not here! Will we be playing games now?" stared, wide-eyed, old Stoilko'. (Romanski, Makedonski pregled, II, 3, 1926, 143, in his review of Weigand's article on the Bulgarian admirative reports, and at the same time casts doubt on the veracity of, the latter's contention that English is not well known in Bulgaria): Tova povtorenie na izkazvanite tam misli i na nemski toj objasnjava i s tova, ce anglijskijat ezik v Balgarija ne bil tvarde poznat. 'He explains this repetition in German of the thoughts expressed there [Slavonic Revue] by the fact that English is not well known in Bulgaria.' Connected with the above is the fact that the use of reported forms to mark past events is more obligatory than their use to mark 'present' events (Maslov, Ocerk, 249). The explanation for this, and for the occurrence of the cetjal form in contexts clearly marking disbelief or doubt, is, we believe, the following: we can speak of degrees of constating marked by the Bulgarian verb. The aorist and imperfect usually combine both the speaker's affirmation of the veracity of the narrated event and the fact of his having witnessed it. This is the most marked degree. When the speaker does not wish to underline his affirmation and/or has not witnessed the event the perfect or eel- or cetjal- reported forms are used. Note that formally there is little difference between the perfect and the ¿¿/-reported, and that there are examples of the complete neutralization of the opposition. Similarly, we have seen examples of cetjal sam, cetjal e used as a durative counterpart to eel sam, eel e, introducing an opposition durative/non-durative within the 'perfect'. Thus, the celand cetjal reported forms, in addition to marking reportedness, are also unmarked for the confirmativeness of the aorist and imperfect. But, when we oppose reported forms to the non-past and perfect, i.e., to forms which are already unmarked for the confirmativeness, the lack of confirmation can become positively marked, i.e., dubitative. For this reason the ce//a/-reported form with present-tense meaning is quite often used dubitatively. We can summarize by saying that when a marked reported form is opposed to a marked confirmative form, the result is a neutralization of paper] Danube has completely unmasked Ognjanov: it gives a full picture of him. He has [had] escaped from Diyarbekir and they're searching for him in every nook and cranny'. Even if these examples are accepted as being the reported equivalent of the perfect there is reason to believe that they belong in the emphatic reported series. See below.

+

« .S

.

TO

+ o >

Q

+

- em . ei

m Oh

CC C O U

i CS 11 3

"Q 3? a

o

Z

D

+

x> 3 Q

(A

- cd

»•ÇÙ

CL,

—

«S a Ph I -d•

o >

+

z

u

3 Q I g. z

o

•o u -t:.

o ft o s

bíl óetjál

•o >o 35 >o

ôetjàl

i> XJ

ôéte

u

bése ôél

The hierarchical structuring of the various grammatical categories marking the Bulgarian 'tenses' can be illustrated by Table I. The same relationships can be illustrated in the form of a distinctive feature chart:

«

35 o S

+ + —

+ + + + + + + + + + + + + + + +

—

—

—

+ + + —

—

—

—

—

—

—

—

—

—

UNIVERSITY OF CHICAGO

98

HOWARD I. ARONSON BIBLIOGRAPHY

Andrejein, Gramm. = L. Andrejein, Grammatika bolgarskogo jazyka (M., 1949). Andrejein, Käm harak. = L. Andrejein, "Käm harakteristika na perfekta v bälgarskija ezik", Ezikovedski izsledvanija v cesi na akademik Stefan Mladenov (S., 1957), 57-64. Andrejein, Kateg. = L. Andrejezin, Kategorie znaezeniowe koniugaeji bulgarskiej (Kraków, 1938). L. Andrejein, "Säätnost i upotreba na minalo neopredeleno vreme v bälgarski ezik", Ezik i literatura, XL (1956), 140-143. Borodie, K voprosu — V. V. Borodie, "K voprosu o znaòenii perfekta v bolgarskom jazyke", Slavjanskaja filologija, vyp. IV (M., 1963), 3-35. Borodie, Ob obseix = V. V. Borodie, "Ob obseix prineipax upotreblenija ölennyx form i glagol'nyx vremen v bolgarskom jazyke", Izvestija na Instituía za bälgarski ezik, VII (1961), 49-80. Borodie, Za funkeiite = V. V. Borodiö, "Za funkeiite na minalo predvaritelno vreme v sävremennija bälgarski ezik", Bälgarski ezik, VII (1957), 430-451. Dejanova, Slavjanskata = M. Dejanova, "Slavjanskata glagolna vidova opozieija imperfekt ~ aorist", Ezik i literatura, XVIII (1963), 106-110. Dejanova, Za otnosenieto = M. Dejanova, "Za otnosenieto mezdu aorista i perfekta v starobälgarski ezik", Bälgarski ezik, XV (1965), 97-109. Demina, Pereskaz. = E. I. Demina, "Pereskazyvatel'nye formy v sovremennom bolgarskom literaturnom jazyke", in: Voprosy grammatiki bolgarskogo literaturnogo jazyka (M., 1959), 313-378. Demina, Sistema = E. I. Demina, "Sistema prosedsix vremen v novobolgarskix tekstax XVIIXVIII vekov", Ucenye zapiski Instituía slavjanovedenija, XIX (1960), 1-48. Georgiev = VI. Georgiev, "Väznikvane na novi slozni formi säs spomogatelen glagol imam", Izvestija na Instituía za bälgarski ezik, V (1957), 31-59. Golqb = Z. Goi^b, Condifionalis íypu balkanskiego w jftzykach poludniowoslowianskich (Kraków, 1964). Jakobson, Shiflers = R. Jakobson, Shifters, Verbal Categories, and the Russian Verb (Cambridge, Mass., 1957). Jakobson, Zur Struktur = R. Jakobson, "Zur Struktur des russischen Verbums", in Charisteria Gvilelmo Malhesio oblata (Prague, 1932), 74-84. Janakiev = M. Janakiev, "Za gramemite, narieani v bälgarskata gramatika 'segaäno vreme' i 'bädeste vreme'", Izvestija na Instiiuta za bälgarski ezik, VIII (1962), 419-432. Kosev = A. Kosev, Hr. Hristov, D. Angelov, Kratka istorija na Bälgarija (S., 1962). Maslov, GlagoVnyj vid = Ju. S. Maslov, "Glagol'nyj vid v sovremennom bolgarskom literaturnom jazyke (znaeenie i upotreblenie)", in Voprosy grammatiki bolgarskogo liieraiurnogo jazyka (M„ 1959), 157-312. Maslov, Käm lipologija = Ju. S. Maslov, "Käm tipologija na glagolnija vid", Ezik i literatura, XVIII (1963), 3-10. Maslov, K voprosu = Ju. S. Maslov, "K voprosu o sisteme form pereskazyvatel'nogo naklonenija", Sbornik v cest na A. Teodorov-Balan (S., 1956), 311-318. Maslov, Ocerk = Ju. S. Maslov, Oierk bolgarskoj grammatiki (M., 1956). Mutafciev = R. Mutafeiev, Segasno isioricesko vreme v sävremennija bälgarski ezik (S., 1964). Stankov = V. Stankov, "Imperfektät v sävremennija bälgarski knizoven ezik — prezens v minaloto", Bälgarski ezik, XV (1965), 193-204. Weigand = G. Weigand, "Der Admirativ im Bulgarischen", Balkan-Archiv, I (1925), 150-152.

ROBERT AUSTERLITZ

TWO GILYAK SONG-TEXTS

l

2

FAR AWAY BEYOND THE RIVER

DRIVEN TOWARD THE OPEN SEA

WHERE IT MERGES WITH THE SEA

ABOVE THE NETS IN THE RIVER HU

WHITE-HAIRED GOFKOF LIKE A GULL

A FLOATING LEAF WAS SWALLOWED

IN THE NORTHERN BREEZE TREMBLING WILLOW LEAVES

BY THE RISING TIDE AS IF

^ H E S E are very free translations of the two Gilyak song-texts which I propose to analyze in some detail below. The fact that these translations only remotely . 1 reflect both the content and the construction of the originals — as will become evident in what follows — supports the contention that poetry is untranslatable, and that the obstacles in the way of translating it are compounded as the disparity between the cultures which the two languages represent increases. A number of versions of these and other song-texts were collected in Japan in 1954, 1956-57, and 1962. The versions given here date from 1954. The informant is Mrs. T. Nakamura, a speaker of the Southeast Sahalin dialect (see Austerlitz, 1956). I will try to show here: how these texts are organized internally (metrically) and what the devices of such internal organization are; how a prose version of a given song-text differs from the sung version; how the immediate 'message' or content of a song-text veils but at the same time suggests one or more deeper and culturally more intricately coded messages; and how the two items chosen for this presentation are genetically related. I will conclude with a few comments on the musical component of these two song-texts (henceforth to be called simply songs). An explication of this type will require a meticulous analysis of the material presented, and more: it will require that whatever is meaningfully related to or tangential to the songs be incorporated into the analysis. Obviously, space will not permit such an exegesis. I will nevertheless try to approximate it by giving detailed glosses for every form in the texts, by providing a prose paraphrase for one of the songs, by presenting the metrical analysis in considerable detail, and by introducing biographical

100

ROBERT AUSTERLITZ

information about personalities who are or were connected with the content or conception of the songs. If the highly elusive nature of poetic (i.e., sung) texts, as they are collected in the field, is not to defeat the collector-analyst's attempts at a full understanding, means must be found to bridge the void between poesy and truth. In my own work, I soon learned to train the informant to make prose paraphrases of her songs. The results were only partially satisfactory. Either the informant retained most of the poetic devices in her prose version, thus helping only minimally to clarify the poetic text (this is the case in the prose version of Song 1, below) or her prose version tended to assume the form of a restatement-at-a-distance which only scantily resembled the original poetic (sung) version. Prose paraphrases are therefore useful but, by their very nature, confirm the autonomy of poetry. The two items presented here belong to the class alx=du+lu or alx-dumt of the genre lu*nt 'song' (lu- 'to sing').1 This class of songs consists of concretely founded idylls, i.e., pictures of an event, person, or scene, generally beheld in a slightly marvelling or surprised perspective, an attitude revealed in our texts by such devices as 'looked', 'saw', 'having the appearance of or'?'; cf. also -ru in the prose version of Song 1. The subjects of some of the other alx*du=nt in my collection are a boat and its two loverpassengers, a traveller in a valley who is caught between the sky and a brook, the movements of a skier, the gauche behaviour of a young swain, an aggressive woman's boasts, looking at a lily-bulb, and similar topics. As a class, the alx^du-nt are opposed to lullabies and to occasional songs (such as a song about a bird in a cage). In addition to its most accessible layer of meaning, each alx=dumt has a deeper significance in that it is either tied to a specific event within the memory of the performer and audience or to a stereotype in the society. This second, inner layer is difficult to elicit for various reasons: the text is by nature vague or the informant's recollection is dim or the informant may simply be unwilling to reveal details. Thus, Gofkof's personality (Song 1) remains vague unless we are willing to speculate on the basis of nontextual evidence. The inner significance of Song 2 will be discussed later. Some (but not all) alx=du=nt also have a third layer or dimension: among some of the highly lyrical specimens there is sometimes a veiled erotic or simply a libidinous component. However, the genre expressely reserved for eros is that of the vacant (see Austerlitz, 1961a) which is quite distinct from the alx-du=nt. According to the informant, the texts of both of the songs which follow were created by her brother around 1920. I interpret this to mean that the text of Song 1 is probably her brother's version of an older song and that the text of Song 2 is her brother's adaptation of the framework of Song 1 to events which he witnessed. My assumptions seem to be borne out by internal evidence. The melody is the same for both songs (and, in fact, is found to occur with other alx-dwnt). It is traditional. 1

For interesting details about and samples of Gilyak poetry see PiJsudski 1912 and 1913.

TWO GILYAK SONG-TEXTS

101

GRAMMATICAL APPARATUS

The transcription used to render the texts is the same as that which I have used in my other publications on Gilyak. The texts are here somewhat ruthlessly segmented into grammatical units. The symbols employed, as well as the classes of grammatical units which the use of these symbols entails, are explained in what follows. + separates the consituents of compounds, gujqi 'willow' + plarjq 'leaf' =gujrji+ flarjq 'willow leaf' {p : / in plarjq : +flarjq is explained below). + + separates those constituents of compounds which already contain sub-constituents: hu+i++amx 'the mouth of the river Hu', where i 'river' and amx 'mouth, estuary'. s indicates that the element which follows is derivational or participial: =i]kcausative, -la- concretizing, -r, =r and =t verbal-noun forming, -rj (and =m, -n, =h) prenominalizers and nominalizers of verbal roots or stems. - indicates that the element which follows is inflectional (in the case of verbs) or an enclitic (in the case of nouns): -nt finite verb, -rja '...-ing and', -i- present-future; -at 'as concerns', -lo '?', -ru 'behold', -do 'toward'. ~ follows the pronominal elements j- and i- (third person singular) and n- (first person singular) which, when they precede nouns, indicate possession and when they precede verbs indicate the object (e.g., i-nty- 'see it' : nry- 'see'). & separates items which co-occur in a reduplicated complex. See Song 1, lines 3-6. The form n—fton&Siof is here exceptionally classed as reduplicative because it corresponds to a non-poetic form co/=f tyr'-rja 3 2 tlewla*r) geñ hup3=r¡kf=f i-nty-ya 3 pard&pard-qk^-f i-nty-ya 4 s ñ~fioñ&ñofLq?+rur3T} 5 gujtji+flarjqa mry&mry=r3 6 lamf+la&l'íqia wyri&wyri PROSE:

1 hu+i++amx daz=yk-t i -nty-rja, 2 tlewla-rj gex hup-n-do e/r-ya, 3 para¶-rjk-t i-nty-rja, 4 3, ñ-fioñ&ñof(q)-ru. 5 gujyi+flarjq mry(w) &mry-yk-t i-nty-ya, 6 lamz+la(lq) wyri&wyri-r defi-nt-ru. * Takahasi reasons that the Gilyak river-name hu+i is from the South-Sahalin Japanese toponyms ho i or ho e (the former is a township and an affluent of the Horonai River some 40 km. north of the coast) because Gilyak place names are rare in South-Sahalin. He further traces the Japanese toponyms to a Gilyak form hui+i which he translates as 'salmon river'. This, I think, is an error: There is a member of the salmon family called hoj 'Hucho Perryi BREVOORT' (Japanese ito, SahalinAinu cirái, Orok nalduma — a commoner Gilyak name for it is guplant) which cannot have anything to do with hu or hu+i. The matter of this toponym is important because if the river is 40 km. inland it is unlikely to have an estuary (as the texts of Songs 1 and 2 indicate) and still less likely to have the features which the text of Song 2 requires. On the other hand, we know certain 'historical facts' about the background to the text of Song 2. How are we to reconcile these with the possibility of a completely fictitious river hut Takahasi also reports that when the Gilyak hear Song 1 performed they are seized by fits of laughter but he cannot explain why. His comment is equally incomprehensible to my informant and remains incomprehensible to me unless the laughter was a reaction to a specific performance.

TWO GILYAK SONG-TEXTS

103

Glosses Line 1 A«+i"the river Hu', i 'river', amx 'mouth; mouth of a river, estuary', raz-'aim at and hit the mark, here visually'; here daz- because the verb governs the preceding noun with final JC; -ijk- normally causative factitive, here poetic; -t non-finite (converb). i-nty- 'see something', tyr- 'look at', -rja '(is/was) ...ing and'. 2 tlewla- 'is white', -y pre-nominal (noun modifier), gex 'gull (Larus crassirostris Viellet; the black-tailed gull)', hup- 'be seated', -n (= nominalizer, -do 'to, toward', here '(thought the gull) TO BE (sitting)'. e%r- 'think'. 3 para¶- 'while someone's arms (or a bird's wings) are extended laterallyhorizontally he sways in such a way that while the right arm goes up, the left goes down, etc. (00 00 00 00)', cf. para- 'arms (wings) are horizontally extended (with no movement, 00)', cf. also far&far=u- 'fuss about nervously', par- 'take one's clothes off, undress [intrans.]'. 4 a for surprise, n - 'my'. The anthroponym as it appears in the text is a poetic distortion of Gof&Gof, said to be an Evenki-Tungus (ywynki) name; it is here preceded by the possessive element n— (which is rare with proper nouns) and followed by the poetic suffix q and by an expletive; the regularly corresponding prose name would be *Gon&Hof. durrj 'shape, form, appearance', here with initial r because it is compounded with the name which precedes it; -ru (in the prose version) 'but; after all it turned out to be'. Note that durrj (-\-rurrj) in the sung version and -ru in the prose version are similar in form, though in content and function they have little to do with each other. — The protagonist's name may harbour a suggestion of various words which can be associated with the color of his hair (white, grey), e.g., Gafc- '(hair) turns grey', Gonu-, Gal&haly- (related meanings), Gana- 'is white'. 5 gujiji 'willow (Salix sachalinensis Fr. Schm.)'. playq 'leaf, mry&mry- (also mry&mry-, mry&myry-) 'thin or delicate objects are trembling or fluttering; small quick repeated movements', an affective verb related to the kinaesthetic form [mrasiSmraa] used to imitate such movements, perhaps connected with mry'walk along (in) a stream'. 6 lams 'north', la 'wind', reduplicated *la&l, with the poetic suffix la&ljj, (and expletives), defi- 'the wind is blowing', wyri&wyri- 'a slight breeze is gently blowing', cf. the kinaesthetic forms wyri&wyri, wyri&wyri which are used to render the sensation of a slight breeze, -ru see line 4. Line-by-Line 1 2 3 4 5 6

Translation

Peering and looking at the mouth of the river Hu (I thought) I saw a white gull sitting [there] [slowly] moving its wings up and down [repeatedly]; but it is the shape of my [friend] Gofkof [instead]. The willow leaves are trembling, the northern wind is gently blowing.

104

ROBERT AUSTERLITZ

Metrical Analysis The musical framework provides the basic metrical skeleton which can serve as a point of departure. Each metrical unit which is here written as a line of verse corresponds to a musical phrase which consists of four bars. In schematic form, this phrase is

4 - + - K

5

1 — r

Both the melodic and the rhythmical structure of this musical phrase easily permit us to segment it into two halves. Carrying the resulting segments over to the text, we emerge with what we will call the nominal and the verbal halves of the line, viz:

1 2 3 4

NOMINAL

VERBAL

hu+i++amx daz=ijk-t tlewla=ij gex para¶=rjk=t 3 n~fion&fiofiq,+

tyr-rja hup-rjk-t i—nty-ya i-nty-ya rurtj

5 gujyi+flayq 6 lamz-\-la&llq1

mry&mry=r wyri&wyri

(The above version is identical with that of the sung text except for the expletives which are here omitted; a has been rewritten as a according to the dictates of the verbal paradigm.) The form +rury in the VERBAL column in line 4 is obviously nominal and will have to be ruled out as an exception (though we can note the existence of a verb dury- 'has the shape of...'). The verbal segment in line 2 consists of two (rather than one) words: the tempo of the recording suggests that the informant forgot to sing hup-rjk't before the end of the nominal segment and compensated by adjusting it into the verbal segment. Similarly, in other versions of this song, a single verbal form occupies the verbal slot; e.g., in a version recorded in 1962 the line is sung as tlew=la=y get? hup'*n-do | e/r'-ya. Let us therefore move hup-yk-t arbitrarily into the nominal column and proceed to concentrate on those grammatical devices which seem to play a role in the metrical structure. The most obvious of these is parallelism. In the schematic representation which follows, N stands for a noun, a nominal root, or a deverbal nominal form and V for a verbal root. Suffixes and the two words which do not participate in parallel units are given in full. Since it seems to have a specific

TWO GILYAK SONG-TEXTS

105

function, the element q is also given: it is part of the root in line 5 and a specifically poetic suffix in lines 4 and 6. 1 N + N + + N Y=yk*t 2 N N N-rjk-t \&V*yk--t 3 4 9 N&N t #j

V-rja y-r/a V-ya rurrj

5 N+N? 6 N+N&N^j

V&V=r V&V(0)

Devices which lend coherence to the text seem to abound. The first three verbal segments end in -rja while the first three nominal segments end in causatives (,-rjk-) and in the suffix -t. The incidence of nouns in these three segments decreases line by line: NNN, NN, 0 . The verbal root in lines 2 and 3 is the same (*nty-). Reduplication is absent at first, then appears in line 4 (in the nominal segment) and in the verbal segments of lines 5 and 6. Is it an accident that rj occurs in the end of the first four verbal segments and q at the end of the last three nominal segments? The last two lines have very similar structures which set them off, as an idyllic couplet, against the slightly more concrete quatrain which precedes them. The couplet, it will be noted, also contains two kinaesthetic verbs which are parallel to each other. It is difficult to ascertain from the recording whether the last verbal form ends in the finite suffix -nt (as sentences would in prose) or in -r (roughly '-ing') as does line 5, or in 0 , as I have indicated. If the last of these is the case, as I believe, we are faced with a strikingly suggestive instance of a naked reduplicated verbal root in utterance-final position. This could never occur in prose and is particularly powerful because the verbal root itself is kinaesthetic. Let us now concentrate exclusively on reduplication and noun-compounding. The text presents the following instances of each: 1 N+N++N 2 3 4 N&N 5 N+N 6 N+N&N

V&V V&V V&V

There is thus an increasing heaping of both compounds and reduplicated forms as the song approaches its end. If we now lump compounding and reduplication with the other devices which participate in parallelism (the poetic element q, the suffix-cluster -rjk-t, and the suffix -tja) we emerge with the following profile which reveals the frequency, the locus, and the nature of the devices involved:

106

ROBERT AUSTERLITZ

N+N Line 1 2 3 4

x

5 6

x x

N&N

q V&V

*i)k--t

x

-rja V&V x x x

x x x

x x x

x x

x x

There are concentrations of devices which run through lines 1-2-3, through 4-5-6, and through 5-6 alone. Line 4, which contains the unexpected element of the narrative, tends to cohere less with lines 1-2-3 than with 5-6. This is also the only line whose verbal half (which is, in fact, not occupied by a verbal form) dangles, i.e., does not cohere with any other segment in the same position. The deployment of the devices, finally, is such that the song can be said to resolve itself (visually) into four quarters (upper/lower, left/right), of which the upper/left is the least marked and the upper/right and lower/left the most marked. I will conclude this metrical analysis with a syllable count. UNE

NOMINAL

VERBAL

1 2 3 4 5 6

9

3 3 3 2 3 4

8

6 4 4 5

Let us remember that the real, temporal length (as is borne out by the music) of each nominal segment and of each verbal segment is the same. It is difficult to say whether the gradual decrease in the syllable count of the first four lines' nominal segments (9-8-6-4) and the comparatively static syllable structure of the last three lines' nominal segments (4-4-5) is accidental or not. Other forms of cohesion between lines l-2-3-(4) and (4)-5-6 have been noted above. Contentual Analysis The quatrain speaks for itself: the observer mistakes his friend Gofkof, at a great distance, for a gull. Contributing factors which, at the same time, provide immediate and simple but powerful images are the color of Gofkof's hair (see also the discussion in the glossary) and the movement of his arms. I have been unable to extract any ad hominem information about Gofkof; at one time the informant hinted that he may have been out of his mind; the fact that he was not a Gilyak may also be of some relevance

TWO GILYAK SONG-TEXTS

107

here (the South-Sahalin Gilyak look upon the Tungus tribes who live in their proximity with a certain amount of condescension). There seems to be a slow increase in the speed of movement in the first four lines: peering (linel), sitting (line 2), movement of extended arms and a moving human shape (3-4), which is abruptly cut after line 4. In line 5 the human-animal frame of reference is replaced by the much more delicate movement of leaves (line 5) and the wind (6). The final couplet is simply a terse idyll. The scene is of special importance. It lends the song gravity: fishing in the estuary is seasonal and essential for survival (for both men and dogs) during the winter. This is an exciting season also for the opportunities which it provides for social contacts and for gossip among populations which are relatively isolated during the long winter months. It is the season when memories are collected for the winter. Further comments, not immediately relevant here, will be found below, under the heading SYNTHESIS. SONG 2 SUNG:

1 2 3 4 5

i

hu+i++am f tajfr-ata dam-la-y ke j—o^-n larku-m platjqa safRo-n gerq3n+cafia wat-i-n j-apu=r3 xec-i-n

tajfr-lo tajfr-lo tajfr-lo dur'y* dur'tf

Glosses Line 1 For the first complex, see Song 1, line 1. tajyr 'a depression within a river, near its mouth'. Such a place is known to be rich in fish and is therefore favoured for fishing (by net). It is close enough to the sea to be affected by the tide; hence it is turbulent within, but calm on the surface. For lack of a better equivalent in English, we will gloss it as 'the hollows.' Both -at and -lo are enclitics, respectively 'as concerns' and '?'. 2 dam- = dam-la- 'there are many', used with count nouns (matyo- with mass nouns); dam*la=y "many a person" = 'many people('s)'. ke 'fishing net'; dam=la=y ke 'many fishermen's nets', j-ox- 'a crowd collects', an animate verb which, in the song, occurs with an inanimate subject (ke). 3 lafku- 'it floats', platjq 'leaf, leaves'; larku-m platjq 'a floating leaf, floating leaves'. sarfro- 'stop working, take a rest' (cf. carfxa- 'is not working [for a few days]'). sarho-n tajyr 'the hollows where [floating leaves] come to rest'. 4 gerqtj 'the sea', cax 'water', wat- 'two persons meet (by chance, on the road); one (liquid) is mixed or mixes with another' (cf. v/at-f 'place where two [intestines] are

108

ROBERT AUSTERLITZ

joined); -/- present-future marker, durrj 'shape', here also 'giving the appearance of [the place or situation where incoming water from the sea merges with the river's water]'. 5 j-apu- 'something comes from another direction and forcibly collides with something else', xec- 'the tide rises, i.e., sea water moves inland in the estuary'. j—apu-r modifies xec-by and both, together, modify durtj. Line-by-Line Translation 1 2 3 4 5

The hollows of the mouth of the river Hu [are] hollows where many [fishermen]'s nets crowd together, hollows where a floating leaf rests [and sinks], like sea-water meeting [the river's], like [sea-water] colliding, rising-toward-the-land. Metrical Analysis

The Gilyak text, as given above, is already cleft into a left and a right half. The prose version provided by the informant was so close to the sung version that it was deemed unnecessary to reproduce it here. The right column, in this case, is clearly nominal. The form tajyr 'hollows' (abbreviated Nx) dominates lines 1-2-3 and plays a special role within line 1 (figura etymologica); durr/ (abbreviated N2) gives coherence to the final couplet. Otherwise, the opening tercet and the closing couplet are not especially marked. There is no reduplication in this song (whereas Song 1 abounded in it). Schematically, including a syllable-count: LEFT

1 2 3 4 5

N+N++Nx V=n N N \--n N N N+N V=n V*r V=w

RIGHT

L

R

Nx-lo Nx-lo Nx-lo Na N2

8 5 7 6 5

3 3 3 3 3

The suffixal parallelism (V-n) in lines 2-3-4-5 is striking; the nominal parallelism, N + N in lines 1 and 4 and N N in lines 2 and 3 (where, in fact, the first N is a deverbal prenominal nominal form), is equally striking. Still, the inner organization of the first and the last lines departs somewhat from that of the interior lines, thus giving lines 1 and 5 their due individuality as boundary markers. In general, however, this song is less complex and its architectonics less elaborate and imaginative than Song 1, perhaps because Song 1, as we shall see, is subject to fewer referential (contentual) constraints than Song 2. Song 1 is also older and perhaps contains an element of multiple authorship.

TWO GILYAK SONG-TEXTS

109

Contentual Analysis The opening line is idyllic and sets the tone for the entire song, especially with the figure 'as concerns the hollows' ~ 'the hollows?' (tajyr-at ~ tajyr-lo). The tajyr and its dual nature provide the key to this song and, indeed, to more than this song. It is described (grammatically modified) in lines 2 and 3: the place where nets are packed during the fishing season, as if they had congregated (in the lower regions of the tajyr), and the place where a leaf is placidly floating and comes to a stop (on the surface). The leaf sinks as a result of the entering high tide which checks the river's (and the leaf's) progress toward the sea (line 4). What was descriptive in line 4 (the pressure of the sea toward the land) becomes almost violent in line 5: the sea collides with the river and wins the struggle: high tide. The mounting pressure in the left column of lines 4 and 5 is mitigated by durtj in the right column: all is only appearance. SYNTHESIS

As might be expected, there is a woman behind the story. According to the informant, Song 2 came into being around 1920 when the South-Sahalin Gilyak C. was ill. It was autumn. A visitor from the North, K., appeared at that time and soon an attachment developed between K. and C.'s wife, L. Not unaware of it, C. once surprises K. and L. at the river, at night, and belabours K. with an oar. K. and L. flee (northward). On the following day, a search party consisting mainly of K.'srelatives(includingthe poet) covers both sides of the river but fails to find K. or L. It was learned later that K. had agreed to meet his brother B. near the hollows and that it was B. who guided K. and L. to safety. In view of this information, Song 2 acquires a new dimension of meaning: Line 1 is the scene, line 2 echoes the turbulence (and passion?) within L., line 3 is an allusion to the intruder K. and to his disappearance, and lines 5-6 portray the tension. In other versions of this song there are two lines between our lines 1 and 2: la lb la lb

gucla foru-n mi+xevri-n

tajyr-lo tajyr-lo

hollows smooth on the outside [= surface] hollows whirling (kevri-) within (mi)

These again portray L. who was known to be quiet and dignified but unpredictable and capricious. (Much of this exegesis was volunteered by the informant.) The relation between Songs 1 and 2 now resembles the relation of two haiku (such as are quoted in Springer [footnote 8], and Austerlitz, 1961b) which share a theme but evoke different responses: The personality of Gofkof (Song 1) remains vague; the scene in which we meet him is depicted in considerable detail. The same locale reappears in Song 2, but it is now secondary to an event to which the poet was witness, an event which probably had considerable shock-value in the society in which it

110

ROBERT AUSTERLITZ

occurred. The interdependence of Songs 1 and 2 is subtle; the devices employed in both songs are refined and powerful. On the formal, metrical level, however, Song 2 suffers as much as it gains on the referential level: the devices are fewer and the coherence is weaker. This is why the assumption of single authorship might be lent credence. MUSIC

The two melodies follow. They are, of course, essentially the same for both songs. In transcribing them from my tape I was torn, as is to be expected, between two poles: on the one hand, it would have been preferable to record every minute ornament and every variation in rhythm with complete loyalty to the tape; on the other, the temptation to integrate the renditions on this particular tape with my memory of all the other renditions of these songs which I have heard was too strong. The result which appears here therefore is a mere approximation of the rhythmical and pitch schemes, in which some of the ornaments and the most obvious deviations from the basic rhythm are nevertheless noted. The metronomic value for one quarter note in Song 1 is ca. 200; in Song 2 it is ca. 164. On the tape (which may not reflect the original situation with complete accuracy), both songs end on C below the treble clef; i.e., I have transposed the melodies upward by a major sixth. Breath pauses are indicated by means of raised commas at the places where they occur. The favourite ornament is a slow trill which starts on the pitch indicated and generally trills the pitch one whole tone below it. I have indicated these trills by means of tr above the staff. When, instead of an expected trill, I heard another configuration, I marked it as, for example, in the last measure of line 5 in Song 1. Line X (between lines 5 and 6) in Song 1 is obviously a mistake. I have placed parentheses around it. In two other instances I followed the informant's opinion rather than the auditory impression gained from the tape: (i) dazaijkdta in line 1 of Song 1 sounds more like dazaijkuta (grammatically: daz-ykm-t, where =u- can occur and would indicate transitivity) and (ii) in Song 2, line 3, where I hear larkumflatjq (with / ) , which is an impossible form. These are my only two emendations. In both songs, the first two vowels (u and i) are sung on one note, with u short and i long. I have transcribed the sequence as ui. All of the bars within a song tend to be of the same length but there is considerable variation. Thus, the first measure in line 4 of Song 1, which consists of a single tone, is not quite four beats long and (probably because of this very reason) upsets the length of the measure which follows it. Since this particular deviation was particularly strong and quite obvious, I wrote the measure in 3/4. Otherwise I have made no changes of this kind. It is quite obvious that each unit which we have called a line in the metrical analysis

111

TWO GILYAK SONG-TEXTS

Song 1

-9

1

i

n i n /—*—*-

hu-i

*

a-ma-ya

M

*

'

fm ta

da-zag-ka-

ty-

ra-

1)3 tr

-4=Htlew- lag

" S i s z t fia

ge-

hu-pag

ka-ta

1

[-

±=± pa-rag-

pa-re—7

^ T

-V-

—

®

t

*

1

ty-

ka-ta 1

3

1

*

A —

— -tfc —

-

—

in-ty- ga -Z ga

©

£

}

±=t -4

flag-

§ (J J v

N

r

O-

guj- gi

m 4 qa

mry-

lam-

za

mry-

tr

i §

deft-lag i = d b

* 3

J la-

I la-

>

i = t *

qa

wy-

±=±

-fr-

ritr

corresponds to a musical phrase. In most instances, the end of the musical phrase is marked by a breath pause: the boundary between phrases is therefore marked by a da capo (of the phrase) as well as by the (quite clearly audible) breath pause. Furthermore, most of the phrases also end in a trilled note. (The trill in the lapse in line X, Song 1, is particularly interesting in this regard.) The first measure of each phrase is

112

ROBERT AUSTERLITZ

Song 2 r

i — R — h— 1 — f

A

i\ A

^ hy-i

.

ma-

79

dam-

I

lag

taj-

ya-

1

laf-kum

i

jo-

gef-qañ

—• P~r

Xan

taj- f c

yai-

lo tr

taj-

ysr-

lo

tr

I

1

ca-

Da'

±=L pu-

ta

I plag-qa

1

ra-

±=±

ke

M— 1i

1

wV

— 1 a-

± = t

ja-

mi

ra

ra-

flon

±=t tin

taj-

•pr-

lo

tr tf—— du-

r|3

i=± du-

os

stationary (i.e., the only pitch, in my notation, is b). The second measure contains a descent and a return to the incipit. The third measure is generally stationary and, as such, an echo of the first measure. Each phrase concludes with a pitch which, up to that point, had only occurred in passing in the second measure (a in my notation). Song 1, the text of which — as we have seen — is much more ornate and 'artistic' than that of Song 2, is also more complex musically: line 5 diverges considerably from the above characterization (thus increasing suspense) and line 6 has a very ornate (and rhythmically complex) ending — perhaps connected with the blowing of the wind? Note also that on occasion the first note in a measure is anticipated by a short (generally eighth) note in the measure which immediately precedes it, e.g., the end of line 3 and the beginning of line 4 in Song 1, lines 1-2 in Song 2, the last two notes in Song 2. In general, words tend not to stretch over one measure in the first half of a phrase (or line) but often do stretch over one measure in the second half of a phrase, e.g., tynrjd, intyijd, ruratj (in Song 1). This is nothing more than a corollary of the syllable counts which were given above in which it was seen that the 'left column' contains a higher number of syllables than the 'right' (pp. 106 and 108). It is at this point that the metrical analysis and the musical analysis converge: what were 'columns' in the

113

TWO GILYAK SONG-TEXTS

metrical analysis are halves of musical phrases. The first half of these musical phrases (i.e., lines of verse) is generally more packed in content (and hence number of syllables), while the second half is the relief from the tension built up in thefirsthalf: contentually, metrically, and musically. COLUMBIA UNIVERSITY

REFERENCES Austerlitz, Robert 1956 "Gilyak Nursery Words", Word, XII, 260-279. 1958 Ob-Ugric Metrics. Helsinki ( = Folklore Fellows Communications, 174). 1959 "Semantic Components of Pronominal Systems: Gilyak", Word, XV, 102-109. 1961a "The Identification of Folkloristic Genres", in Poetics, 505-510. Warszawa (Instytut Badari Literackich) and The Hague (Mouton). 1961b "Parallelismus", ibidem, 339-443. 1966

"Questions of Text and Music in Mansi Songs", Current Musicology, Spring, 37-57.

Jakobson, Roman 1957 "Notes on Gilyak", in Studies Presented to Y. R. Chao on His Sixty fifth Birthday ( = Academia Sinica, Bulletin of the Institute of History and Philology), Part I, 255-281. Panfilov, V. Z, 1962/1965 Grammatika nivhskogo azyka. 2 vols. Moskva-Leningrad (Akademia Nauk S.S.S.R. Institut azykoznania). Pilsudski, Bronislaw 1912 "Poezya gilakow", Lud, XVII, 95-123. 1913 "The Gilyaks and Their Songs", Folk-Lore, XXIV, 477-490. Springer, George P. 1956 "Language and Music: Parallels and Divergencies", in For Roman Jakobson. on the Occasion of His Sixtieth Birthday. The Hague (Mouton), 504-213.

Essays

Takahasi, Moritaka 1942 Karahuto giriyaku go [The Gilyak Language of Sahalin]. Osaka (Asahi sinbunsya).

ROBERT AUTY

THE GOSPEL A N D PSALTER OF CHERSON: SYRIAC OR RUSSIAN?

I

N 1944 Roman Jakobson 1 took up the conjecture put forward in 1935 by André

Vaillant,2 to the effect that the famous poycbCKaia nucMena in which the Gospel and Psalter were written that St Cyril found on his visit to Cherson3 were in fact not Russian (still less Gothic or Scandinavian) but Syriac. Vaillant had noted that two MSS of the VC alter CoypH into PSCH in ch. XVI, and Jakobson brought to light a similar confusion in a 13th-14th-century MS of Chrabr's treatise, where an original asur- had become rus-. Moreover, he sketched in the historical background of the incident and showed the credibility of the Syriac interpretation in the light of the circumstances of the mid-9th century. In an authoritative article D. Gerhardt 4 then summarized a century's research on this puzzling passage, accepting the Syriac explanation as the most satisfactory of all those previously advanced and showing how this solution arose as a natural and acceptable conclusion of the long drawn out scholarly debate on the subject. There, it might have been thought, the matter would rest. The Syriac explanation proposed by Vaillant, reinforced by Jakobson and Gerhardt, offered an acceptable solution to a historical problem. Like all such solutions, it was a probability, not a certainty, a conjecture based on historical premisses and sound philological method, but one which depended less on imaginative or even fanciful hypotheses than any solution previously advanced. Nevertheless, the Syriac explanation has not met with general acceptance. F. Liewehr, writing in 1951,5 had not even acknowledged its existence, and Gerhardt's rejoinder was necessitated by the realization that the articles of Vaillant and Jakobson had been overlooked. In 1956 Emil Georgiev published his compendious survey of the life and literary activities of Saints Cyril and Methodius.6 In it he dealt on

1 R. Jakobson, "Saint Constantin et la langue syriaque", Université de Bruxelles, Annuaire de l'Institut de Philologie et d'Histoire Orientales et Slaves, VII: 1939-1944 (New York, 1944), 181-6. a A. Vaillant, "Les 'lettres russes' de la Vie de Constantin", Revue des Études Slaves, XV (1935), 73-7. 8 Vita Constantini (hereinafter = VC), VIII, 15. * D. Gerhardt, "Goten, Slaven oder Syrer im alten Cherson? Ein Nachtrag", Beiträge zur Namenforschung, IV (1953), 78-88. 5 F. Liewehr, "Wie Konstantin-Kyrill mit Wulfilas Bibelübersetzung bekannt wurde", ibid., III (1951/2), 287-90.

'

E. Georgiev, Kupuji u Memoduü

ocHoeono/iootcnuifu Ha cAaenncKume Aumepamypu

(Sofia, 1956).

115

THE GOSPEL AND PSALTER OF CHERSON

pp. 60-62 with the question of the books in 'Russian letters'. He rejected the Gothic hypothesis and accepted the view that the Gospel and Psalter received by St Cyril were indeed written in proto-Russian. It is remarkable that he did not discuss the Syriac theory, even though he mentioned Gerhardt's article in a footnote. 7 It is equally strange that T. Lehr-Splawinski, in his edition with Polish translation of the Lives of Saints Cyril and Methodius, 8 while accepting the Gothic hypothesis and acknowledging the existence of the Russian interpretation, does not so much as mention the possibility of the books in question being written in Syriac. The Russian interpretation was argued at length in 1963 by V. A. Istrin. 9 His arguments fall into two classes, textual and historical. "All 23 copies" of VC, Istrin points out, 10 use the phrase 'Russian letters' in the relevant passage. Although there are in fact 32 extant copies of the text 11 Istrin's main point is not invalidated : none of the known MSS shows the form coypbcicbiMH or any form with s-r- rather than r-s-. The variant forms pocbCKbi and poyxaMu are, justifiably enough, disregarded by Istrin. But against the numerical preponderance of the witnesses we may set the immense distance of time that separates them from the date at which VC was composed. Not a single preserved MS of this work was written down earlier than 1400; 12 all are at least half a millennium younger than their original source. In these circumstances the metathesis sur- > rus- can be accepted as highly probable. The examples of this metathesis given by Vaillant and Jakobson have moreover since been augmented by H. G. Lunt, 13 who pointed out an example of coypbCKbiu > poycKbiu in the First Novgorod Chronicle. In addition, F. V. Mares 14 has noted other cases where Old Russian scribes transposed the unfamiliar ethnic adjectivespoyMbCK-, npoycKinto poycK-. To restore the phrase coypbCKbiMu nucMenu is an act of sober textual criticism, not of arbitrary fantasy. Istrin's historical arguments are based on the fact that at the time of St Cyril's journey to the Crimea Syria was under Arab dominion and that the Christian Syriac language was not in public use. This has little relevance to circumstances in the towns of the Byzantine Empire (such as Cherson) where, as Jakobson clearly showed, Syrians were commonly found in the 9th century. ' Op. cit., 61, n. 4. • T. Lehr-Splawinski, ¿ywoty Konstantyna i Metodego (obszerne) (Poznan, 1959), 30, n. 45. * V. A. Istrin, 1100 Aem CAaenHCKOu a30yxu (Moscow, 1963), 18-19, 103-6. 10

" B X e p c o H e c e ace, c o r j i a c H O CBHfleTejibCTBy B c e x

flBaauaTH

Tpex

floineflinHX

a o H a c CQUCKOB

'naHHOHCKOrO 3KHTHH', KOHCTaHTHH OÔHapyHCHJI ' E B a H T e j I H e ' H ' I T c a J I T H p b ' , HanHCaHHbie pyCCKHMH 6yKBaMH CpOyCfaCKHMH nHCMeHbl IIHCaHO')." 11

( O p . cit.,

18.)

F. Grivec-F. Tomsic, Constantinus et Methodius Thessalonicenses. Fontes (= Radovi Staroslavenskog instituta, 4) (Zagreb, 1960), 16. 12 Op. cit., 17. 13 Horace G. Lunt, "Again the pycbCKUMH nHCMeHbl", Cercetâri de lingvisticâ, III (1958), Supliment (= Mélanges linguistiques offerts à EmitPetrovicipar ses amis étrangers à l'occasion de son soixantième anniversaire) (actually appeared 1962), 324-6. 14 "CnaBHHCKa (jHUiononw", Mamepuajtu 3a V Meotcdynapoden Kompec Ha cjtaeucmume, II (Sofia, 1963), 69.

116

ROBERT AUTY

More recently E. Georgiev has returned to this question,15 adducing the argument that the text of VC speaks of pocbCKam nucMena. Now rosbsk- cannot, of course, be explained as a metathesis of surbsk-; but it is difficult to see why this reading, contained in three of the MSS of the south Slavonic branch, 18 should be regarded as original or authoritative. The MSS of the Russian branch universally have poyebeK- or something similar; and two of the South Slavonic ones have poyutKUMU (poyuiKbiMb).17 Moreover, F. Tomsiö confirms the view of P. Lavrov that the MSS of the Russian branch preserve the original text more reliably than the south Slavonic ones.18 It is thus clear that the forms in ros- must be regarded as a further modification of those in rus-, due no doubt to the interference of Byzantine Greek forms (Pö?, etc.) such as are quoted by Georgiev. F. Grivec' account of the life and work of Saints Cyril and Methodius,19 which must now be regarded as the standard monograph on the subject, accepts the view that Syriac, not Russian, was the language referred to in the disputed passage. It is unfortunate that the judicious conclusion of this great scholar has been ignored or overlooked in some recent publications. K. Ericsson,20 for instance, having built a carefully argued, if somewhat brittle, structure of hypotheses in order to prove that the Poljane had been converted from Byzantium in the early 9th century, accepts the Russian character of the Gospel and Psalter of Cherson without seriously considering the evidence for the Syriac interpretation. Philological method and historical plausibility alike must lead the unprejudiced student of the question to accept the probability — in this kind of investigation there are no certainties — that St Cyril's biographer wrote of eeamenbe u y/aamupb, coypbCKbiMU nucMemi

nucano.

It was not the aim of the present article to recapitulate the arguments adduced by Vaillant, Jakobson and Gerhardt, which, though disregarded by many students of the passage in question, can speak clearly to those prepared to hear. One further consideration may however be advanced in support of the Syriac interpretation. Gerhardt criticised Vaillant's extreme scepticism concerning the literal truth of VC. To doubt, as Vaillant had done, whether the Samaritan of VC VIII, 13-14, was anything more than a projection of the Good Samaritan of the parable was, Gerhardt reasonably remarked, "reichlich überspitzt". Yet we cannot doubt that formal principles of style and structure had a profound influence on the text and content of VC. The work has so far been too little studied from this point of view. The 15 E. Georgiev, "ÜHCbMeHHocTb poccob", Cyrillo-Methodiana: Zur Frühgeschichte des Christentums bei den Slaven 863-1963 (Cologne-Graz, 1964), 372-81. " See the variant readings given by Grivec-Tomsiö, op. cit., 109, 111. 17 Ibid., 111. 18 Ibid., 84. 19 Franz Grivec, Konstantin und Method: Lehrer der Slawen (Wiesbaden, 1960), especially 48-9. ,0 K. Ericsson, "The earliest conversion of the Rus' to Christianity", Slavonic and East European Review, XLIV (1966), 98-121, especially 118.

THE GOSPEL AND PSALTER OF CHERSON

117

importance of numerical symbolism in the Middle Ages is well known; and J. Vasica has shown the importance of the number 7 in VC.21 A similarly significant function attaches to the number 3. Constantine-Cyril's path to the goal of wisdom, prefigured in his boyhood dream (III, 1-7) proceeds in three stages: the study of Gregory Nazianzen (III, 17-20), the encounter with the foreign grammarian (III, 21-26), and finally study at Constantinople as the protege of the logothete (III, 27-32, IV, 1 ff.). Constantine-Cyril's public career consists of three missions, to the Saracens, to the Khazars and to the Slavs of Moravia. It is possible that the account of Constantine's stay in Cherson may have been influenced by a numerical principle of the same kind. Constantine first learns Hebrew (VIII, 10), then Samaritan (VIII, 11-14), and finally a third language (VIII, 15). It may well be that a third Semitic language was added by the learned hagiographer to complete the triad. That Constantine learned Hebrew at Cherson is probable; that he learned Samaritan is possible. Maybe he learned Syriac too; but it is equally possible that the famous gospel and psalter are fictional, part of the 'grammar of poetry' (if the author of this term will permit its extension thus far!) which informs the whole of the Vita Constantini. OXFORD

11

Sec V. ChaloupeckJ etc. (ed.), Na usvitu kfest'anstvi (Prague, 1942), 243.

P. H.

ABAHECOB

K HCTOPHH MOPOOHOJIOrHHECKHX qEPEflOBAHHH H OOHEMHOrO COCTABA KOPHEBblX MOP o p M a x OTpa3HJiHCb BO

B c e x cjiaBHHCKHx H3biKax H B e f l y T c B o e H a n a j i o OT e m e 6 o j i e e / i p e B H H x a n o x HH/IOeBponeñcKoro. O f l H a K o B i t e j i H x KOMnaKTHOCTH H3JioaceHHa MM n 0 3 B 0 J i H M c e 6 e H e i c o T o p y i o B o j i b H o c T b H HaHHeM C B o e n o B e c T B O B a H H e

c

'anafora',

c

Toro,

HTO M H H a x o ^ H M

B

c o B p e M e H H O M p y c c K O M A3biKe. I l y T e M B H y T p e H H e ñ p e K O H C T p y K u i i n Mbi ycTaHOBHM o c H O B H b i e S T a n t i B pa3BHTHH p a c c M a T p H B a e M b i x HBJICHHH, n o / i K p e n n B HX a a H H H M H HCTOpHH pyCCKOrO H3bIKa. B c o B p e M e H H O M p y c c K O M H3biKe y M e H b m H T e j i b H b i e OT HMCH c y m e c T B H T e j i b H b i x OCHOBOH

Ha

TBepAyio

napHyio

corjiacHyio

MoryT

06pa30BaTbca

npH

c

noMoum

(C ' ô e r j i b i M ' o ) . H a n p H M e p , eo3—eo3ÓK, zod—ZOÒÓK, zópod— zopodÓK, dy6—dyòÓK, 3ad—3adÓK, uac—uacÓK, CUH—CUHÓK, ifeem—tfeemóx, Komè'A—KomeAÓK, mépeM—mepeMÓK, H T.FL.; pù6a—pùÒKa ( c p . púóoK — p o f l . n . MH.H.), 6á6a—6á6xa, Kopóea—Kopóem, mpaeá—mpáetca, déea—déeKa, zoAoeá— zoAÓeKa, óopodá—SopóÒKa, 6epë3a—6epë3Ka, MOZÚAÜ—MOZÚAKÜ, cmená—cméma, mponá—mpóma, zopa—zópna, uápa—mpKa, ceúma—ceúmKa, H T.FL. Cytjx})HKC -(Ó)K-, KAK U3eecmm, BOCXOAHT K a p e B H e p y c c K O M y -SK-. cyhk—

dyaá—dyotcxa,

KHÚaa—

coxá—cóuixa; C

pyKá—pyuKa.

6WK—ÒUHÓK,

pexá—

TOHKH 3pcHHH CHH-

XpOHHblX OTHOIIieHHH OCTaeTCH HÊaCHblM, nOHCMy B 06pa30BaHHHX COBpeMeHHOrO pyccKoro a3biica c cycJxjjHKCOM -(O)K- npn OTCYTCTBHH nepefloBaHHa napHbix TBepaux corjiacHHX c cooTBeTCTByiomHMH MsrKHMH 3aflHeH3biHHbie qepeflyioTca c IUHIIHmHMH. Beflb npH HajiHHHH HepeflOBaHHH a ¡I Me, x/¡tu, K¡/H cjieflOBaJio 6bi oacHflaTb Hepe^oBaHHH w / / w ' ,

C//C', H T.A. HJIH: npn

c//c', H T.fl. (tfeem—ifeemoK,

OTCYTCTBHH nepeflOBaHHH m¡¡rrC, CJieflOBaJio 6bi oxcHflaTb Taicxce

esuma—ceumxa)

OTcyTCTBHH nepeflOBaHHa 3aflHea3biHHbix c uiHnamHMH (T.e., 6oa—óoaox,

coxa—

COXKÜ, H T.fl.). OTMeneHHaa 'aHOMajraa' MoaceT 6biTb noHHTa TOJibKo B TOM cjiynae, ecjin Mbi flonycTHM, MTO B npouiJioM nocjie napHbix TBepflbix corjiacHbix, c oflHoñ cTopoHbi, H 3aflHeH3biHHbix, C flpyroií, B flpeBHepyccKOM 6MJIH pa3Hbie cy({)HKCbi OTJIHHaJIHCb B CBOeM (J)OHeMHOM COCTaBe H TeM CaMblM npeflCTaBJISJIH C 0 6 0 Í Í flBa pa3Hbix cymHxca Ha mnnamne. Ilpeacfle NEM M H nonbiTaeMca oôiacHHTb 3Ty 'aHOMajiHio', paccMOTpHM o6pa3oBäHHe YMEHBIHHTEJIBHBIX OT HMCH Ha Mancyio napHyio corjiacHyio. YMeHbuiHTejibHbie OT HMCH cymecTBHTejibHbix Ha -a c OCHOBOH Ha Maricyio napHyio

corjiacHyio

OÖHHHO

KaneAbKa, do AH—doAbxa,

coxpaHaiOT

mwpn—mwpbxa,

ee.

Cp.

nedenn—HedenbKa,

TâKHx cjioB OTHocaTca TaKace coßcTBeHHbie HMeHa, HanpHMep, BaAH—BaAbKa, a Tame

Baca—Bacbxa,

Kama—KambKa,

TaKHe cjioBa KaK dndn—dndbKa,

xanAH—

DBWH—dbiHbKa. K x p y r y ,

6ÜHH—6amxa,

edn—edbKa, mnmn—mnmbm,

Bap

H—Bapbw,

Ilemp—IJembKa,

6amn—ôambxa,

H

HeKOTOpbie flpyrne. OflHaKO cymecTBHTejibHbie, HMeiomne nepea < H ' ) corjiacHyio, a TaKxee , o6pa3yK)T yMeHbHiHTejibHoe c 3aMeHoñ < H ' ) Ha : commi—comeHKa,

Keaumn—KeaiuoHKa,

KOAOKOAbHH—KOAOKOAeHKü, 1

euiunn—euuiema,

ÓOÜHH — ÔoeHKa,

H T.fl. 1 C

neenn—necema. cnaAbHR—cnaAema,

3aMeH0H MarKOH IiapHOH

EflBa JIH He eflHHCTBeHHWM HCKJIIOHCHHCM HBJMeTCH ,

120

P . H. ABAHECOB

corjiacHoñ

Ha cooTBeTCTByromyio

oT^EJIBHME

cjioBa

memepa—memepna

c

oflHoñ

TBepayio

corjiacHoñ

B

o6pa3yioT

yMeHbniHTejibHbie

Henp0H3B0flH0H

OCHOBC,

(HO HMEETCA T a o c e 06PA30BAHHE memepbKa),

mémn—merma.

P H ^ O M C npHBejteHHbiMH yace o6pa30BaHHHMH mama—msimbKa, Hejii>3H He oTMeTHTb dumn—demm,

a Taicace

Taicace

HanpHMep,

6amn—óambKa

Mamb—Mamita.

H e T n0CJiefl0BaTejibH0CTH B 06PA30BAHHH yMem>uiHTejibHocTH OT HMCH c y m e CTBHTejibHbix MyaccKoro

poaa

6e3

0K0HHaHHa

c

OCHOBOÍÍ

Ha M a r a y i o

napHyio

corjiacHyK). CorjiacHbie < H ' > , , , < T ' )

Henp0H3B0flHbix cymecTBHTejibHbix B npoH3BoflHbix HajiHie-

cTByeT nepeflOBaHHe —

3aMeHa STHX

TBepflbie:

zÓAyñb—zoAySÓK,

AOKomÓK,

Aánomb—AanomÓK,

Medeédb—MedeedÓK, (npHMepw

MJITKHX corjiacHbix Ha

nózomb—HozomÓK,

zpy3db—zpy3ÓÓK.

flHajieKTHoe)

cooTBeTCTByiomHe

KÓzomb—KozomÓK,

zeo3db—zeo3ÓÓK.

CM. Bbmie) OT 3eepb

cTopenHoe H

mn3b—KHH3SK

(?). O^HaKO O6HHHO B 0 6 p a 3 0 B a H M x c Maricón OCHOBOÍÍ

ripii

MHe

AÓKomb—

H3BecTHbi

xopb—xopeK,

¿maneKTHbie:

napb—AapéK,

o6pa3yeTC» He TOJIBKO 3eepeK,

H

HO Taicace

flp.

(npo-

3eepÓK.

OTCYTCTBYET noejiefloBaTejibHOCTb Taicace B 06pa30BaHHH yMeHbniHTejibHMx OT HMEH cymecTBHTejibHbix aceHcicoro p o j i a c OCHOBOÍÍ Ha Maricyio napHyio corjiacHyio. C p . cmynéhb—cmynémKa, amapéjibua, UlUHéAb

HO

— lUUHéAbKa

o 6 m e e KádKa, OT deepb

KOAbióéAb—KOAu6éAbKa,

peAb—peAKa,

H lUUnéAKO.

enb—eAtca,

amapénb—• SyméiAb—óymúAKa,

C l O f l a M03KH0 OTHeCTH TaKKe AHajieKTHOe Kadb

a Taicace Mamb—MámKa.

: deépna

a Tanate flHajieKTHoe deépbKa. B 06pa30BaHHax

CorjiacHbie

paccMaTpHBaeMoro

Bceraa nepe^yioTca c oooTBeTCTByioineH TBep/joñ: mempádb—mempádxa, AoiuáÓKa,

nAÓufadb—nAotifadna,

zpydb—zpydna,

6poeb—ópóexa,

ceAbdb—ceAeÓKa,

nAemb—nAemtca,

tfépKoeb—tfepKÓeKa,

YJCA3AHHME HENOCJIEFLOBATEJIBHOCTH H OTKJIOHCHHH — dpyoKÓK, 3eepÓK,

zopá—zópKa, 3nmb—3xmeK,

H T.FL. —

H

HO HOZÓ—HÓMCKÜ, HO nózomb—HozomÓK,

3eepb—3eepeK, 6ámn—SámbKa,

Tana

AÓuiadb— cemb—cémKa,

MopKÓeb—MopKÓena. Aec—AÉCÓK, HO Taicace HO

HO H

dpyz— seepb—

Mamb—MámKa,

Jienco O6T>HCHHK»TCH, e c j i n M U npHMeM BO BHHMaHHe 0CH0BHbie 3Tanbi B

HCTOpHH yMeHbUIHTejIbHblX 06pa30BaHHH, C OAHOÍÍ CTOpOHbl, H HCTOpHH (¡DOHOJIOrHHecKoií cHCTeMW, c a p y r o H —

B HACTHOCRA OTFLEJIEHHE uiHnamHX OT 3AFLHEA3BIHHBIX

H3BCCTHoro eme B ApeBHepyccKOM H3biKe H oTHOCHBiuerocfl K MHTKOÜ pa3H0BHZiH0CTH cymecTBHTenbHHX xceHCKoro pofla Ha -a. Cp. (JiopMy pofl.n.MH.H. . BnpoieM, B flnajieKTax HMeeTca H .

K H C T O P H H MOP3>OHOJIOrH1ECKHX KaK ocoGbix (J)OHeM H pa3BHTHe KoppejiaiiHH 4)0HeM006pa3yK)meM KaTeropHH.

121

HEPEflOBAHHH

MHTKHX

TBepflbix corjiacHbix KaK

~

H a OCHOBe flaHHblX CpaBHHTejIbHO-HCTOpHHeCKOH rpaMMaTHKH CJiaBHHCKHX H3bIKOB H HCTOpHH pyCCKOrO 33bIKa (nO IIOKa3aHH5IM KaK nHCbMeHHblX naMJITHHKOB, T A K H FLHANEKTHBIX H C T O H H H K O B ) MOXCHO Y T B E P A C A A T B , H T O yMeHbiiiHTejibHbie H M C H A cymecTBHTejibHbie 06pa30BajiHCb c 3neMeHTOM -K-, KOTopbiií npncoe/iHHsuicH K TeMe í HJIH ü. .UojiroTa H KpaTKocTb, a TaKace KanecTBo rjiacHoñ TeMbi 6 H J I H , B H J I H M O , 06ycJi0BJieHbi xapaKTepoM rjiacHoñ HenpoH3BoaHoro cymecTBHTejibHoro: rjiacHbie HeBepxHero N O ^ T E M A — a o j i r a a ( T . e . , á) 3aMeHa:iacb A O J I R O Ñ rjiacHOH i, KpaTKaa (T.e., ó) — KpaTKoñ rjiacHoií í; rjiacHbie BepxHero noflteMa ú n ü , a TaK»e l coxpaHajiHCb. TaK nojiyrajiHCb cy(})(J)HKCM - i k - > - i c - ; - i k - > - b c - , -ük- > - y k - (npea-ük> -bk(HanpHMep, cbmbKb). CTaBjieHbi e / i H H H H H b i e cjiynan, HanpHMep, xaMbiKb), 3aecb M H He KacaeMc» a p y r a x 06pa30BaHHÜ — ™ n a i f b p K b e b i f a , ifbpKbeuifa; cbmoebifb. C TeneHHeM BpeMeHH — oflHaKo, B H A H M O , eme B no3^HeM npacjiaBHHCKOM — B pe3yjibTaTe Mop^ojiorHHecKoñ aHaJiorHH HMeeT MecTO B3anMHoe npoHHKHOBeHHe cy(J>4)HKca -bk- B cjioBa c O C H O B O H Ha ü (cp. .ap.-pyc. d o M b i f b ) , H cycjx^HKca -bkB B M . 6oJiee paHHero tferbmbtfb), a 3aTeM H B O C H O B M H a ó (cp. HanpHMep, iferbmbKb flpyrae O C H O B W . B pe3yjibTaTe He BnojiHe acHbix nponeccoB CYTJXJMKC -ic- npoHHKaeT B Mancyio pa3HOBHflHOCTb 6 OCHOB. Cp. eme B CTap0CJiaBHHCK0M KopaÓAb— a TaKvKe ap.-pyc. HOMCb — HOMCUlfb, Meub — Menuifb, H APKOpaÓAUlfb, E c T e c T B e H H O , HTO 3 a O T e a 3 b i H H b i e n e p e a p e i f m e K c a M H í c M a r n a j i H C b , B CBSBH C n e M B

no3flHeM

npacjiaBHHCKOM

Mbi

HaxoflHM

k / j c ,

nepefloBaHHe

g / / d z > z , x / / s .

O c T a j i b H b i e H e c M a r i e H H b i e c o r j i a c H b i e KaK (J)oHeMbi c o x p a H H J i H C b B S T O M n o j i o x c e H H H 6e3

H3MeneHHH, n p n o ó p e T a a

nojiyMHTKHx

BapnaHTax.

SMaHCHnHpOBajIHCb HanpoTHB,

OT

nojiyMHTKHe

no3HimoHHyio

H3BecTHo,

HTO

3a,HHefl3bIHHbIX corjiacHbie

nojiyMarKOCTb, T.e.,

yace

H

B

CTaJIH

eme

B

Bbicrynaa

xtonHCbMeHHyio

nopy

CaMOCTOaTeJIhHbIMH

paHHioio

nHCbMeHHyio

B CBOHX

UMNAIUHE (J)OHeMaMH.

nopy

npe^-

CTaBJIJIJIH C 0 6 0 Í Í n 0 3 H U H 0 H H b i e B a p H a H T b l C O O T B e T C T B y i O m H X H e C M K T H C H H h l X 4 > O H e M . TaKHM

06pa30M,

6 M J I O : déva—dévica

H r é k a — r é c i c a , lésb—lésbcb

méxb—mésbcb,

H

z v é b r b — z v é r b c b ,d v b r b — d v b n c a , s y m — s y m k b , k a m y — k a m y k b .

IlpHBeAeM Bbi6opoHHC> MaTepnajibi apeBHepyccKoro H3WKa.2 O C H O B O H H a HecMarneHHyio c o r j i a c H y i o (KpoMe 3 a / J H E A 3 B I H H B I X ) : eoda—eoduifa, K03utfa,

z o p a — z o p u i / a ,

¿paMoma—zpaMomuifa,

K o / i o d a — r c o A o d u i f a , Kbpcma—Kbpcmutfa,

M H 3 d p a — M A j d p u i f a , neAeua•—neAemifa, p u ó u t f a , ceuma—ceumuifa,

rjiacHyio, a TaKxce :

3eMAH—3eMAutfa,

HMCH

C

HMCH

Ha

-a

c

óbueAa—óbueAuya,

dn>ea—dmeutfa, naea—Aaeutfa,

nonnea—nonneui^a,

y d a — y d u i f a ; OT

OT

ko3ü— MemAa—MemAui^a, nbma—nbmui/a,pbi6a—

OCHOBOH

etbKbuia—erntcbuiuifa,

Ha cMarneHHyio coK a u i a — K a u i u t f a ,

1 B H a c T o a m e f t c T a n . e , HOCHIUHA xapaKTep n p e f l B a p i r r e j n . H o r o cooómeHHH, Hcn0Jib3OBaHbi JIHIHB HeMHorHe aaHHbie ^peBHepyccKoro H3biKa, pyccKHXfluajreKTOBH yxpaHHCKoro H3biKa, B OCHOBHOM noiepnHyTbie H3 cjioBapeK H . H . Cpe3HeBCKoro, B . H . U a n á , k B . fí. rpamemco. H p y r m MaTepHam.i no 3THM jBbixaM HflaHHweflpyrnxcJiaBJiHCKHx H3biKOB 6yayT npeflCTaBJieHH B 6onee npocTpaHHoñ nySjiHKauHH Ha TeMy 0 6 o6pa30BaHHH yMem>nmTein>HMX HMCH cyiuecTBHTeJibHbrx B cnaB«HCKHx H3bLKaX.

122

P. H. ABAHECOB

KOMCÜ—xoofcuifa,

nauta—uatuuifa,

xbiMcutfa, odeMcda—odexcduifa, B

06pa30BaHHax

Abotca—MMcuya,

Kyvya—Kyufuija,

OT HMeH « e H C K o r o

posa

poma—poufuifa,

3mwh—3Mn>utfa, H a -a

xbiotca—

xenun—xeAbuifa.

c OCHOBOH H a

3aflHea3MHHbie

HaJiHnecTByeT n e p e a o B a H H e nocjieflHiix c u i H n a m H M H : eepuza—eepuomiiia,

Kombivcutfa,

Kbmza—KbHUMCuifa,

prbKa—prbHutfa,

(OT

coponuifa

Myxa—Mynuifa, copoxa),

xomuza—

paxa—pauma, (OT

cpauuifa

pyxa—pyuuifa,

cpaxa),

Myxa—Mbtuuifa,

dbCKa—dbiuuifa. K

npHBefleHHbiM

cnoBa

npHMepaM

apeBHepyccKoro

MONCHO n p n 6 a B H T b ,

a3biica,

fljia

KOTopbix

HanpHMep,

eme

MHe H e H 3 B e c T H b i

cjieayiomne

MOTHBHpyromHe

Henp0H3B0flHbie cjioBa: 3acmaeuifa (3acmaea), xopuifa (B cneuHajibHOM 3HaneHHH, BHflHMO, OT Kopá), xpadutfa ( ' K o p o 6 ' ; BHAHMO OT xpaóa), naAuifa (OT nanal), (OT y nal), ymuifa (OT yma npH 6ojiee CTapoM ymbi—ymbeé), (HenocpeflcTBeHHo

nyzbeuifa

OT

tfbpKbeutfa (HenocpeACTBeHHO

npa

nyzbea

6ojiee

npn

OT tfbpxea

CTapoM

Sojiee

y/iuya

nepenutfa (OT uepenal), nyzbi—nyztee),

CTapoM

tfbpxbi—ifbpKbee),

HHrbUlfa (OT HHTbH). 3 T H c T a p b i e 0 6 p a 3 0 B a H H a H a -uita pyccKHX

AHaJieKTax

6o;ibmeH

B 3HaHHTejn>HOM R O J i n n e c T B e c o x p a H a K > T c a B

OTMECTH B C T a p o M

HJIH M e H b m e ñ

3HaneHHH y M e n b u j H T e j i b H b i x ,

CTeneHH o T o p B a B u i H C b

OT c e M a H T H K H

a

name,

B

MOTHBHpyiomero

C J i o B a , H M e i o T TO HJIH H H o e c n e u H a j i H 3 H p o B a H H o e 3 H a n e H H e . I I p H B e f l e M HecKOJibKO npHMepoB

( n a c T b i o H3 c j i o B a p a

B.

H.

¿fajia, nacTbio

cTBeHHo no Ha6jiio,zieHHaM Ha/i roBopaMH):

H 3 B e c T H b i x MHe

6á6utfa,

zoAoeútfa,

Henocpefl-

eÚAa—eÚAUifa,

déeúifa, úzAuifa, usóuija, uxpúifa, xúmuifa (OT xúma — CTeóejib), xoAÓduifa,

xocútfa,

Aáeuifa (OT Aáea), uopúifa, náAuifa, nemAÚifa, nopówuiia

ywcúifa

(OT yoicá), ympóóuifa,

(OT nopóuia), púóuifa,

(OT yxá), MDfcútia (OT Mza).

yuiúifa

OT HMeH Ha -o 3 c OCHOBOH Ha HecMarneHHyio corjiacHyio (KpoMe 3aflHea3biHHbix) flpeBHepyccKoro

OTMCTHM: ópamt—ópambifb,

He BCTpeieHHoro B naMaTHHKax enm), zpadbifb,

deopb—deopbtfb,

zeopb—zeopbtfb,

doMb—doMbtfb,

erbHbtfb (OT

eopom—eoponbifb,

zopodb—zopodbtfb,

xAum—XAunifu

zpadb—

(MH.H.), KOAOKOAb—

KOAOKOAbifb, KOAb—KOAbifb, KOHb (HJIH KOHb c HecMarneHHOH cornacHoñ nepe,a 4>HKCOM -IC- > -bC- BOCXOflHT e m e K IIpaCJiaBHHCKOMy, (J)OpMy KbHXJKblfb

MO»HO

CHHTaTb ApeBHeHIIIHM H HOpMaJIbHblM 0 6 p a 3 0 B a H H e M C HepeflOBaHHeM gl¡Z

Ot

cymecTBHTejibHbix

MyaccKoro

Mcepaeb—wcepaebifb,

poaa

c

ochoboíí

Ha

i nocjie

Aapb—Aapbtfb,

3TO CJIOBO paHO ot cymecTBHTejTbhhx

n p H 0 6 p e j I 0 MarKOe

Hbpeb—Hbpebifb;

»eHCKoro

HecMHrneHHbix c o r j i a c H b i x : ebpeb—ebpebifa, KJirbmb—KATbtnbifa,

TaK»ce Mbiuibtfa CTapbie

a

hjih b

pa3Horo

caxapéif,

mneca

ceemétf, nymétf b

Sopéif

b

BHfle -(e)if-

Taioice

b

coBpeMeHHOM

nocjie

uemebpmb—nemeb-

a

oecéif,

xeocmétf,

pyccKHX

opAétf,

ot

HanpHMep:

6ypaeétf,

ch6hpckhx,

eoAnéif.

/icá T ' u ' a /

jiHTepaTypHOM

M0p(f)0H0Ji0rsiHecKHx

ot hmch

yMeHbiiiHTejibHbie

flpyrnx

necéif,

óymyn—

KaBKa3CKHx"),

npyméif,

pxdéif,

O t m c t h m 3,uecb Taicace

z'em'el'ka, hjih

hjih sucok;

kon'ok, zv'ër'—

131

K HCTOPHH MOPt>OHOJIOrHHECKHX HEPEflOBAHHH

zvërok, golub' mys—myska.

(HJIH

golub)—golubók,

dv'er'—dv'erka

dv'orka,

HJIH

klët'—klëtka,

KaK BHflHo H3 npHMepoB B 3TOT nepHOfl KaK H paHee TBep^aa napHaa corjiacHan HenpoH3BOflHoro cJioBa coxpaHaeTca B np0H3B0flH0ñ; 3aflHe«3biHHaa n e p e a y e T c a C n i n n a m e l i . XapaKTepHoñ ocoôeHHocTbio 3TOH snoxH aBjiaioTca

Henp0H3B0flHbie

ocHOBbi Ha MflrKyio napHyio corjiacHyio (J)OHeMy : OHH B OAHHX c j i y i a a x coxpaHaioTca B np0H3B0,HH0M yMeHbiiiHTejibHOM (kon'—koríok,

car'—car'ok),

3aMeH5HOTCH cooTBeTCTByiomcH TBepAOH (p'ecerí—peconka, golubok).

Mop4>OHOJiorHiecKoe

HepeflOBaHHe < t ' / / t )

B apyrnx

zvêr'—zvërok,

cTaHOBHTca

—

golub'—

cymecTBeHHoii

npHMeTOH 06pa30BaHHH yMeHbuiHTejibHbix OT 3HaHHTejibHbix r p y n n cymecTBHTejibHblX.

( 6 ) B 6 o j i e e no3flHHH n e p n o f l 3TOH s n o x H n p « coxpaHeHHH Bcex nepeHHCJieHHbix Bbirne THiioB o6pa30BaHHH yMeHbuiHTejibHbix noaBjiaeTca HOBMÍÍ THII — C ycTpaHeHHeM nepeflOBaHHH < t ' / / t > , T.e., C coxpaHeHHeM Mancoìi n a p H o ñ corjiacHoñ B yMeHbuiHTejibHOM. IIoaBjiaiOTca TaKHe 06pa30BaHHa KaK zv'ër'—zv'ër'ok,

p'en'ók,

stup'én'—stup'en'ka,

akvar'el'—akvar'el'ka,

/majieKTHoe

p'en'—

dv'er'—dv'er'ka.

CnejiyeT o T M e n r n . , HTO B KaacflbiH nocjieayiomHH n e p n o f l npo^ojiacaioT c y m e CTBOBaTb MHorae H3 THIIOB yMeHbuiHTejibHbix 06PA30BAHHÌI npe^biAymnx s n o x , HO OHH npHoôpeTaioT HJIH flHajieKTHbra xapaKTep, HOT Ty HJIH HHyio CTHjiHcraHecKyio oKpacKy, HJIH 3aKpenjiaioTca B onpeaejieHHbix 4>pa3eojiorHiecKHx o ô o p o T a x , HJIH, HaKOHeii, cneuHajiH3HpyioTca B 3HaHeHHH, HanpHMep, cTaHOBHTca HA3BAHH6M pacTeHHH, »CHBOTHORO, npHcnocoôJieHHa, n a c r a o p y a n a , H T.A., T.e., nepecTaiOT 6biTb yMeHbuiHTejibHbiMH. HepeflKO CTapbie yMeHbuiHTejibHbie 3aKpenjiJHOTca B TOIIOHHMHK6 H aHTponoHHMHKe. C p . HanpHMep, Ha3BaHHH peK —

ffouetf,

BoAxoeetf,

flyôetf,

HmyAeif (cp. p. Mmyji), Ko/ieuya (MypMaH. O6JI. ; cp. p. Koma TOH xce o6jiacTH), Cbueutfa (cp. p. CbiAeà), Ha3BaiiH5i r o p o j t o B Peuutfa, reo3Òeif (HepHHr. O 6 J I . ) , rpodetf (MorHJi. O 6 J I . ) , ffeopeii (Ha p. A H r a p a , KpacHoap. OÔJI.), KaMeneif (B HecKOJibKHX MecTax), Ko3ejiei{ (HepHHr. O6JI.), Koponetf (HepHHr. OÔJI., cp. Kopon TOÎÎ »ce o6jiacTH), KpeMenetf (TepHon. O6JI. Y C C P H MHHCKOÍÍ O6JI. B C C P ) , Kpecmifbi ( H o B r o p . O6JI. ; cp. Kpecmbi TOH ace OÔJI.), Kypeneif (B pañoHe r. MOJIO^ENHO, B C C P ; cp. Kypem, ^lepHHroB. OÔJI.), Ocmpoeetf (B panoHe r. MojioaeHHO, B C C P ) , Ilaee/iey (Pa3aH. OÔJI.). OcoôeHHO HHTepecHO Ha3BaHne Ocmpootcey (POBCH. OÔJI., Y C C P ) , coxpaHHBinee CTapoe yMeHbuiHTejibHoe c cyaMHJiHH Tpyôutfbin, TaKxce

HenocpeACTBeHHO

jiHHHbie HMeHa B pojiH (J)aMHJiHH — CeMenetf,

IIuAunetf,

CTapbie PoManetf,

MeaHifoeo,

JIxuMifee,

yMeHbuiHTejibHbie H T.a.

HaxoHeii,

OHH COXpaHHIOTCH B pH^MaX H aCCOHaHCaX HapOflHOH n033HH, B nOCJIOBHIiaX,

132

P. H. ABAHECOB

3araflKax, h T.fl. Cp., HanpHMep, b pyccKOM a3biice: nee y H neeetf nauieA xAeeeif,

nnmb mbicm

oeetf (y/ieü);

KOAOH

XAeeei} 6eAUX

pomoK He ceoü xAeeoK, ne 3ameopuuib;

MOKU

e MM

oeeif {pom u 3yôbï); c p . : uyotcoü

MOU

Mwoeutfbi, 30A0mbie eoAoeuifbi;

b yKpaHHCKOM : MOAOÔÏÏ MOAoduifi, 3aeueaüme eoAoeutfi; ebinuüme do dentin, — maM, HeMa

EEMOHIFN

; eede

KOHH

y cmamuifio, a

MUAOZO

y ceimnuijw ; npuütuoe uepneifb,

noAoxcue nüifM e yzneifb — nexaü mnAnrrib, nom npuüdy orìnmb. OflHaKO Bee STO BbixoflHT

flajicKo 3a

CToameM

cooömeHHH.

HaMeneHHbie

STanbi

npeflejibi BonpocoB, KOToptie MoryT 6hti> ocBemeHbi

b hctophh

yMeHbiUHTejibHbix

cymecTBHTejibHbix,

KaK

b Ha-

HaM

KaaceTCH, IIOflTBepXCAaiOTCH AaHHbIMH He TOJIbKO BOCTOHHOCJiaBHHCKHX H3HKOB, HO TaKace

AaHHbie

h, b ocoôchhocth, ioacHocjiaBHHCKHX npeACTaBjieHbi b A p y r o ì i p a ö o T e .

3ana£H0CJiaBHHCKHx 6yayT

a3biKOB.

3th

MOCKBA

A N D R E I AVRAM

DE LA LANGUE QU'ON PARLE AUX ENFANTS ROUMAINS

1. INTRODUCTION

O

N dispose, à l'heure actuelle, d'une grande quantité de matériaux concernant LE LANGAGE DES ENFANTS et aussi de quelques études précieuses consacrées aux problèmes théoriques d'un domaine de recherches qui s'est avéré extrêmement important pour la linguistique générale; il suffit de mentionner, dans cet ordre d'idées, le célèbre travail de M. Roman Jakobson, Kindersprache, Aphasie und allgemeine Lautgesetze,1 dont le caractère novateur a été, à juste titre, souligné maintes fois. Le domaine connexe des recherches sur LA LANGUE QU'ON PARLE AUX ENFANTS (angl. baby talk)2 est dans une situation moins favorable. Les matériaux recueillis systématiquement jusqu'en ce moment proviennent, à ce que nous savons, d'un nombre restreint d'idiomes. Sur le plan théorique, on a déjà valorifié une partie des données existantes et plusieurs auteurs (parmi eux, M. Roman Jakobson, dans son travail précité et surtout dans "Why 'Marna' and 'Papa'?" 3 ) ont attiré l'attention sur l'importance de ce genre de recherches. Nous croyons qu'à présent l'enrichissement des données est absolument nécessaire. Notre article — que nous dédions au grand maître de la linguistique enfantine — contient un nombre de faits caractéristiques de la langue qu'on parle aux enfants roumains; 4 nous y avons ajouté quelques observations qui pourraient, éventuellement, être d'un intérêt plus général. 1

Voir maintenant Roman Jakobson, Selected Writings, I: Phonological Studies (La Haye, 1962), pp. 328-401 (les renvois des notes qui suivent se rapportent aux pages de ce recueil). ' Nous allons employer, dans les pages suivantes, l'abréviation LE pour désigner 'la langue que parlent les adultes en s'adressant aux enfants' (et qui ne doit pas être confondue avec 'le langage enfantin', que parlent les enfants eux-mêmes); par LN nous allons désigner la langue 'normale'. Voir, pour plus de détails concernant la définition de la LE et les questions de terminologie, Roman Jakobson, op. cit., 331, 538-539; Robert Austerlitz, Word, XII (1956), 260; Charles A. Ferguson, in For Roman Jakobson... (La Haye, 1956), 121; Ashok R. Kelkar, Word, XX (1964), 40-41. 3 Roman Jakobson, op. cit., 538-545.

134

ANDREI A VRAM 2. P H O N É T I Q U E ET P H O N O L O G I E

Avant de présenter nos remarques sur les particularités phoniques de la L E roumaine, nous soulignons la nécessité de faire une distinction entre les sons rencontrés dans les mots qui constituent le lexique spécial de la LE et les sons qui apparaissent dans les mots 'normaux' (c'est-à-dire appartenant à la LN) lorsque ceux-ci sont altérés par l'adulte s'adressant aux enfants. Pour ce qui est de la première catégorie de faits, notons, tout d'abord, l'absence presque totale, dans l'inventaire phonique de la LE, des consonnes fricatives; seules [f] e t [3] s °nt un peu mieux représentées, dans ar§, àr§ [arf], (h)îr§ [(h)irf], bîja ['bÏ3a] — tous avec le sens 'ne touche pas!, ça brûle!' (aucun de ces mots n'est d'un emploi général). Chose assez surprenante, la consonne [r] est présente dans l'inventaire, quoique très peu fréquente: en dehors de ar§ et de ses variantes, nous pouvons citer le mot bri-bri [,bri'bri] 'propre, beau, joli, agréable' (très peu répandu du point de vue géographique; les mêmes notions sont, beaucoup plus souvent, exprimées par titi, gigea, etc., discutés ci-dessous). Il n'y a rien d'intéressant à remarquer au sujet de l'inventaire et de la fréquence des voyelles. Nous signalons seulement le détail suivant: les mots enregistrés par nous avec le sens de 'propre, beau, joli, agréable' contiennent au moins une voyelle prépalatale,5 tandis que le mot qui désigne les notions de 'sale, laid, désagréable' contient une voyelle médio-vélaire; on a ainsi, d'une part, titi, cici ['tjitji], cicea ['tjitja], cecea, gigea ['d3¡d3a], avec une voyelle 'claire' sous l'accent (et, parfois, aussi dans la syllabe inaccentuée) — d'autre part, cî(h) [ki(h)], avec une voyelle 'sombre'. On est tenté de voir ici — donc chez les adultes — un fait de symbolisme phonétique semblable à celui qui apparaît — pour donner un exemple frappant — chez la fillette norvégienne qui a baptisé Mump og Mippe [momp â mi(p)ps] deux statuettes grotesques dont la première avait un aspect 'sombre', l'autre un aspect 'souriant'.8 Notons cependant qu'on rencontre une situation différente ailleurs; cf., dans la LE arabe, [daHH] 'pretty, nice, good, clean', d'une part — [PiSî] 'dirt, dirty', [kixx] 'dirty, bad', de l'autre (à côté de [kaSS] 'dirt, dirty').7 En ce qui concerne l'adaptation phonétique des mots qui proviennent de la LN, en laissant de côté les faits connus presque partout (comme le remplacement de la consonne [r] par [1] ou la simplification des groupes consonantiques), nous allons signaler deux phénomènes: (a) L'emploi de l'occlusive glottale à la place de la fricative (sourde ou, parfois, 4 Les faits cités en exemple représentent une partie de ceux que nous avons recueillis à l'aide d'une enquête ou que nous avons trouvés dans l'Atlas linguistique roumain, dans les dictionnaires et dans diverses études. 5 Presque tous contiennent aussi des affriquées palato-alvéolaires ou alvéolo-palatales (nous n'avons pas noté tous les détails d'ordre phonétique) qui, selon certains spécialistes, peuvent exprimer l'idée de 'petit' ou de 'joli'; voir Iorgu Iordan, Stilistica limbii române (Bucarest, 1944), 85.

•

A l f S o m m e r f e l t , Norsk tidskrift for sprogvidenskap, I ( 1 9 2 8 ) , 30.

' Charles A. Ferguson, loc. cit., 122-124. Selon J. Cassar-Pullicino, Orbis, VI (1957), 194, dans la LE maltaise il y a un [u :] 'denoting endearment and affection'.

DE LA LANGUE QU'ON PARLE AUX ENFANTS

135

sonore) correspondante. Lorsqu'on parle aux enfants, des interjections comme hopa ['hopa] (à diverses valeurs), haï [haj] 'viens!, allons!' deviennent, très souvent, ['?opa] et, respectivement, ['?aj]. La même occlusive apparaît aussi dans les mots spécifiques de la LE ['?a'?a], ['?â'?â] et ['?ï'?i] 'matières fécales, défécation'; elle est rare dans la LN. (b) Le remplacement des consonnes [tj], [d3] par [ts], [dz]. Des phrases comme Ce face? ['tje fatje] 'Qu'est-ce qu'il fait?', Cine plînge? ['tjine ,plïnd3e] 'Qui pleure?' deviennent ['tse fatse], ['tsine ,p(l)indze] (à noter que l'affriquée [dz] manque dans la langue littéraire et dans la LN de la plupart des régions de la Roumanie). 8 Ajoutons, à propos de ce phénomène, une remarque sur la 'dialectologie' de la LE roumaine. Il y a des parlers caractérisés par la transformation, dans certaines conditions, des affriquées [tj], [d3] en fricatives du type [f], plus ou moins palatales; par conséquent, les formes face, plînge, citées ci-dessus, sont prononcées ['faee], ['plinze] ou ['pli^e]. A ces parlers correspond une variante de la LE dans laquelle [0] et [z] ou [3] sont remplacés par [s], [z], suivis, d'habitude, d'un faible élément semi-vocalique: ['faste], ['p(l)inzte]. Les faits discutés dans les lignes précédentes, sous (b), nous rappellent un phénomène semblable rencontré dans certains idiomes de l'Asie du Nord-Est, où un même phonème se réalise comme [te] chez les hommes et comme [ts] chez les femmes.9 D'autre part, il paraît que la préférence pour les consonnes [s], [z] (seules ou combinées avec les occlusives ayant le même point d'articulation) puisse être comparée au phénomène rencontré par A. Meillet chez une mère française et qui consistait dans le remplacement de [3] (en parlant des petits genoux de l'enfant) par une consonne très proche de [j]. Les prononciations de ce type seraient, selon le linguiste français, des "diminutifs de prononciation". 10 En effet, c'est l'idée de 'petitesse' qui est impliquée dans les formes roumaines ['fatse], ['p(l)indze], etc. Il est vrai que l'altération phonétique discutée atteint, à coté des substantifs (par exemple, ['mindze] ou ['minzte], pour minge ['mind3e] 'balle'), aussi les verbes et les autres catégories de mots — ce qui résulte, d'ailleurs, des exemples que nous avons donnés ci-dessus ; mais 'diminution' et, disons, 'verbe' sont des notions qui s'associent facilement dans le domaine de la LE. 11 Le type de syllabe le plus fréquent dans les mots spécifiques de la LE est 'consonne + voyelle (ou diphtongue)' (environ 90 % de toutes les syllabes rencontrées dans nos matériaux). Les mots que nous avons recueillis ont, pour la plupart, deux syllabes, dont la première est, en général, accentuée. Parmi les monosyllabes on 8

Le changement analogue de [J], [3] en [s], [z] est plus rare dans la LE correspondant aux parlers envisagés ici ; voir cependant ce qui est dit ci-dessous sur les parlers caractérisés par la transformation en fricatives des consonnes [tj] et [«I3]. • N. S. Troubetzkoy, Principes de phonologie (Paris, 1964), 20. 10 A. Meillet, Bulletin de la Société de linguistique de Paris, XXII (1920-21), 167. M. Roman Jakobson, op. cit., 544, parle, lui aussi, d'un "typically diminutive sound symbolism", à propos du rus. njanja 'nanny', opposé à mama. 11 Voir ci-dessous, p. 139.

136

ANDREI AVRAM

rencontre surtout des mots à caractère interjectionnel, employés soit comme saluts (pa! 4au revoir!', tai! 'au revoir!'), soit comme moyens par lesquels on demande à l'enfant d'exécuter une action (ai! [?aj] 'viens, allons, regarde, écoute!') ou on lui interdit de toucher à quelque chose (voir ci-dessus âr§, hîr$; on pourrait ajouter ici cîh, qui remplit la même fonction que âr$, mais qui a aussi — comme nous l'avons déjà vu — une valeur adjectivale). On constate donc que les mots à caractère interjectionnel ou ayant la valeur d'un impératif sont, en général, les plus courts (ce qu'on peut observer aussi dans la LN); beaucoup d'entre eux s'éloignent, par leurs syllabes fermées, des structures simples, 'primitives',12 qui caractérisent la LE et, à la fois, la langue des enfants dans sa première étape de développement. En dehors des mots du type papa '(la) nourriture', 13 faisant partie de ce qu'on a nommé les 'international baby words',14 il y a quelques formes comme pa-pal, na-nal qui, à proprement parler, sont des variantes des formes simples pal 'au revoir!',15 na! 'tiens!' (lorsqu'on menace l'enfant ou on lui applique, plus ou moins sérieusement, une correction). Dans pa-pa !, na-na !, l'accent principal (sur la deuxième syllabe) peut être accompagné d'un accent secondaire sur la syllabe initiale, mais, très souvent, ces formes composées n'ont qu'un seul accent: [pa'pa], [na'na]. On arrive, de cette façon, à l'utilisation des contrastes accentuels (qui autrement ne sont pas employés pour différencier des mots): [pa'pa] 'au revoir!': ['papa] '(la) nourriture'; [na'na] 'tiens!': ['nana] '(la) soeur aînée, (la) tante, (la) marraine'.16 Quelques autres problèmes de phonétique, concernant surtout la structure des mots, seront discutés dans le chapitre consacré au lexique. 3. GRAMMAIRE

Un trait caractéristique de la grammaire de la LE, par rapport à la LN, consiste dans l'emploi, surtout au singulier,17 de la troisième personne à la place des deux autres. la

Cf. Bertil Malmberg, Proceedings of the Ninth International Congress of Linguists, Cambridge, Mass., August 27-31, 1962 (La Haye, 1964), 81-83. 13 Pour la double valeur, déterminée et indéterminée, de la forme avec article défini papa 0a forme sans article est papa), voir le chapitre consacré à la grammaire. 11 Robert Austerlitz, loc. cit., 276. 15 Ce mot est employé aussi dans la LN (cf. Iorgu Iordan, op. cit., 49). Notons cependant que, tandis que dans le pa 'normal' l'intonation est descendante et le mot a une durée qui ne dépasse pas la durée de da 'oui', par exemple — dans le pa appartenant à la LE l'intonation est nettement ascendante et la voyelle est, d'habitude, allongée; ajoutons encore que dans la LE on emploie, le plus souvent, la forme avec réduplication. 16 La réduplication est très fréquente chez les noms d'animaux (le point de départ en est, d'habitude, une onomatopée): ham-ham ou hau-hau 'chien', oac-oac [,wak'wak] 'grenouille', etc. Même si l'on laisse de côté les formations de ce genre il nous reste, parmi les mots que nous avons recueillis, une vingtaine du type marna, papa, ce qui confirme l'opinion selon laquelle la réduplication est "a favorite device in nursery terms" (Roman Jakobson, op. cit., 542). 17 Au pluriel on emploie fréquemment la première personne des verbes (et des pronoms), dans des phrases comme Mergem la culcare, littéral. 'Nous allons nous coucher', mais ayant, en réalité, le sens 'Tu vas te coucher' ou 'Va te coucher!'. Cf. Academia Republicii Populare Romane, Gramatica limbii române, I, édifia a Il-a, revàzutâ çi adâugitâ (Bucarest, 1963), 152, 243.

DE LA LANGUE QU'ON PARLE AUX ENFANTS

137

En général, on évite les pronoms eu 'je, moi', tu 'tu, toi', le sujet de la proposition étant exprimé par un substantif ou par un pronom de la troisième personne. On dit, par exemple, Vine mama (tata, etc.) îndatâ 'Maman (papa, etc.) vient ( = va venir) tout de suite', au lieu de Vin îndatâ '(Je) viens tout de suite'. De même, on dit A fost cuminte bâiatul? 'Le garçon a-t-il été sage?', au lieu de Aifost cuminte? 'As-tu été sage?'.18 La deuxième personne se maintient mieux à l'impératif; toutefois, la deuxième personne de l'impératif est souvent remplacée par la troisième personne de l'indicatif (accompagnée, obligatoirement, d'un substantif): Merge la cuîcare fetija 'La fillette va se coucher', au lieu de Mergi la culcare! 'Va te coucher!'. Les cas de ce genre démontrent non seulement la force de la tendance de l'adulte à employer la troisième personne, mais aussi son souci d'atténuer l'ordre qu'il donne à l'enfant; le caractère affectif de la construction est évident.19 Une autre particularité grammaticale de la LE consiste dans l'emploi fréquent de la forme de nominatif-accusatif dans des contextes où les règles de la LN exigent la forme de génitif: carte mama 'livre maman', au lieu de cartea mamei 'le livre de maman'. Un peu différent est le cas du datif analytique avec la préposition la (fréquent, d'ailleurs, dans la langue populaire), qui remplace le datif synthétique, normal dans la langue littéraire: Spune la mama...!, au lieu de Spune mamei... ! 'Dis à maman...!'. On pourrait supposer que l'adulte évite les désinences du génitif-datif afin que le mot garde son aspect 'fondamental', mieux connu par l'enfant. Il faut cependant tenir compte du fait que, très souvent, on évite aussi l'article indéfini, dont l'emploi n'altérerait aucunement l'aspect des mots qu'il précède; par exemple: Dâ mama creion (au lieu de un creion) 'Maman va te donner [un] crayon'. D'autre part, il faut mentionner que les formes avec l'article défini -a des substantifs féminins ou des masculins comme tatâ, nene sont, souvent, préférées là où la grammaire de la LN exigerait les formes sans article: altâpapa '(un) autre plat', un nenea 'un monsieur'. On voit donc que le phénomène est plus complexe que ne le laissaient voir les cas du type carte mama. Il s'agit, en effet, d'une tendance générale de l'adulte à employer une grammaire considérée (à raison ou à tort) plus simple; c'est ce qui explique aussi la suppression de la copule ou du verbe a face 'faire', employé dans beaucoup d'expressions : Bâiatul cuminte 'Le garçon [est] sage', Copilul nani (ou Nani copiluï) 'L'enfant [fait] dodo' ou — avec valeur d'impératif — 'Dors!'. A propos de l'exemple Copilul nani, ajoutons que la construction complète, avec 'faire' est, elle aussi, fréquente (sinon même plus fréquente); la différence entre les deux formules consiste dans le fait que la première (la plus courte) est plus 'babyish' que l'autre, sans représenter, cependant, le degré extrême de 'babyishness' de tous les 18

II s'agit ici d'un phénomène qui est loin d'être spécifique pour la LE roumaine; cf., par exemple, les observations de Ashok R. Kelkar, loc. cit., 47, sur la LE marathe. 19 Pour l'emploi, avec valeur affective, de la troisième personne, à la place des deux autres, voir S. Puçcariu, Dacoromania (Cluj), V (1927-1928), 748, note 2, et surtout Iorgu Iordan, op. cit., 126127. Dans les faits qui viennent d'être cités on pourrait voir aussi le résultat d'une tendance générale de l'adulte à restreindre — lorsqu'il parle aux enfants — le nombre des formes verbales employées.

138

ANDREI A VRAM

points de vue (phonétique, grammatical, lexical).20 En partant de la phrase Dormi (copile) ! 'Dors (mon enfant)!', qui appartient à la LN, on passe dans le domaine de la LE par l'application de trois procédés, dont l'un est lexical (le remplacement de a dormi 'dormir' par l'expression a face nani) et les deux autres sont grammaticaux (le remplacement de l'impératif par l'indicatif et de la deuxième personne par la troisième) : Copilulface nani (ou Face nani copiluï). Dans Copilul nani (Nani copilul) on supprime le verbe et on obtient ainsi quelque chose de 'plus enfantin'. Enfin, il est possible de faire appel à un procédé phonétique pour obtenir le maximum de 'babyishness': dans Pilul nani! ['pilu 'nani] (cf. tifa ['titsa] = fetifa 'la fillette') le substantif provenant de la LN est modifié — par la suppression de la syllabe initiale — de telle manière qu'il reçoive une structure plus proche de celle qu'on rencontre le plus souvent chez les mots spécifiques de la LE; nous allons revenir sur cet aspect du problème dans le chapitre suivant. 4. LEXIQUE

Nous présentons, dans ce qui suit, quelques remarques sur les mots de la LE roumaine, sans en donner la liste complète. Commençons par une observation sur les parties du discours auxquelles appartiennent les éléments de ce lexique spécial : la chose la plus frappante est l'absence presque totale des verbes; seul a papa (dérivé de papà 'nourriture') = LN a mînca 'manger' est d'un emploi général. En s'adressant aux enfants, l'adulte est contraint, dans la grande majorité des cas, de recourrir aux verbes de la LN. Quand même, certaines actions s'expriment à l'aide des mots spécifiques de la LE (substantifs, adjectifs, adverbes, interjections), employés soit seuls (par exemple, nani peut signifier 'dors!', 'tu vas dormir', etc.), soit précédés du verbe a face 'faire': a face nani, a face ni-ni (ou titi, gigea, etc.) 'caresser, cajoler', a face pipi, etc. Les expressions a da na-na (à côté de a face na-na) 'battre' et a da un pup( ic) 'baiser', avec le verbe a da 'donner', doivent être considérées — du point de vue de leur formation — moins spécifiques pour la LE, étant donné qu'elles sont, pour ainsi dire, des 'calques' d'après des expressions appartenant à la LN : a da bâtaie 'battre', a da o sârutare 'baiser'. Les sphères sémantiques auxquelles appartiennent les mots spécifiques de la LE sont — on le sait — peu nombreuses; nous n'insistons pas sur ce point. Pour le reste, c'est la LN qui fournit le lexique nécessaire à la communication. Cependant beaucoup de mots 'empruntés' à la LN sont soumis à des modifications qui leur donnent un 'air enfantin' ; dans ce but, on utilise soit la 'simplification' phonétique, soit la dérivation à l'aide des suffixes diminutifs. Le premier cas est représenté par des mots comme bobo, don, coco, mumos,21 20

Voir, sur les degrés de 'babyishness', Ashok R. Kelkar, loc. cit., 41 ; cf. Roman Jakobson, op. cit., 331. " Cf. la même substitution de la syllabe initiale dans *mumuçat (< frumuçat), d'où le roum. dial. (dans la LN) muçat 'beau' (A. Graur, Bulletin linguistique, Bucarest, V, 1937, 70-71).

DE LA LANGUE QU'ON PARLE AUX ENFANTS

139

n

pisu (pisi), pupa, tita, pour bomboanâ, 'bonbon', respectivement, picior 'pied', cocos 'coq', frumos 'beau', pisicâ 'chat(te)', pâpu§â 'poupée', atîta 'tant'. Les formes auxquelles on aboutit par la suppression de certains sons et par le changement d'accent sont, pour la plupart, dissyllabiques et paroxytones ; en éliminant ou en modifiant certaines syllabes, on réalise, parfois, aussi l'identité du vocalisme (pisi), du consonantisme (pupa) ou même l'identité complète des deux syllabes de la forme nouvelle (bobo). La modification par dérivation est très bien représentée dans nos matériaux; voici quelques exemples : apâ 'eau', pîine 'pain', somn 'sommeil' deviennent apicâ ou apifâ, pîinicâ ou pîini{â, somnel ou somnicâ23 (dans ce dernier cas la forme féminine du suffixe, -icâ, est ajoutée à un substantif neutre ; s'explique-t-elle par l'expressivité plus grande de la voyelle [i]?). Les mots qui font partie du vocabulaire spécial de la LE et qui ont déjà des traits 'enfantins' peuvent devenir encore 'plus enfantins' par l'adjonction d'un suffixe diminutif : papa devient pûpicâ, pâpitâ ou même pâpificâ. A noter que les suffixes diminutifs s'attachent souvent à des racines verbales: à coté de pâpicâ 'nourriture', il existe un autre pâpicâ, dérivé du verbe a pâpa 'manger': Pâpicâ bâiatul! 'Mange! (en s'adressant à un garçon)', littéral, 'le garçon mange + suffixe diminutif'. Certaines formations de ce type (dérivées des verbes) sont employées aussi lorsqu'on s'adresse aux adultes (avec une valeur affective), mais elles apparaissent d'habitude dans la LE, ainsi que l'a noté B. Cazacu,24 qui cite des parallèles comme le russe IIoùdeM cnamm 'Allons nous coucher!' (LE cnamm = LN cnamb) ou l'allem. ich willele (= ich w///).25 La dérivation diminutivale impliquant, en règle générale, un allongement des mots et un changement d'accent, les formes obtenues par le procédé examiné ci-dessus s'éloignent, très souvent, de la structure typique des mots appartenant à la LE. On élimine, parfois, ce désavantage par une 'amputation' phonétique du même type que celle que nous avons déjà rencontrée dans tîtaOJIOrHH'

npexme Bcero Ha

MaTepHaJie

KOHKpeTHbix e c T e c T B e H H b i x H3biKOB. B e 3 3 T o r o y n p a a c H e H H a B o G j i a c T H n o p o x m a i o m e r o CHHTaKCHca j i n u i b 6 y a y T yBejiHHHBaTb HHCJIO (yace H T e n e p b

3HaiHTejibHoe)

KaJiaMGypOB H JIHHTBHCTHHeCKHX aHeKflOTOB. B CB83H c o CKa3aHHbiM BaacHo o 6 p a T H T b BHHMaHHe e m e H H a c j i e a y i o m n e MOMCHTM. ECJIH JieKOHHecKHM (J)yH^aMeHTOM ^ j i a M H o r o o 6 p a 3 H b i x p e a j i H 3 a u H H T a r M e M b i c j i y x c a T u e j i b i e G o J i b i n n e TpaflHUHOHHbie K j i a c c b i CJIOB, n a c r a p e n H , TO AJia K a a c a o i í

144

O. C. AXMAHOBA

rnnepTarMeMW

(hjih

fljia KaKHx-To C T p o r o ycraHOBjieHHbix

bhaob

ranepTarMeM)

JieKCHieCKOH OCHOBOHflOJiaCHMflBJMTbCH,ÜO-BH,ZtfIMOMy, r0pa3£0 GoJiee OrpaHHHeH-

cjiob, ocoSbie h cneuK(J)HHecKne ; y i a Kaac^oro JBbiKa H a G o p b i KoHKpeThmx aitJiorHnepjieKC. BMecTe c TeM, noHaTHe ranepjieKceMbi n o T p e ó y e T c a pacnpoCTpaHHTb TaKace h H a Te C0B0KynH0CTH l pa3HOHMeHHbix' 'ajiJiorraepjieKc' ('chhohhHbie Kjiaccbi

m o b ' ) , K 0 T 0 p w e , TeopeTHHecKH p a c c y a c ^ a a , M o r y T 3aMeHaTb flpyr « p y r a n p a n o p o a c fleHHH

BbICKa3bIBaHHÍÍ, T X . ,

eflHHHií, Ha

kotopmx

IipeflCTaBJiaTb C 0 6 0 Ü TEKHe psflbl H

MoacHo 6 a 3 H p o B a T b

Hnacecjieflyiomaa nacTb p a 6 o T b i

oGcjie^oBaHHH oahoh

hbhtch

ranepTarMeMHyio nonbiTKoñ

COBOKyiIHOCTH

MexaHHKy.

bo3mohcho

6ojiee a e T a j i b H o r o

pa3HOBH,ziHOCTH r n n e p j i e K c e M H o r o c o o T H o m e H H a eAHHHu;

b

COBpeMeHHOM p y c c K O M «3buce.

rjiaroJibHaa

ranepjieKceMa,

CHHTaKCHca, npeflCTaBJiaeT

ochoboh nopoamaiomero a3MKe coBOKynHocxb o6pa30BaHHH,

aBjiaiomaaca ecTecTBeHHoií

co6oh b

pyccKOM

MOp(})OHOJIOrHHeCKH H CeMaHTHHeCKH HaCTOJlbKO CJIOaCHblX, HTO HHOrAa HeBOJIbHO B03HHKaeT cjieAyioinHH Bonpoc:

hto

ecJiH 6 h aBTopw

nopoacflaiomeH rpaMMaTHKH HMejia nepBOHanajibHo

h ctopohhhkh ochobhmx H/iefí acjio He c aHTJiHHCKHM, a c

B03HHKJIH JIH 6bl y HHX Cpa3y »Ce OHeHb Cepbe3Hbie COMHeHHa? jih ecTecTBeHHbifi HejioBeHecKHií a3biK mohcho ce6e npeflCTaBHTb KaK coBOKynHocTb TpaHccJjopMaiiHH HeKOTopwx aflepHbix Mo/jejieñ hjih CTpyKTyp, nocjiefloBaTejibHO h 3aK0H0MepH0 CBa3aHHbix flpyr c ApyroM ccphhhbimh othopyCCKHM H3bIKOM, He ^eñcTBHTejibHO

HieHHaMH, npHHeM TaKHMH, KOTopbie noaaaioTca cTporoMy h nocjiefloBaTejibHOMy onHcaHHio? H e

bo3hhkjio jih

6 h cpa3y coMHeHHe oTHocHTejibHo, BooGme, ocyme-

CTBHMOCTH B033peHHa Ha H3bIK, KaK Ha CpaBHHTeJIbHO HeCJIOaCHblH MeXaHH3M, fljia ynpaBJieHHa KOTOpblM flOCTaTOHHO JIHHIb C(J)OpMyjIHpOBaTb H no B03M0HCH0CTH (J)0pMajiH30BaTb HeKOTopyio C0B0KynH0CTb 6ojiee IlpH

3tom

HeoóxoflHMo o6paTHTb BHHMaHHe eme

h

hjih

MeHee

cjiohchhx

npaBHJi?

Ha cjieayiomee oócToaTejibCTBo.

üo-BHflHMOMy, 'nopoayjaiomaa rpaMMaraKa' CHHTaeT TaKHe npaBHJia yace 3aaaHHbiMH, H3BecTHbiMH. OHa He MynaeT ceóa BonpocoM o t o m , KaK y 3 H a T b Te peajibHbie 4>aKTbi, KOTOpbie npeAJiaraeTca onncaTb. X o T a

3to skchjihijhtho h He

pa3bacHa-

eTca, oHa Mojinajimo npeanojiaraeT, h t o Bce a3biKH o6a3aTejn>Ho KOMy t o

yace

nOJIHOCTbK) H HCHepnblBaiOme H3BeCTHbI. OflHoií H3 cjioacHeiíiiiHX

ctopoh

rjiaroJibHoñ

rHnepneKceMbi

b

pyccKOM,

KaK,

no-BHflHMOMy, h b flpyrax cjiaBHHCKnx a3biKax, aBJiaeTca nepenjieTeHHe pa3JiHHHbix epeflCTB BbipaaceHHa,

BHfla,

KOTopbiñ

ocHOBaH Ha p a 3 H o o 6 p a 3 H b i x

coneTaHHax

npncTüBOK h B H « o o 6 p a 3 y i o m H X h h ( | ) h k c o b . HeKOTOpbie acneKTbi 3 t o h n p o 6 j i e M b i MoacHO noacHHTb H a c j i e a y i o m H X npHMepax: ( 1 ) T j i a r o j i b i c n p H C T a B K O H ' 3 a ' — TaKHe KaK 3a6umb, 3azoeopumb, zamb, 3adoji6umb, 3aoKumb, 3au¿pamb, 3üKmamb,

3a.Konmumb, 3anemamb,

3adep3apa6o-

mamb, 3axodumb, h AP- Y 3 t h x r j i a r o J i o B npoHCxo/jHT cBoeoópa3Hoe pa3flBoeHHe, KQTopoe AOBOJibHQ j i e r a o npocjieacHBaeTca He

tojibko b m i a H e omohhmhh

npHCTaBOK,

145

HEKOTOPblE OCOEEHHOCTH rJIArOJIbHOH rHIIEPJIEKCEMH T.e.,

fle3HrHaTHBHoñ

'3a-',

HO H B

HecoBMecTHMocTH p a 3 H u x c j i y i a e B ynoTpeGjieHHH npncTaBKH

njiaHe

pa3a,

3auzpaA TpaHcno3HUHH B HecoBepmeHHbiH B H A He

ecezda 3aKanueaem Ha Mope, 3aneM 6MAO

saxanueamb

noanaexcH

coBceM.

peóema

KaK 6 y f l T o BO3MO5KHH, HO, KOHCHHO, S y a y T 3 B y i a T b HecpaBHeHHo S o j i e e

—

HCKyccTBeHHo, neM, H a n p H M e p ,

3a6ueamb ceau. B M e c T e c T e M E/jHHCTBeHHoií B O 3 -

M o a c H o ñ TpaHccJjopMauneíí a n a , H a n p H M e p , jieKCHKO-4)pa3eojiorHiecKaa:

3aKanamb HOZOÜ G y f l e T TpaHC(j)opMauH0

HanpHMep,

He nado 6UAO, 3aneM 6buo

HaiHHaTb

HOZOÜ.

Kauamb

He nado 6MAO 3aMambieamb wapcfi eoKpyz ZOAOBU — (})pa3a 6e3ycJioBHo OTMeneHHaa, a 3aKanmeamb n03T0My

mm0A0K

—

HeT. JlaMna

3üKonmuAa —

a BepHee, B o o 6 m e , n p o c T o

Konmumb,

KonmuAu. T a K » c e OH 3apa6omaA (3apa6ambieaA

pyÓAeü e Mecny, HO Momop 3apa6omaA (nanunaA paGomamb)

(2)

nanana

neAb3H óbiAo donycKamb, nmodbi AaMna, Kepocuma H n p . HOHUHOAU Konmumb,

Tjiarojibi

c

npHCTaBKoñ

'Ha-' —

nazoeopumb, naMcamb, naümu, Hammamb,

TaKHe KaK

no) 500

c caMOZO ympa.

nañumb, naeecmu,

Haeexmb,

HaKOAomb, naAemamb, nanecmu, H ap. Ha

n e p B b i ñ B 3 r j i a a BnoJiHe c o n o c T a B H M b i c p a c c M O T p e H H H M H B b i m e r j i a r o ü a M H Ha ' 3 a - ' , T . K . , OHH Toace n o c j i e a o B a T e j i b H o p a c m e m i a i o T c a — c

BHTeKaiomHMH

HanpHMep

CJIOKHOCTHMH

naóumb (móueamb)

BAIOTCA r o p a 3 « o KÜK

oTCiofla

6 o j i e e cjioacHbiMH. H a n p H M e p ,

BCjieacTBHe

naóueamb CM0ABK0 aM6HBajieHTHocTH

nynejia'? HJIH

3amuKoe

( 1 ) C B e p x y , H ( 2 ) B KOJiHHecTBe

njiaHe

oópyn Ha óoixy.

u3HocumcR luyóoHbKa cmapax

ne nado óbiAo

B

cepHiÍHbix

Ho

OTHOUICHHH.

B u e j i o M OTHOUICHHH

0Ka3bi-

KaK 6 b i T b B TaKOM c j i y n a e :

HOSUX

uaSbeml

MOXCHO

—

Tax

Kmo

JIH C K A 3 A T B :

noeux 3aüifeel I I o B H f l H M O M y , HeT, n p n n e M TaKace 3TOH

pa3bi:

naueamb

e meMHome ezo naee3AU Ha cmoAÓ. —

3ammoe 3aueM 6HAO,

3HaHHT

'aejiaTb

Kmo no3eoAUA

146

O. C. AXMAHOBA H o uaeommb

Haeo3umb ezo na cmojiól 3 t o ecTecTBeHHo?

flajiee,

mohcho jih uazoeopumb

cmojibxo

nodapnoe,

c mpu Kopoóa

nenyxul

( h nazoeapueamb).

3aneM 6bvio nazoeapueamb

H a KHomcy moncho naMcamb h naotcuMamb.

nenyxul

h nanambieamb.

Lfeembi

jih

MoacHo jih ecezda

nymb, h T.n. MoacHo

HOKambieamb

noMnozy

jianwul

opden Ha AaifKan h T.n. mohcho h HaKOAomb h HaKaAbieamb.

na uuixny,

MoacHo

Ho

Ho

cmoAbKO

H o mohcho jih CKa3aTb, h t o b

óbuibie zodbi OHU HawciiMQAU no cmo eedep eunal JJopozy, cambiü HaKamamb

— Tan ace jih

zocmunifee

MOMCHO uazoeopumb

nAacmumu

CKa3aTh,

HanpHMep,

3, e m e pa3, ( 2 ) Bce, ckojimco ecTb, M H o r o — T.e., cooTBeTCTBeHHo: nepeóenumb

h T.n., h

ex3amb ecex óynmoeufUKoe,

h T.n. hjih ece

nomoAOK,

cmenu,

ece KpecAa. IJepeGo/iemb

ÓOACSHHMU. üepeeH3amb

6oAe3Hbw h eceMu

mnMceAoü

cmeny,

cmyA, KpecAO h ece cmyAbn,

nomoAKU. üepeóumb

eepeeKoü,

ece

zpunnoM, h

peMHeM

nepe-

c cooTBeTCTByiomHMH MHoroHHCJieHHbiMH

BapHaHTaMH H pa3HOBHflHOCTHMH CeMaHTHHeCKHX CB5!3eH, 06yCJTC>BJIHBaeMbIX KOHKpeTHOH JieKCHKOH pa3Hbix TJiarOJIOB. H o KaK 3 t o Bce BHrjia^HT b paccMaTpHBaeMOM 3nccb njiaHe, T.e., c t o h k h 3peHHH TpaHC^opMauHH, k o t o p m m M o r y T 6biTb noABeprayTbi coOTBeTCTByiomHe HCxoflHbie 06pa30BaHHa? HanpHMep, M0acH0 CKa3aTb: OH 3anoeo nepeóeAueaA

B c e KOMHamu, ecepaMbñ

nepeóeAueaA

OH COM

paMy;

ITo-BH/iHMOMy, MoacHo, 0flHaK0 6HceMaHTH3M

CHHMaeTC»? T o ace o t h o c h t c h k nepeóueamb,

nepeñajiueamb

(ecjin TOJibKO nocjieflHHH

He oKaaceTca cjihiiikom 'ncKyccTBeHHbiM': HanpHMep, OH NECKOABKO pa3 nepeóaAueaA zpunnoM, demu nepeóanueadu B

cjiynae ace, HanpHMep,

nepeen3bieaA

eceMu 6oAe3HHMU y otee e panneM ueModami

eepeexoü,

3a 3UMy eo3pacmel). h

peMHHMU

ecezda

— He 6yaeT jih HanSojiee ecTecTBeHHoií a c c o i m a i m e ñ

nepeeH3bieaA ecex 6yumoeu¡uKoe

fljia B T o p o r o npHMepa cobccm a p y r o e 3HaneHHe, a hmchho — lnepeBH3MBaJi paHbi', 'ocTaHaBJiHBan KpoBOTeneHHe'? C KeM-Au6o

nepezoeopumb,

HUMU nepezoeapueamb nepeecmb

SpycoK,

ecezo

ne ycneeuib

peóeuoK

—

nepeecmb

euiuem.

npaBHJibHo H ecTecTBeHHo. MoacHo nepeuzpamb (He) nepeuzpbieamb. nAacmumu,

nepezoeopumb.

HenpHeMJieMo coBepmeHHo. PoKaenuna

Ho

TaKoe

B

H o 3aieM

6UAO C

u KucAoma

MoryT

O6OHX c j i y i a a x

(3aH0B0) napmuw,

nepeedamb

HO TaKace ee H

npefljioaceHHe KaK OH AWÓUA nepeuzpbieamb

no-BHflHMOMy, 6 y a e T BocnpHHHTo TOJH>KO KaK *3aH0B0', 'cHOBa'?

ace, no-BHflHMOMy, 6yn,eT oTHocHTbca H K TaKHM (J>pa3aM, KaK nepemmbieamb 6eAbe, nepenAaeAnmb

cmoAbKo

ece To ece

(MHOTO) uyzyna, H T.n.

(4) T j i a r o j i b i c npHCTaBKoií

' c - ' . 3/iecb s b h o 0M0HHHHbie npiicTaBKH c I h

c I I , npHflaiomHe cooTBeTCTByiomHM r j i a r o j i a M npaMo npoTHBonojioacHbie 3HaneHHH, KaK 6 y f l T o 6 w He nopoacaaioT BHfloBbix ocjioacHeHHH, noflo6Hbix nepenHCJieHHbiM

HEKOTOPblE OCOEEHHOCTH IMIArOJIbHOH rHIIEPJIEKCEMBI

147

Bwme. TaK HanpHMep cóumb wamy c zoAoebi H cóumb ÓOCKU SMecme 0AHHAK0B0 TpaHCíJjopMHpyioTCH B cóueamb. T o ace OTHOCHTCH K ceecmu, cee3mu, H T.n. OflHaKo h 3 f l e c b B03HHKaK)T 6 o j i e e c j i o a c H b i e c j i y i a H . TaK H a n p H M e p , y n i a r o J i o B CKdAbieamb, cKjiadbieamb, mecmu, cpaSomambcn o T H o m e H H a r o p a 3 A O M e H e e a c H b i e ; HX M o a c H o PACCMATPHBATB KAK THNHHHBIIÍ CJIYNAH TOH H C B U S C H C I I H O C T I I (JiaKTOB,

o

KOTOPOÍÍ 6HJIO

HccJiefloBaHHH.

CKa3aHo

Bbirne.

HeacHbiM

B e f l b e c j i n n o c j i e f l H e ñ HHCTaHUHeñ,

ocraeTca

OCHOBOÍÍ B c e r o

CSMHX

H caM

MeTofl

a3biK03HaHHa

0KA3BIBAETCH, B KOHCHHOM c n e T e , ' a K C H O M a p e a K U H H HHopMaHTa', TO B p a a JIH 6 y a e T n p e y B e n H H e H H e M C K a 3 a T b , HTO H c c j i e A O B a H H e B STOH o 6 j i a c r a p y c c K o r o a 3 b i K a

eme

TOJIbKO HaHHHaeTCH. ( 5 ) r j i a r o j i b i c n p H C T a B K O H ' n p o - ' : npoeodumb, npoeo3umb, npozjiadumb, npoKOJiomb, npoAemamb, npoxodumb, nponocumb, nponaxamb. P a c m e n j i e H H e 3 f l e c b B H a H Ó o J i e e o 6 m e M BIME MONCHO o n p e a e j i H T b KaK ( 1 ) iepe3, MHMO, B o r p a m m e H H o ñ n a c T H HJIH y n a c T K e , H ( 2 ) a o j i r o ,

flojiroe

B p e M a , u e j i b r a a e H b , H T . n . , T . e . , COOTBCT-

npoeo3umb uepe3 zpanuify, B Kop3une H ifeAbiü dem. IIpoAemamb MUMO eopoda,pndoM H eao &cu3Hb UAU nnmb Aem. IJpoxodumb MUMO H eecb dem. ílponocumb dupy H dea ce3ona. Tlponaxamb 6opo3dy H YIOAÓHH H T.A. B m i a H e ' c e p H i Í H b i x OTHONIEHHÑ' K a p T H H a 3 f l e c b o K a 3 M B a e T c a B e c b M a c i i o a c H o ñ . T a K HanpHMep npoeo3umb I a B j i a e T c a H e c o B e p m e H H b i M BHAOM K n p o B e 3 T H , B TO B p e M a KaK B npoeo3umb ifeAbiü CTBEHHO:

dem

n p H C T a B K a ' n p o - ' o H e m > OTHCTJIHBO M o p 4 > o J i o r H 3 H p y e T c a KaK HMCHHO JICKCH-

necKaa

MopcJieMa. B p e 3 y j i b T a T e H e c o B e p m e H H b i ñ

BHA .zjjia TaKHX

raarojioB

KaK

npoeo3umb I I , npoxodumb I I ( ' i í e j i b i ñ fleHb'), npozopemb I I ( ' a e j i y i o H O H b ' ) , nponaxamb I I , H T . n . 0KA3BIBAETCA HCKJHONEHHBIM, a TPAHC4>OPMAUHA n o T p e 6 y e T cjioaaHbix o n n c a T e J i b H b i x BbipaaceHHH.

(6) r j i a r o j i w c npHCTaBKOH ' n o - ' . H 3aecb CBoeo6pa3He noeedenun Toro HJIH apyroro rjiarojia npH nonbiTKax BH/IOBOH TpaHC(J)opMauHH 0Ka3biBaeTca B 3 H a i H Tejn>HOH CTeneHH o6yc.JioBJieHHbiM TeM, HacK0JibK0 OTHCTJIHBO flaHHaa npHCTaBKa MopojiorH3yeTca HMCHHO KaK caMOCToaTejn>Haa JieKCHHecKaa Mop4>eMa. M M e t o T ca B BHay TaKHe rjiarojibi, KaK noeodumb, noxoáumb, nozonnmb, nonocumb, H T.N. PacmenjiaioTca KaK (1) no+eooTBeTCTByiomHH rjiaroji, co 3HaneHHeM 'HeKOTopoe BpeMa', H ( 2 ) c He Mop o j i o r H 3 a n H H 3 j i e M e H T 0 B ,

cocTaBJiaiomHX

A a H H o e c j i o » H o e 0 6 p a 3 0 B a H H e . 0 6 p a 3 0 B a H H a STH c j i e a y i o m n e : áoóueambcn, óueambcn;

áoóupambcn,

saóupambcx;

nepeóupambCH;

doeodumbCH,

nepe-

nepeeoóumbcn;

148

O. C. AXMAHOBA

eMeiuambCJi,

CMeuiambCH;

cmebi,

ycnexa,

pu6a

moxcho ToiibKo doóumb. nepeóueambcx

nepeóumbcn: zopbi,

óepeza

doóueaemcñ

Mootcem

t o oHa He MoaceT doóumbcn

na 6epezy,

nepeóueaemcn

6bimb

nepe6nTa

h j i h omdeAKa.

h o doKAadvma

oóoüiquxoM,

MejioBeK MoaceT doóupambcn

nepeÓHTb.

h t . a - , h cooTBeTCTBeHHo do6pambcn

Mopx,

nodKAadica

— ee

h o MoxceT jih

oóoüufKaMU,

H yac k o h c h h o h h t o t h h a p y r a a He M o r y T

npedcedameneMl

Meóenb

doóupambcn

nepenucbieamb-

TaK HanpHMep, doóueambcn

MHZKÜH Meóe.ib

MOMcem moAbKo

cedame/ib

nepeóupambcn,

saAueambcn;

moncho yem, Mecma, h t . ^ - , h n p a 6jiaronpH«THbix oScToaiejibCTBax h x moncho doGumbcn.

E c j i h ace 8bi/i08AeHHan opamop

3muMambCH,

oóxodumbcn.

CR\ Haxodumbcn,

ftzodu

nped-

do

eepiuumi

T y a a . K mambio

TaKace M o r y T doóupambcn

HOZO eeca, n T.n. O/iHaKo CKa3aTb, h t o nodKAaÓKa doópanacb

do

MoaceT onpedeAenh hto

e HOBOM Mazarme

HeB03M05KH0. Mo»cho TOJibKO doópamb noÓKAaÓKy, K.tyÓHUKy. To-xce o t h o c h t c h k t e k h m napaM, KaK móupambcn na 3a6op, eepuimy, Kpuuiy, b otjihhhc o t moeap, xneó, mepcmb 3a6upawmcfi na cKAade. fíoeodumbCH moncho ópamoM, cecmpoü, podcmeeHHUKOM. fíopoza, tuocce, napK M o r y T doeodumbCH do Mopn, do MazucmpaAU, h np., npaneM b o 6 o h x cjiynaax doópaAacb

KAyóhUKQ

do

noAHoü

Kop3UHbi

BHflOBhie MOflH^HKaiXHH OKa3bIBaK>TCH HeorpaHHHeHHblMH OHa aoBoflHTca MHe cecmpoü;

ópamoM,

dopozy h np. MOaCHO doeecmu

(doeodumb)

dopoza

doeodumcn

OH doeoduttlCH

MHe

h T.n. H o ecjin

do Mopn,

h T.fl., t o n o f l o 6 H a a

do Mopn, MazucmpaAu,

TpaHc4>opMauHfl flJiH n e p B o r o H3 3HaneHHH s t o t o r j i a r o n a coBepmeHHo HeB03M0acHa. B c j i y i a a x Bpo^e nepeeodumbcn

na dpyzyto

doADKHocmb,

e dpyzoü

leTKaa

zopod

Mop4>ojiorH3aiíHH flonycKaeT pa3Hoo6pa3Hbie TpaHCfJjopMauHH. B c j i y i a a x ace Bpo,ae ece

denbzu,

dpo3dbi,

puóa

nepeeeAacb

nocTpoeHHe

CHHTeTHHecKHM. E c j i h ueAoeeK BMeiuueaemcH

0Ka3biBaeTca r o p a 3 ^ o

e uyoKue

B M e u i a T b c a e HUX, s t o He 3HaHHT, h t o e r o m o x c h o eMeiuamb MÜCAO h

cTopoHM caxap, eMeiuamb

b Hero, h o

zocmuHHyw,

e

flpyrae

MoryT j i h o h h

HeoGbmnoü

MejioBeK; cnupm

oócmaHoene,

b Hero eMeuiambcxl e

MoaceT CMeiuambCH c eodoü

óoAbuioü

h x mohcho

e mecmo, CMeiuambch,

aydumopuu

h pa3Hue

c apyroñ

epa3zoeop;

HHrpeflHeHTbi eMeutueawmcM

3anaxu

6ojiee

h MoaceT

deAa, e pa3zoeop

eoüdx

MoaceT

e

Kaacflbiñ

cMeuiambcn

e

odm,

h o NEAOEEK He C M e n i H B a e T C H b 3 t o m ace cMbicjie, h t c m 6 o j i e e , KOHeiHO, e r o Hejib3a cMeiuamb,

T o r f l a KaK o 6 e s t h xpaHcc^opMamni BnojiHe ecTecTBeHHW

BToporo

H3 nepeHHCJieHHbix 3HaneHHH. CmydeHm

Mambcn.

H o b s t o m CMMCJie e r o Hejib3H 3anuMamb,

He TOJibKo 3aHUMambCH nod Konmopy,

oeoufnyio

Mowcem,

xouem,

pemuA

T o r a a KaK noMeufenue

6a3y, demcKuü

wih 3am-

MoaceT

cad, h T.n., h o e r o

MOaCHO TaKHM 06pa30M 3ÜHUMamb H 3ÜHHmb, X O T H OHO H He MOaCeT 3ÜHambCS. CMeXOM eatomcn

H CM3ÜMU MOaCHO KaK 3ÜAUeambCM, TaK H 3ÜAUmbCH. Jlyz,

HU3UHÜ 3ÜAU-

eodoü, h o 3üAumbCH ero Bpaa j i h cnoco6nbi. O f l H a K o soda MoaceT h x 3aAumb,

b t o BpeMfl KaK CMex h cAe3bi HHKoro 3aAumb He M o r y T . HeAoeeK nepeSupaemcn dpyzyto

deapmupy,

h T.n., a oeoiyu

nepeóupawmcn

e oeonfexpanuAuufe.

OTJiHHHe o t nepee3acaiomHX Ha a p y r y i o KBapTHpy oeoufu h o h x M o r y T nepeóupamb. ohh

K o r f l a nepenucbieatomcn

He M o r y T nepenucambcx.

He M o r y T

OflHaKo,

na b

nepeópambcn,

Atodu, t o , b o t j i h h h c o t pyKonHceñ,

PyKonncb ace MoacHo e m e h

nepenucamb.

Toeapu

149

HEKOTOPBIE OCOEEHHOCTH TJIArOJIbHOH rffllEPJIEKCEMbl

cöbieamb. OHH x o p o r n o HJIH ÜJIOXO cöbieatomcn,

MOJKHO

HO B OTJIHHHC OT

npea-

CKa3aHHH H MeHTaHHH OHH He MOryT CÖbimbCM, B TO BpeMH KaK nOCJieflHHe HejIb3S

cßbimb. CuocumbCH (cnecmucb) öeAbe HOCUAU

UAU

HaXOÖUmbCH zde-HUÖydb He

HJIH HTO OHO OKa3aJIOCb HeflOCTaTOHHO HOCKUM.

COOTHOCHMO

xoMcdeme,

c HanajibCTBOM, HO ecjiH

MOJKHO C a p y r H M r o p o a o M ,

nAambe cuocuAocb, TO STO n p o H 3 o u i J i o n o T O M y , HTO e r o CJIHIIIKOM .ziojiro c Haümucb, TORAA KAK Hdxodumbcn

XOÖOK

H np.

Oöxodumbcn

3a dem

xodumb,

COOTHOCHTCH C

( 6 e 3 n e r o - j i H Ö o ) c o B e p m e H H O He

M0p c T B 0 , HTO

flajibHeiiiiiaa

p a 3 p a 6 o T K a npiiHUHnoB H Hfleä

A a r o m e f i rpaMMaTHKH MoaceT 6 b i T b npoflyKTHBHoä H TJiy60K0r0

nOHHMaHHH TOH TOJIH, KOTOpyiO

BHCTHKH a r p a e T H c c j i e a o B a H H e j i e K c e M T o r o HJIH

TOJibKo

n p n ycjioBHH

nopoacnojiHoro

B pa3BHTHH 3TOH OÖJiaCTH JIHHT-

rnnepjieKceMHbix

oTHoiueHHH H B o o ö m e

ranep-

apyroro » 3 b i K a . E e 3 3 T o r o nocTpoeHHe n o p o a c A a i o m H X M O f l e j i e ü

He M o a c e T BMHTH H He B b i f i f l e T 3 a paMKH a ö c T p a K T H o r o T e o p e T H H e c K o r o PACCYXCAEHHH. IIpeflCTaBJiHJiocb npHHUHnnajibHyio

TaK»ce

KpafiHe c y m e c T B e H H b i M

pa3HOTanHOCTb

pa3Hbix

eme

H eme

pa3

noflHepKHyTb

ecTecTBeHHbix HejioBenecKHX J B M K O B ,

npexwe Bcero HMCHHO B OTHOIHCHHH CBOHCTBeHHbix HM ranepjieKceM H, cjieaoBaTejibHO, B njiaHe npHMeHHMocTH K HHM Haeii 'nopoac/iaiomeH rpaMMaTHKH'. Il03T0My jiHHrBHCTHHecKaa Teopna, He yHHTbißaiomaa Bcen CJIOÄHOCTH H MHOTOo6pa3H« HejioBenecKHX H3MKOB — "Die Verschiedenheit des menschlichen Sprachbaues" — c caMoro Hanajia oöpeneHa Ha Heycnex. MOCKBA

R O L A N D BARTHES

P R O U S T ET LES N O M S

O

N sait que la Recherche du Temps Perdu est l'histoire d'une écriture. Cette histoire, il n'est peut-être pas inutile de la rappeler pour mieux saisir comment elle s'est dénouée, puisque ce dénouement figure ce qui, en définitive, permet à l'écrivain d'écrire. La naissance d'un livre que nous ne connaîtrons pas mais dont l'annonce est le livre même de Proust, se joue comme un drame, en trois actes. Le premier acte énonce la volonté d'écrire: le jeune narrateur perçoit en lui cette volonté à travers le plaisir érotique que lui procurent les phrases de Bergotte er la joie qu'il ressent à décrire les clochers de Martinville. Le second acte, fort long puisqu'il occupe l'essentiel du Temps Perdu, traite de l'impuissance à écrire. Cette impuissance s'articule en trois scènes, ou, si l'on préfère, trois détresses : c'est d'abord Norpois qui renvoie au jeune narrateur une image décourageante de la littérature: image ridicule et qu'il n'aurait pourtant même pas le talent d'accomplir; bien plus tard, une seconde image vient le déprimer davantage: un passage retrouvé du Journal des Goncourt, à la fois prestigieux et dérisoire, lui confirme, par comparaison, son impuissance à transformer la sensation en notation; enfin, plus grave encore, parce que portant sur sa sensibilité même et non plus sur son talent, un dernier incident le dissuade définitivement d'écrire : apercevant, du train qui le ramène à Paris après une longue maladie, trois arbres dans la campagne, le narrateur ne ressent qu'indifférence devant leur beauté; il conclut qu'il n'écrira jamais; tristement libéré de toute obligation envers un vœu qu'il est décidément incapable d'accomplir, il accepte de rentrer dans la frivolité du monde et de se rendre à une matinée de la duchesse de Guermantes. C'est ici que par un renversement proprement dramatique, parvenu au fond même du renoncement, le narrateur va retrouver, offert à sa portée, le pouvoir de l'écriture. Ce troisième acte occupe tout le Temps Retrouvé et comprend lui aussi trois épisodes; le premier est fait de trois éblouissements successifs: ce sont trois réminiscences (Saint-Marc, les arbres du train, Balbec), surgies de trois menus incidents, lors de son arrivée à l'hôtel de Guermantes (les pavés inégaux de la cour, le bruit d'une petite cuiller, une serviette empesée que lui tend un valet); ces réminiscences sont des bonheurs, qu'il s'agit maintenant de comprendre, si l'on veut les conserver, ou du moins les rappeler à

PROUST ET LES NOMS

151

volonté : dans un second épisode, qui forme l'essentiel de la théorie proustienne de la littérature, le narrateur s'emploie systématiquement à explorer les signes qu'il a reçus et à comprendre ainsi, d'un seul mouvement, le monde et le Livre, le Livre comme monde et le monde comme Livre. Un dernier suspens vient cependant retarder le pouvoir d'écrire: ouvrant les yeux sur des invités qu'il avait perdus de vue depuis longtemps, le narrateur perçoit avec stupeur qu'ils ont vieilli : le Temps, qui lui a rendu l'écriture, risque au même moment de la lui retirer: vivra-t-il assez pour écrire son œuvre? Oui, s'il consent à se retirer du monde, à perdre sa vie mondaine pour sauver sa vie d'écrivain. L'histoire qui est racontée par le narrateur a donc tous les caractères dramatiques d'une initiation; il s'agit d'une véritable mystagogie, articulée en trois moments dialectiques : le désir (le mystagogue postule une révélation), l'échec (il assume les dangers, la nuit, le néant), l'assomption (c'est au comble de l'échec qu'il trouve la victoire). Or, pour écrire la Recherche, Proust a lui-même connu, dans sa vie, ce dessin initiatique; au désir très précoce d'écrire (formé dès le lycée) a succédé une longue période, non d'échecs sans doute, mais de tâtonnements, comme si l'œuvre véritable et unique se cherchait, s'abandonnait, se reprenait sans jamais se trouver; et comme celle du narrateur, cette initiation négative, si l'on peut dire, s'est faite à travers une certaine expérience de la littérature: les livres des autres ont fasciné, puis déçu Proust, comme ceux de Bergotte ou des Goncourt ont fasciné et déçu le narrateur; cette 'traversée de la littérature' (pour reprendre en l'adaptant un mot de Philippe Sollers), si semblable au trajet des initiations,empli de ténèbres et d'illusions, s'est faite au moyen du pastiche (quel meilleur témoignage de fascination et de démystification que le pastiche?), de l'engouement éperdu (Ruskin) et de la contestation (Sainte-Beuve). Proust s'approchait ainsi de la Recherche (dont, comme on sait, certains fragments se trouvent déjà dans le Sainte-Beuve), mais l'œuvre n'arrivait pas à 'prendre'. Les unités principales étaient là (rapports de personnages,1 épisodes cristallisateurs2), elles s'essayaient à diverses combinaisons, comme dans un kaléidoscope, mais il manquait encore l'acte fédérateur qui devait permettre à Proust d'écrire la Recherche sans désemparer, de 1909 à sa mort, au prix d'une retraite dont on sait combien elle rappelle celle du narrateur lui-même, à la fin du Temps Retrouvé. On ne cherche pas ici à expliquer l'œuvre de Proust par sa vie; on traite seulement d'actes intérieurs au discours lui-même (en conséquence, poétiques et non biographiques), que ce discours soit celui du narrateur ou celui de Marcel Proust. Or l'homologie qui, de toute évidence, règle les deux discours, appelle un dénouement symétrique: il faut qu'à la fondation de l'écriture par la réminiscence (chez le narrateur) corresponde (chez Proust) quelque découverte semblable, propre à fonder définitivement, dans sa continuité prochaine, toute l'écriture de la Recherche. Quel est donc l'accident, non point biographique, mais créateur, qui rassemble une œuvre déjà 1

Par exemple : le visiteur intempestif des soirées de Combray, qui sera Swann, l'amoureux de la petite bande, qui sera le narrateur. 2 Par exemple: la lecture matinale du Figaro, apporté au narrateur par sa mère.

152

ROLAND BARTHES

conçue, essayée, mais non point écrite? Quel est le ciment nouveau qui va donner la grande unité syntagmatique à tant d'unités discontinues, éparses? Qu'est-ce qui permet à Proust d'énoncer son œuvre? En un mot, qu'est-ce que l'écrivain trouve, symétrique aux réminiscences que le narrateur avaient explorées et exploitées lors de la matinée Guermantes? Les deux discours, celui du narrateur et celui de Marcel Proust, sont homologues, mais non point analogues. Le narrateur VA écrire, et ce futur le maintient dans un ordre de l'existence, non de la parole; il est aux prises avec une psychologie, non avec une technique. Marcel Proust, au contraire, écrit; il lutte avec les catégories du langage, non avec celles du comportement. Appartenant au monde référentiel, la réminiscence ne peut être directement une unité du discours, et ce dont Proust a besoin, c'est d'un élément proprement poétique (au sens que Jakobson donne à ce mot) ; mais aussi il faut que ce trait linguistique, comme la réminiscence, ait le pouvoir de constituer l'essence des objets romanesques. Or il est une classe d'unités verbales qui possède au plus haut point ce pouvoir constitutif, c'est celle des noms propres. Le nom propre dispose des trois propriétés que le narrateur reconnaît à la réminiscence : le pouvoir d'essentialisation (puisqu'il ne désigne qu'un seul référent), le pouvoir de citation (puisqu'on peut appeler à discrétion toute l'essence enfermée dans le nom, en le proférant), le pouvoir d'exploration (puisque l'on 'déplie' un nom propre exactement comme on fait d'un souvenir) : le nom propre est en quelque sorte la forme linquistique de la réminiscence. Aussi, l'événement (poétique) qui a 'lancé' la Recherche, c'est la découverte des Noms; sans doute, dès le Sainte-Beuve, Proust disposait déjà de certains noms (Combray, Guermantes)-, mais c'est seulement entre 1907 et 1909, semble-t-il, qu'il a constitué dans son ensemble le système onomastique de la Recherche: ce système trouvé, l'œuvre s'est écrite immédiatement.3 L'œuvre de Proust décrit un immense, un incessant apprentissage.4 Cet apprentissage connait toujours deux moments (en amour, en art, en snobisme): une illusion et une déception; de ces deux moments, nait la vérité, c'est-à-dire l'écriture; mais entre le rêve et le réveil, avant que la vérité surgisse, le narrateur proustien doit accomplir une tâche ambiguë (car elle mène à la vérité à travers bien des méprises), qui consiste à interroger éperdument les signes : signes émis par l'œuvre d'art, par l'être aimé, par le milieu fréquenté. Le Nom propre est lui aussi un signe, et non bien entendu, un simple indice qui désignerait, sans signifier, comme le veut la conception courante, de Peirce à Russell. Comme signe, le Nom propre s'offre à une exploration, à un déchiffrement: il est à la fois un 'milieu' (au sens biologique du terme), dans lequel il faut se plonger, baignant indéfiniment dans toutes les rêveries qu'il porte,5 et un objet précieux, comprimé, embaumé, qu'il faut ouvrir comme une fleur.6 Autrement dit, si 8

Proust a donné lui-même sa théorie du nom propre à deux reprises : dans le Contre Sainte-Beuve (chapitre XIV: Noms de personnes) et dans Du Côté de chez Swann (Tome II, 3e partie: Noms de Pays: le Nom). 1 C'est la thèse de Gilles Deleuze dans son livre remarquable: Proust et les Signes (Paris, P.U.F.). 5 "... ne pensant pas aux noms comme à un idéal inaccessible, mais comme à une ambiance réelle dans laquelle j'irais me plonger" (Du Côté de chez Swann, Paris, Gallimard, 1929, in-8, tome II, p. 236).

PROUST ET LES NOMS

153

le Nom (on appelera ainsi, désormais, le nom propre) est un signe, c'est un signe volumineux, un signe touj ours gros d'une épaisseur touffue de sens, qu'aucun usage nevient réduire, aplatir, contrairement au nom commun, qui ne livre jamais qu'un de ses sens par syntagme. Le Nom proustien est à lui seul et dans tous les cas l'équivalent d'une rubrique entière de dictionnaire : le nom de Guermantes couvre immédiatement tout ce que le souvenir,l'usage, la culture peuvent mettre en lui; il ne connaît aucune restriction sélective, le syntagme dans lequel il est placé lui est indifférent ; c'est donc, d'une certaine manière, une monstruosité sémantique, car, pourvu de tous les caractères du nom commun, il peut cependant exister et fonctionner hors de toute régie projective. C'est là le prix — ou la rançon — du phénomène d"hypersémanticité' dont il est le siège,7 et qui l'apparente, bien entendu, de très près, au mot poétique. Par son épaisseur sémantique (on voudrait presque pouvoir dire: par son 'feuilleté') le Nom proustien s'offre à une véritable analyse sémique, que le narrateur lui-même ne manque ni de postuler ni d'esquisser: ce qu'il appelle les différentes 'figures' du Nom, 8 sont de véritables sèmes, doués d'une parfaite validité sémantique, en dépit de leur caractère imaginaire (ce qui prouve une fois de plus combien il est nécéssaire de distinguer le signifié du référent). Le nom de Guermantes contient ainsi plusieurs primitifs (pour reprendre un mot de Leibnitz) : "un donjon sans épaisseur qui ri était qu'une bande de lumière orangée et du haut duquel le seigneur et sa dame décidaient de la vie ou de la mort de leurs vassaux" ; "une tour jaunissante et fleuronnée qui traverse les âges" ; l'hôtel parisien des Guermantes, "limpide comme son nom" ; un chateau féodal en plein Paris, etc. Ces sèmes sont, bien entendu, des 'images', mais dans la langue supérieure de la littérature, elles n'en sont pas moins de purs signifiés, offerts comme ceux de la langue dénotante à toute une systématique du sens. Certaines de ces images sémiques sont traditionnelles, culturelles: Parme ne désigne pas une ville d'Emilie, située sur le Pô, fondée par les Etrusques, grosse de 138.000 habitants; le véritable signifié de ces deux syllabes est composé de deux sèmes : la douceur stendhalienne et le reflet des violettes.9 D'autres sont individuelles, mémorielles: Balbec a pour sèmes deux mots dits autrefois au narrateur, l'un par Legrandin (Balbec est un lieu de tempêtes, en fin de terre), l'autre par Swann (son église est du gothique normand, à moitié roman), en sorte que le nom a toujours deux sens simultanés: "architecture gothique et tempête sur la mer". 10 Chaque nom a ainsi son spectre sémique, variable dans le temps, selon la chronologie de son lecteur, qui ajoute ou retranche de ses éléments, exactement comme fait la langue dans sa diachronie. Le Nom est en effet CATALYSABLE: on peut le remplir, le dilater, combler les interstices de son armature 6

"... dégager délicatement des bandelettes de l'habitude et revoir dans sa fraîcheur première ce nom de Guermantes, ..." (Contre Sainte-Beuve, Paris, Gallimard, 1954, p. 316). 7 Cf. U. Weinreich, "On the Semantic Structure of Language", in: J. H. Greenberg, Universals of Language (Cambridge, Mass., The M.I.T. Press, 1963; 2nd ed. 1966), p. 147. 8 "Mais plus tard, je trouve successivement dans la durée en moi de ce même nom, sept ou huit figures différentes..." {Le Côté de Guermantes, édition citée, I, p. 14). 8 Du côté de chez Swann, édition citée, II, p. 234. 10 Ibid., p. 230.

154

ROLAND BARTHES

sémique, d'une infinité de rajofits. Cette dilatation sémique du nom propre peut être définie d'une autre façon: chaque nom contient plusieurs 'scènes' surgies d'abord d'une manière discontinue, erratique, mais qui ne demandent qu'à se fédérer et à former de la sorte un petit récit, car raconter, ce n'est jamais que lier entre elles, par procès métonymique, un nombre réduit d'unités pleines: Balbec récèle ainsi non seulement plusieurs scènes, mais encore le mouvement qui peut les rassembler dans un même syntagme narratif, car ses syllables hétéroclites étaient sans doute nées d'une manière désuète de prononcer, "que je ne doutais pas de retrouver jusque chez l'aubergiste qui me servirait du café au lait à mon arrivée, me menant voir la mer déchaînée devant /' église et auquelje prêtais l'aspect disputeur, solennel et médiéval d'un personnage de fabliau".11 C'est parce que le Nom propre s'offre à une catalyse d'une richesse infinie, qu'il est possible de dire que, poétiquement, toute la Recherche est sortie de quelques noms.12 Encore faut-il bien les choisir — ou les trouver. C'est ici qu'apparait, dans la théorie proustienne du Nom, l'un des problèmes majeurs, sinon de la linguistique, du moins de la sémiologie: la motivation du signe. Sans doute ce problème est-il ici quelque peu artificiel, puisqu'il ne se pose en fait qu'au romancier, qui a la liberté (mais aussi le devoir) de créer des noms propres à la fois inédits et 'exacts' ; mais à la vérité, le narrateur et le romancier parcourent en sens inverse le même trajet; l'un croit déchiffrer dans les noms qui lui sont donnés une sorte d'affinité naturelle entre le signifiant et le signifié, entre la couleur vocalique de Parme et la douceur mauve de son contenu; l'autre, devant inventer quelque lieu à lafois normand, gothique et venteux, doit chercher dans la tablature générale des phonèmes, quelques sons accordés à la combinaison de ces signifiés; l'un décode, l'autre code, mais il s'agit du même système et ce système est d'une façon ou d'une autre une système motivé, fondé sur un rapport d'IMITATION entre le signifiant et le signifié. Codeur et décodeur pourraient reprendre ici à leur compte l'affirmation de Cratyle: "La propriété du nom consiste à représenter la chose telle qu'elle est." Aux yeux de Proust, qui ne fait que théoriser l'art général du romancier, le nom propre est une simulation, ou comme disait Platon (il est vrai avec défiance) une 'fantasmagorie'. Les motivations alléguées par Proust sont de deux sortes, naturelles et culturelles. Les premières relèvent de la phonétique symbolique.13 Ce n'est pas le lieu de reprendre le débat de cette question (connue autrefois sous le nom d'HARMONIE IMITATIVE), OÙ l'on retrouverait, entre autres, les noms de Platon, Leibnitz, Diderot et Jakobson. 14 On rappelera seulement ce texte de Proust, moins célèbre mais peut-être aussi pertinent que le Sonnet des Voyelles : "... Bayeux, si haute dans sa noble dentelle rougeâtre et dont le faîte est illuminé par le vieil or de sa dernière syllable ; Vitré, dont l'accent aigu lo11 12

Ibid., p. 234.

"C'était, ce Guermantes, comme le cadre d'un roman" (Le côté de Guermantes, édition citée, I, p. 15). 13 Weinreich (op. cit.) a noté que le symbolisme phonétique relève de l'hypersémanticité du signe. 14 Platon: Cratyle; Leibnitz: Nouveaux Essais (III, 2); Diderot: Lettre sur les sours et muets ; R. Jakobson : Essais de Linguistique Générale.

PROUST ET LES NOMS

155

songeait de bois noir le vitrage ancien', le doux Lamballe qui, dans son blanc, va du jaune coquille d'œuf au gris perle, Coutances, cathédrale normande, que sa diphtongue finale, grasse et jaunissante, couronné par une tour de beurre", etc.15 Les exemples de Proust, par leur liberté et leur richesse (il ne s'agit plus ici d'attribuer à l'opposition i/o le contraste traditionnel du petit¡gros ou de Vaigu/rond, comme on le fait d'ordinaire: c'est toute une gamme de signes phoniques qui est décrite par Proust), ces exemples montrent bien que d'ordinaire la motivation phonétique ne se fait pas directement: le décrypteur intercale entre le son et le sens un concept intermédiaire, mi-matériel, mi-abstrait, qui fonctionne comme une clef et opère le passage, en quelque sorte démultiplié, du signifiant au signifié : si Balbec signifie affinitairement un complexe de vagues aux crêtes hautes, de falaises escarpées et d'architecture hérissée, c'est parce que l'on dispose d'un relai conceptuel, celui du RUGUEUX, qui vaut pour le toucher, l'ouïe, la vue. Autrement dit, la motivation phonétique exige une nomination intérieure : la langue rentre subrepticement dans une relation que l'on postulait — mythiquement — comme immédiate: la plupart des motivations apparentes reposent ainsi sur des métaphores si traditionnelles (le RUGUEUX appliqué au son) qu'elles ne sont plus senties comme telles, ayant passé toutes entières du côté de la dénotation; il n'empêche que la motivation se détermine au prix d'une ancienne anomalie sémantique, ou, si l'on préfère, d'une ancienne transgression. Car c'est évidemment à la métaphore qu'il faut rattacher les phénomènes de phonétisme symbolique, et il ne servirait à rien d'étudier l'un sans l'autre. Proust fournirait un bon matériel à cette étude combinée : ses motivations phonétiques impliquent presque toutes (sauf peut-être pour Balbec) une équivalence entre le son et la couleur: ieu est vieil or, é est noir, an est jaunissant, blond est doré (dans Coutances et Guermantes), i est pourpre.16 C'est là une tendance évidemment générale: il s'agit de faire passer du côté du son des traits appartenant à la vue (et plus particulièrement à la couleur, en raison de sa nature à la fois vibratoire et modulante), c'est-à-dire, en somme, de neutraliser l'opposition de quelques classes virtuelles, issues de la séparation des sens (mais cette séparation est-elle historique ou anthropologique? De quand datent et d'où viennent nos 'cinq sens'? Une étude renouvelée de la métaphore devrait désormais passer, semble-t-il, par l'inventaire des classes nominales attestées par la linguistique générale). En somme, si la motivation phonétique implique un procès métaphorique, et par conséquent une transgression, cette transgression se fait en des points de passage éprouvés, comme la couleur: c'est pour cela, sans doute, que les motivations avancées par Proust, tout en étant très développées, apparaissent 'justes'. Reste un autre type de motivations, plus 'culturelles', et en cela analogues à celles que l'on trouve dans la langue: ce type règle en effet à la fois l'invention des néo16 Du côté de chez Swann, édition citée, II, p. 234. On remarquera que la motivation alléguée par Proust n'est pas seulement phonétique, mais aussi, parfois, graphique. 16 "La couleur de Sylvie, c'est une couleur pourpre, d'un rose pourpre en velours pourpre ou violacé.... Et ce nom lui-même, pourpre de ses deux I — Sylvie, la vraie Fille du Feu" (Contre SainteBeuve, édition citée, p. 195).

156

ROLAND BARTHES

logismes, alignés sur un modèle morphématique, et celle des noms propres, 'inspirés', eux, d'un modèle phonétique. Lorsqu'un écrivain invente un nom propre, il est en effet tenu aux mêmes règles de motivation que le législateur platonicien lorsqu'il veut créer un nom commun; il doit, d'une certaine façon, 'copier' la chose, et comme c'est évidemment impossible, du moins copier la façon dont la langue elle-même a créé certains de ses noms. L'égalité du nom propre et du nom commun devant la création est bien illustrée par un cas extrême : celui où l'écrivain fait semblant d'user de mots courants qu'il invente cependant de toutes pièces: c'est le cas de Joyce et de Michaux; dans le Voyage en Grande Garabagne, un mot comme arpette n'a — et pour cause — aucun sens mais n'en est pas moins empli d'une signification diffuse, en raison, non seulement de son contexte, mais aussi de sa sujétion à un modèle phonique très courant en français.17 Il en est ainsi des noms proustiens. Que Laumes, Argencourt, Villeparisis, Combray ou Doncières existent ou n'existent pas, ils n'en présentent pas moins (et c'est cela qui importe) ce qu'on a pu appeler une 'plausibilité francophonique' : leur véritable signifié est: France, ou mieux encore la 'francité'; leur phonétisme, et au moins à titre égal leur graphisme, sont élaborés en conformité avec des sons et des groupes de lettres attachés spécifiquement à la toponymie française (et même, plus précisément, francienne) : c'est la culture (celle des Français) qui impose au Nom une motivation naturelle : ce qui est imité n'est certes pas dans la nature, mais dans l'histoire, une histoire cependant si ancienne qu'elle constitue le langage qui en est issu en véritable nature, source de modèles et de raisons. Le nom propre, et singulièrement le nom proustien, a donc une signification commune: il signifie au moins la nationalité et toutes les images qui peuvent s'y associer. Il peut même renvoyer à des signifiés plus particuliers, comme la province (non point en tant que région, mais en tant que milieu), chez Balzac, ou comme la classes sociale, chez Proust: non certes par la particule anoblissante, moyen grossier, mais par l'institution d'un large système onomastique, articulé sur l'opposition de l'aristocratie et de la roture d'une part, et sur celle des longues à finales muettes (finales pourvues en quelque sorte d'une longue traîne) et des brèves abruptes d'autre part: d'une côté le paradigme des Guermantes, Laumes, Agrigente, de l'autre celui des Verdurin, Morel, Jupien, Legrandin, Sazerat, Cottard, Brichot, etc.18 L'onomastique proustienne parait à ce point organisée qu'elle semble bien constituer le départ définitif de la Recherche : tenir le système des noms, c'était pour Proust et c'est pour nous, tenir les significations essentielles du livre, l'armature de ses signes, sa syntaxe profonde. On voit donc que le nom proustien dispose pleinement des deux 17 Ces mots inventés ont été bien analysés, d'un point de vue linguistique, par Delphine Perret, dans sa thèse de 3e Cycle: Etude de la langue littéraire d'après le Voyage en Grande Garabagne d'Henri Michaux (Paris, Sorbonne, 1965-1966). 18 II s'agit, bien entendu, d'une tendance, non d'une loi. D'autre part, on entend ici LONGUES et BRÈVES, sans rigueur phonétique, mais plutôt comme une impression courante, fondée d'ailleurs en grande partie sur le graphisme, les Français étant habitués par leur culture scolaire, essentiellement écrite, à percevoir une opposition tyrannique entre les rimes masculines et les rimes féminines, senties les unes comme brèves, les autres comme longues.

PROUST ET LES NOMS

157

grandes dimensions du signe: d'une part il peut être lu tout seul, 'en soi', comme une totalité de significations (Guermantes contient plusieurs figures), bref comme une essence (une 'entité originelle', dit Proust), ou si l'on préfère, une absence, puisque le signe désigne ce qui n'est pas là; 19 et d'autre part, il entretient avec ses congénères des rapports métonymiques, fonde le Récit: Swann et Guermantes ne sont pas seulement deux routes, deux côtés, ce sont aussi deux phonétismes, comme Verdurin et Laumes. Si le nom propre a chez Proust cette fonction œcuménique, résumant en somme tout le langage, c'est que sa structure coïncide avec celle de l'œuvre même: s'avancer peu à peu dans les significations du nom (comme ne cesse de le faire le narrateur), c'est s'initier au monde, c'est apprendre à déchiffrer ses essences : les signes du monde (de l'amour, de la mondanité) sont faits des mêmes étapes que ses noms; entre la chose et son apparence, se développe le rêve, tout comme entre le réfèrent et son signifiant s'interpose le signifié: le nom n'est rien, si par malheur on l'articule directement sur son référent (qu'est, EN RÉALITÉ, la duchesse de Guermantes?), c'est-à-dire si l'on manque en lui sa nature de signe. Le signifié, voilà la place de l'imaginaire : c'est là, sans doute, la pensée nouvelle de Proust, ce pour quoi il a déplacé, historiquement, le vieux problème du réalisme, qui ne se posait guère, jusqu'à lui, qu'en termes de référents : l'écrivain travaille, non sur le rapport de la chose et de sa forme (ce qu'on appelait, aux temps classiques, sa 'peinture', et plus récemment, son 'expression'), mais sur le rapport du signifié et du signifiant, c'est-à-dire sur un signe. C'est ce rapport dont Proust ne cesse de donner la théorie linguistique dans ses réflexions sur le Nom et dans les discussions étymologiques qu'il confie à Brichot et qui n'auraient guère de sens si l'écrivain ne leur confiait une fonction emblématique.20 Ces quelques remarques ne sont pas seulement guidées par le souci de rappeler, après Claude Lévi-Strauss, le caractère signifiant, et non pas indiciel, du nom propre.21 On voudrait aussi insister sur le caractère cratyléen du nom (et du signe) chez Proust: non seulement parce que Proust voit le rapport du signifiant et du signifié comme un rapport motivé, l'un copiant l'autre et reproduisant dans sa forme matérielle l'essence signifiée de la chose (et non la chose elle-même), mais aussi parce que, pour Proust comme pour Cratyle, "la vertu des noms est d'enseigner" :22 il y a une propédeutique des noms, qui conduit, par des chemins souvent longs, variés, détournés, à l'essence des choses. C'est pour cela que personne n'est plus proche du Législateur cratyléen, fondateur des noms (demiourgos onomatôri), que l'écrivain proustien, non parce qu'il qu'il est libre d'inventer les noms qu'il lui plait, mais parce qu'il est tenu de les inventer 'droit'. Ce réalisme (au sens scolastique du terme), qui veut que les noms soient le 'reflet' des idées, a pris chez Proust une forme radicale, mais on peut se demander s'il n'est pas plus ou moins consciemment présent dans tout acte d'écriture et s'il est 19 "On ne peut imaginer que ce qui est absent" (Le Temps Retrouvé, Paris, Gallimard, III, p. 872). — Rappelons encore que pour Proust, imaginer, c'est déplier un signe. 20 Sodome et Gomorrhe, II, ch. 2. 21 La Pensée Sauvage (Paris, Pion, 1952), p. 285. 22 Platon, Cratyle, 435 d.

158

ROLAND BARTHES

vraiment possible d'être écrivain sans croire, d'une certaine manière, au rapport naturel des noms et des essences: la fonction poétique, au sens le plus large du terme, se définirait ainsi par une conscience cratyléenne des signes et l'écrivain serait le récitant de ce grand mythe séculaire qui veut que le langage imite les idées et que, contrairement aux précisions de la science linguistique, les signes soient motivés. 23 Cette considération devrait incliner encore davantage le critique à lire la littérature dans la perspective mythique qui fonde son langage, et à déchiffrer le mot littéraire (qui n'est en rien le mot courant), non comme le dictionnaire l'explicite, mais comme l'écrivain le construit. ECOLE PRATIQUE DES HAUTES ETUDES PARIS

,s

Mythe invétéré, de Pythagore à Vico, et au-delà.

C. E. BAZELL

ON THE DECLENSION OF BATS SUBSTANTIVES

I. THE GENITIVE SINGULAR AND RELATED CASES

I

N his recent account of the declension of Bats substantives, R. R. Gagua 1 wishes to explain a difference between the stems in back vowel and in front vowel in phonological terms. At first sight the explanation may seem evident. We have e.g. Nom. ca Gen. cai Dat. cain

'bear'

ze zè zen

'sheep'

and this difference runs through most of the singular declension. Gagua (p. 83) considers that an -/- affix characterised these oblique cases generally, but was eliminated by assimilation in stems with front vowel. What makes this, even at first sight, less than obvious, is the fact that ei is quite a normal sequence in Bats in words where it is not suspect of more recent origin. More relevant however is the evidence that in polysyllabic words, in particular in nominal e-stems, an assimilation led not to e but rather to i. This evidence is rather obscure in modern Bats, but it is fairly clear in the language as described over a hundred years ago by Schiefner.2 Though this scholar records many genitives in -i of nouns which are certainly not stems in back-vowel, his information cannot be accepted uncritically. At the same time it cannot be disregarded. I shall anticipate my conclusions by saying that Schiefner failed to distinguish two different cases of the e-stems, the genitive (formed by nasalisation of the stem-vowel) and the 'directive' (formed by addition of i). In other nouns there was syncretism of these two cases, and in modern Bats the distinction has been lost. Schiefner was an observant scholar and clues to the case-distinction may be found in his text. He hints (p. 69) that the 'genitive' for 'Ziel' terminates more often than could otherwise be expected in -i with nouns otherwise showing another form of genitive. This is confirmed by examples in his specimens of the language. Admittedly 1 1

R. R. Gagua, Voprosy izucenija iberijsko-kavkazskih jazykov (Moscow 1961), 72-88. A. Schiefner, Versuch über die Tusch-Sprache (St. Petersburg, 1856).

160

C. E. BAZELL

these are rather literal translations from Russian, but this hardly affects their value for the present purpose. I cite only p. 99: Sinna lai ist da, cru wecer waxa oxu denih, quw ma silgeco denih; ai xeten o si baxe chang bazri. Dagi cruen, me chain uirwas dakardie itt bah okrui ... sairwa waxe bazri. Cru tquih wocwali, bazre kotleh titen oxun fiila maxkiw ... I have omitted diacritics from the citation since they could be of no use; students of Bats can supply them. I have not inserted any either; the three forms italicized should be understood as having final nasalised vowels which for some reason (perhaps the peculiarity of informants) were overlooked by Schiefner. The translation bears out a distinction between the 'directive' (bazri) and the genitive (bazre): "Both agreed to do it, the trickster had to go on that day and the thief on the next day; they got up and went together to the market. The trickster saw that a Jew was counting a thousand pieces of gold ... he went himself to the market. The trickster followed, and in the crowding of the market he cut open his breast-pocket with a knife." This distinction is quite clear in 19th-century Bats. But it also affords a clue to the interpretation of modern Bats. Gagua's 'consonant-stems' to which he devotes the first half of his article are not to be so classified. They are in fact e-stems, and there is no reason for his 'genitive terminations' in -e; the inflection consists merely in the nasalisation. In modern Bats there is normally no terminal vowel in the nominative at the phonetic level, but even the most primitive morpho-phonemics would supply its equivalent. I suggest that all the nouns of Bats are to be regarded as vowel-terminal. But this raises some questions in respect of the plural of nouns, which has no simple relation to the singular. It may be preferable to bypass the question for the moment, but it is interesting to note that the terminal stem-vowel of the singular noun has relevance to the plural formation. Stems in -a- tend to have a plural in -(a)si- while there are very few similar formations in stems with any other vowel. Stems in -o- tend to form the plural by simple addition of -i; about half of the native words have this termination, which is the only one applied to loan-words with this vowel, e.g. zagno 'book', ezo 'yard' (from Georgian), plural zagnoi, ezoi. oi is phonetically ui in non-initial syllables, and this is how the forms are cited in previous writings on Bats. The commonest plural of e-stems (also the most frequent and productive type among nouns in general) is similarly cited, e.g. sing, knaf (gen. knat'e) 'son', plural knafi. It is surely apparent that -i derives from ei, which became ii by a raising parallel to that of oi to ui in non-initial syllables, and hence i (a sequence of two identical vowels both in non-initial syllable being impermissible in Bats). Indeed this is also the only reasonable synchronic interpretation; the basic form is knafe-i, which phonetic rules transform to knat'i. The nominative singular must also be regarded as knat'e, which phonetic rules reduce to knat\ This is rather more problematic, since there are a few nouns re-

ON THE DECLENSION OF BATS SUBSTANTIVES

161

taining -e in the nominative, e.g. ale 'lord' and ase 'calf'. But these are quite exceptional, and must be dealt with as exceptions rather than be allowed to interfere with the rule as such. I am inclined to set up a special 'stable vowel' E distinguished by its retention in circumstances which eliminate the normal vowel e at the phonetic level. This is not an ad hoc solution: we can recognise a whole set of 'stable vowels' (representable by capitals) which are retained in environments (especially the environment final pause) in which vowels are otherwise represented by zero or (in the case of i and u) displaced into the previous syllable. The rules for 'unstable' vowels do not involve any loss or displacement of the final i of the plural, at least in the case of -ei and -oi: it belongs to these rules that the diphthongs are not affected in the first instance, and since these rules apply before the reduction of diphthongs, the final -i is not further affected. Whereas the plurals in -si, very rare except among the a-stems, show the expected displacement of the final vowel, e.g. k'alik'asi (phonetically k'alik'ais) 'towns'. It seems reasonably well established that the common plural in -i of e-stems goes back to -ei, and that the directive singular (differing only by nasalised final) has a corresponding origin. It also seems obvious to take the final -i in the genitive of stems in other vowels, which answers both to the directive and the genitive of the e-stems, to be identical with the vowel which in the e-stems is reserved for the directive alone. It is not surprising that the e-stems, the most frequent class, kept a distinction elsewhere lost. Any attempt to go beyond these reasonable assumptions must be speculative. There is no obvious cause for the generalisation of -i- throughout the other declensions. But we are on quite firm ground in stating that there has been such a generalisation. In Bats as described by Gagua, there is an ergative cakuiv (read cako-i-v) beside the older cakov. There is no trace of -i- in any earlier description: clearly it has been introduced from the genitive, probably via the directive with which there is syncretism. It is of course hazardous to extrapolate too far back, but even the short attested history of Bats suggests that an earlier account would have less to make of the 'intrusive -i-'. It is almost absent from the pronominal declensions of Schiefner; cf. his dative okxun with the okxu-i-n of Deseriev3 a hundred years later. Whatever the origin of 'intrusive -/-', it is not early in most of the forms and is probably late even in forms attested by the earliest descriptions of Bats. It is absent from the participial declension, even in the genitive (-0 not oi). When so much is dubious, it is all the more worth stressing the little that is certain: namely that the 'intrusive -J" is never found in any other case unless it is also found in the genitive or directive. Its existence here is a starting-point for explanations rather than something which can be explained. The first stage in the spread of -/- seems to have been the syncretism of genitive and directive, normally in favour of the latter, in substantival stems other than the -e-stems. The at first purely phonetic coalescence in the /'-stems could have helped this development. A form like xi (from xi 'water') would have represented both the genitive ' J. D. DeSeriev, Bacbijski Jazyk (Moscow, 1953).

162

C. E. BAZELL

(basic xi-n) and the directive (basic xi-in), but, re-interpreted as a morphological identity with the latter basis, could have led to the subsequent genitive-directive (basic -in) of stems in -o -u and -a. The far more frequent e-stems were less likely to be affected. Since stems in -i are virtually lacking outside the substantives, non-substantival declensions such as the participles and adjectives would remain unaffected, as in fact is the case. But many problems remain. There are some stems in -o and -u (though probably not in -a) whose genitive is formed by simple final nasalisation without intrusive -/-. The suggestion of Gagua (p. 84 footnote 11) that the vowel i of the genitive case has been lost, is of course quite irresponsible. The only noticeable common characteristic of such nouns is that they mostly share either a stem containing a nasal or a plural formed with nasal consonant. This may be a clue, but at the moment I can make nothing of it. II. THE PLURAL OF THE /4-STEMS

The only productive plural termination of the a-stems appears to be that in -si, as attested by relatively recent loans such as kveq ana-si ('globe') and masina-si ('machine'). Here may also belong mus-i ('labourer'), probably representing musaSi; cf. the adverb moisior mosi-s from mosi 'bad'. These are all stems in -A by our convention. The commonest termination for stems in unstable a, represented by zero finally and before i, is -is, e.g. sing, mar(a) 'husband', plural mar(a)is, and similarly from the stems basa- 'flower' (plurals also bas(a)bi, bas(a)mi) matfct- 'tongue' batta- 'month', 'moon' etc. It would very roughly account for the data if one were to posit a normal termination -si for the o-stems in general, while this sequence would suffer an obligatory permutation in stems with unstable a. The basic forms would be such as 'urdnA-si (phonetically 'urdnasi or optionally 'urdnais) 'mara-si (phonetically maris) Obviously some such connection must be made between two phonologically similar plural formations both characteristic of (though not peculiar to) the a-stems. At the same time this solution is unconvincing so long as it cannot be extended beyond the stems in -a, much as a recent interpretation4 of Latin erat as from es-ba-t is unconvincing so long as it cannot cover any instances in which a 'consonant' other than b is involved. We cannot feel satisfied with an account which lacks phonetic analogy in a way which is demanded as a matter of course in historical linguistics. Certainly, we do not have to demand of a synchronic order of rules that it actually reproduces the historical order, but we have to demand that it is plausible in much the same way — 4

James Foley, "Prothesis in the Latin verb sum", Language, 41, 59-64.

ON THE DECLENSION OF BATS SUBSTANTIVES

163

that, so to speak, it could be looked at as a 'fictional' historic development conforming to the canons of the developments we know. Now is there any analogue to the distinction between unstable ¿¡-stems and stable .4-stems in respect of the development considered above? I cannot see that there is any such analogue. Plurals with s are relatively rare outside the a-stems, and fall roughly speaking into these categories: (i) monosyllables, e.g. qo-is ('crow'); (ii) verbal derivatives, such as fegdienois, plural of t'egdieno (from the verb fegdo- 'make') 'deed' and (iii) adjectival plurals. The two latter categories are hardly relevant, since neither a distinction of stem-vowel nor a distinction between stable and unstable vowels plays a role here. As for the monosyllabic substantives, it is perhaps relevant that the great majority (though still not a large number) are in fact in -a. It is not unreasonable to suspect that the relative frequency of a monosyllabic stem is in some way connected with the relative normality of a bi-phonemic termination, as opposed to the mono-phonemic plural terminations of stems which are basically disyllabic. But this is little more than a guess. It is interesting to note that distinctions of terminal vowel also play a role in the verb. Unfortunately the information here is very slight. So far as it goes, it appears that verbs with a terminal vowel -u have a present tense with inserted -/'-, e.g. vo/w-5 'go' (imperfect votur), present voitu; this vowel is lacking in other tenses. Verbs with a terminal vowel -o often have ablaut; e.g. pres. mel'o 'drink', future maVo. Verbs with the terminal vowels -/' and -e have no normal future formation: instead the auxiliary lat'e (in independent use 'stand') is used with the infinitive. There is only one verb with terminal -a (apart from compounds with -ala- and an homophonous simple verb), namely 'be' which consists simply of the phoneme a. This, like several other verbs, has a different stem for the future, coincident with x/7'w (probably to be interpreted as xiil'u on the analogy of voitu) = German werden. In view of this general tendency, one might expect the substantival /-stems and w-stems to show some sort of pattern in respect of the plural formation, even if only a negative one like the absence of a future formation in the verbal /-stems. Many of the /'-stem substantives are mass-nouns, probably lacking a plural: at least none is attested for these. Others are words for animals having a plural in -aire, a common formation for this semantic category whatever the stem: e.g. daxkairc from the stem daxki- 'mouse', kotairc from koti- 'cat'. The two outstandingly frequent /-stems, k'ort'i- ('head, beginning, self') and deni- 'day' have an anomalous declension in the singular; their plurals k'ort'mi and denis are similarly uncharacteristic. Stems in -u are few, and their plurals are even more rarely attested, but some pattern emerges here. s

Strictly speaking this should be -otu- since the initial v- is a gender-morpheme. The final -u is not normally regarded as part of the verbal stem since it is lacking in the past tenses. It seems to me more natural to regard these tenses as involving a deletion of the stem-vowel with substitution of other tense-indicators, the stem-vowel of the 'present' to be regarded as that of the verb. However this interpretation is disputable and I should not wish to insist on it if there are good arguments to the contrary.

164

C. E. BAZELL

The three best attested «-stems are lamu- 'sky', 'mountain', gagu- 'knee', and k'mat'u'cliff'. Each of these has a plural in -ni, though the first two seem also to have plurals analogous to those of the o-stems, as described above. Plurals in -ni are otherwise in a small minority, most examples being merely apparent since the nasal is part of the basic form of the singular noun. Hence the w-stems may also be regarded as having a characteristic plural form. UNIVERSITY OF LONDON

IRENA BELLERT

A SEMANTIC APPROACH TO GRAMMAR CONSTRUCTION

O

N E of the most important problems in linguistics is that concerning the approach that should be pursued in constructing a grammar of a natural language in order that an adequate description be achieved. The problem may be summarized in terms of the following questions: (1) Can we postulate syntactical rules with the fewest possible semantic assumptions or without any at all? (2) Does semantics begin where syntax ends? (3) Is everything in deep syntax semantically relevant? It seems that enough evidence has been presented in recent literature that would support answers to these three questions, all of which would suggest one preferable approach to grammar construction, which I shall refer to as a semantic approach. As regards the first question, the answer might be in the affirmative, if and only if we were concerned with the surface structures. Since enough evidence has been presented by the proponents of transformational grammars to show that descriptive adequacy calls for the postulation of deep structures, it would seem unreasonable to insist any longer on the apparent advantages of arriving at syntactical rules without the aid of semantics. As regards the second question, it appears more and more clear from recent publications pertaining to the problem of the interrelation between syntax and semantics 1 that no such boundary can be established. More than that, the conclusions which suggest themselves are that any effort towards establishing such a boundary would be injurious to both the syntactical and semantic representations. 2 In order to answer the third question, which I believe should be in the affirmative, 1

Noam Chomsky, Aspects of the Theory of Syntax (Cambridge, Mass., M.I.T. Press, 1965); J. J. Kate and P. M. Postal, An Integrated Theory of Linguistic Descriptions (Cambridge, Mass., M.I.T. Press, 1964); Uriel Weinreich, "Explorations in Semantic Theory", in Current Trends in Linguistics, ed. by Thomas A. Sebeok, Vol. Ill (The Hague-Paris, Mouton and Co., 1966); A. K. 2olkovskij and J. A. MelCuk, "O vozmoznom metode i instrumentax semantiCeskovo sinteza", Naucno-Texniieskaja Informatsja, No. 6 (Moscow, 1965). ' We use the terms 'syntax' and 'syntactical representation' in the sense understood in the generative transformational grammar, where the deep structure representation belongs to syntax, rather

166

IRENA BELLERT

it is necessary to say a few words about the empirical semantic data we have at our disposal, the extent to which such data have been used in grammar construction in recent proposals, and the question of the adequacy of grammars in the light of a discussion concerning semantic data available empirically. It seems evident that no sentence or phrase, and probably hardly any morpheme of a natural language, is understood otherwise than as an abbreviation standing for a complex semantic structure. Each natural language has its specific abbreviations corresponding to certain 'chosen' composite semantic relations, which are usually untranslatable into analogical abbreviations in another language, but are always translatable in a larger context in a loose way by means of different abbreviations 'chosen' for certain composite relations in the other language. However, most linguistic signs, such as words, phrases, or even sentences which are not complete utterances, are ambiguous when taken out of context. Accordingly, we will consider complete utterances as the basic units of a semantic system, for evidently such units are the most meaningful and the least ambiguous. We may assume with good reason that every speaker-hearer (to whom I shall simply refer as the speaker) of a natural language understands an arbitrarily large number of complex linguistic signs (sentences or strings of sentences which constitute complete utterances), that is to say, he is completely aware of what situations, states of affairs, phenomena, etc., they may apply to. We will assume that to understand a complex sign is to know the corresponding composite semantic relation which comprises any relevant relation within the narrated event, including any relevant relation between the narrated event and the producer of the sign, the addressee of the sign, the time of producing the sign, and the 'things' referred to by the producer of the sign. We assume that the universe of discourse may be conceived of as an infinite class of composite relations, each corresponding to a complex linguistic sign. 3 We thus assume that a human being is capable of extrapolating a composite relation underlying a sign type, that is to say, he is able to associate a sign type with the complexity of relevant underlying relations without being aware of or being able to specify precisely the underlying elementary relations. Similarly, when we observe a horse running over a period of time, we have a different picture of the running horse at any moment within that time. Accordingly, we could speak of an infinite number of composite relations corresponding to different compositions of the horse's movements. However, we recognize any such composite relation at a given time as the same in terms of the abbreviation: "A horse is running", even though it would be hard, if not impossible, to specify all the individual invariant elementary relations of which the composite relation consists. than in the sense employed in logic, where by syntax we understand only the formal relations on the surface of the text. 3 For a more precise description, see I. Bellert and W. Zawadowski, "Preliminary Attempt at Outlining a System of Linguistic Signs in Terms of the Theory of Category", paper presented at the Conference on Semiotics, Kazimierz nad Wisl^, September 1966, to appear in Sign - Language Culture (The Hague-Paris, Mouton & Co., forthcoming).

A SEMANTIC APPROACH TO GRAMMAR CONSTRUCTION

167

Thus, although the occurrence of a linguistic sign such as "This boy is singing" may be correlated with the infinite number of situations to which it could be applied, we may assume that the signification of this sign type depends on the underlying composite relation which remains the same for a number of such situations and for the producer of the sign, the time when the sign is produced, the young human male being referred to by the producer of the sign, and the action designated by the abbreviation used. The nature of the interrelations between any such composite relations associated with the corresponding sign types is determined by grammar, that is, the speaker's tacit knowledge of his language, without which he obviously would not be capable of understanding all the sentences of his language; but the signification of an utterance is clearly determined by his overt knowledge, as evidenced by his ability to paraphrase or simply to use his language properly. If a speaker knows exactly what a complete utterance signifies (to what possible situations, states of affairs, phenomena, etc. it can be applied), this means that he will interpret correctly any structural relations involved in that utterance — without being able to specify them explicitly. This is obviously a result of his tacit knowledge of grammar, although such facts yield data which are available empirically. Every speaker of a natural language is able to paraphrase utterances of his language, that is to say, he is able to identify the corresponding composite relations without being able to specify the elementary relations common to the composite relations in question. We do not claim that an average speaker is precise to the same extent in paraphrasing; this certainly depends on a capacity for deep introspection and a proper semantic training. However, since the problem of postulating a grammar is the task of a linguist rather than of an average speaker, we may assume that precise statements as to the equivalence of significations can be arrived at and assumed to be empirically available to linguists. The well-known method for discovering the phonemic system of a language in terms of which we can identify certain sequences of sounds on the phonemic level is evidently based on the equivalence of significations. For, in general — with the exception of the problem of homonymy and synonymy — we can discover minimum phonemic differences between two sequences of sounds only if they are associated with differences in the corresponding composite relations; otherwise, we ignore the differences in the two sequences of sounds and identify them as the same on the phonemic level. If therefore we can assume, that the knowledge of the significations of a set of complete utterances can be taken as empirically given to a linguist, we may go a step further and assume that we are able not only to identify the composite relations corresponding to two different utterances, but also to identify parts thereof. Accordingly, the differences in the significations of certain utterances can also be taken as available empirically. Thus, for instance, when considering the following utterances:

168

IRENA BELLERT

(a) John went to Warsaw

(a') John did not go to Warsaw

(b) Mary has read this book

(b') Mary has not read this book

we are able to identify a part of the composite relation associated with a' (or b') with that associated with a (or b). Consequently, we are able to identify the semantic change between the significations of a and a' as the same as that between b and b'. Hence, the notion of constant semantic change seems to be empirically justified and can serve as a sound basis in a heuristic approach to grammar construction. In fact the capability of identifying parts of significations was evidently a basis for postulating deep structures in Chomsky's generative transformational grammar. For, when considering the utterances: (a) Mary has a book

(a') Mary's book is red

(a") Mary's book was lost

(b) Jim has a cat

(b') Jim's cat is white

(b") Jim's cat was sick,

we are able to identify the composite relations associated with a (or b) with the parts of the composite relations associated with a' and a " (or b' and b " ) ; hence, we have empirical grounds for postulating a type of grammar in which this identity is displayed in the deep structure representation for different surface structures. If we can state that an utterance is correctly understood by a speaker, then obviously we have to assume that every structural relation is properly interpreted. In other words, it is obvious that a speaker has a tacit knowledge of the grammar which assigns a structural description to each sentence, as well as to a connected sequence of sentences in a discourse; owing to this tacit knowledge, he can properly understand any utterance or discourse. The task of a linguist is to proceed in the opposite direction and to postulate an abstract grammar on the basis of data derived from that aspect of his tacit knowledge which is not available empirically and not susceptible to direct investigation even by introspection. Thus we assume as empirically given a system which can be described as a finite set of complete utterances (those which a linguist chooses for investigation) put into correspondence with a set of semantic composite relations (significations), which can be compared and identified partly or as a whole in terms of invariant semantic relations. 4 Such a finite system corresponds exactly to what we may assume as empirically given to a linguist, and accordingly it represents what a speaker overtly knows about his language. Obviously we know nothing or very little about the grammar of a natural language as internalized by a speaker of that language. However, certain requirements have to be met by any adequate grammar as postulated by a linguist. By now it is perfectly clear that whatever else a grammar of a natural language might be, it must at least define all the sentences and only the sentences of that language together with their structural descriptions, and it must, therefore, be based on a generative system containing some rules which are recursively applicable. In fact, it should also define all 4

For more details, see I. Bellert and W. Zawadowski, op. cit.

A SEMANTIC APPROACH TO GRAMMAR CONSTRUCTION

169

the possible sequences of sentences (connected discourse), that is, it should assign structural descriptions to larger units than sentences. Suppose that a generative grammar is postulated according to Chomsky's recent proposal. A grammar would then contain a basis and a transformational component; the basis would generate deep structures of sentences on which transformations would operate to produce surface structures of sentences. The basis would consist of a categorial component which would be a simple phrase-structure grammar and a lexicon containing lexical entries each consisting of a set of phonological features, a set of syntactic features, and a set of semantic features. 5 The most important principle underlying Chomsky's latest proposal is that the semantic interpretation of sentences is based on the deep structure representation, that is, all the features relevant semantically are represented in the deep structure of a sentence. This principle is clearly the one which should be consistently pursued to its logical conclusion with the object of arriving at a grammar for which the following statement would hold: ANY TWO EXACT PARAPHRASES HAVE THE SAME DEEP STRUCTURE REPRESENTATION.

We do not say that exact paraphrases 8 have precisely the same meaning in the broadest sense of this term. WHAT we say is the same, but HOW we say it, is different; hence we assume that between paraphrases the distinction is one of HOW and never of WHAT. Thus, whatever the difference of meaning in its broadest sense may be between two exact paraphrases, it should be assigned to the corresponding transformational rules. If there is no difference of meaning in its narrowest sense (i.e., 'nuclear meaning' rather than 'cognitive meaning' which term may be misleading), then we should expect the deep structure representations to be the same. This is tantamount to saying that a transformational grammar should be revised from the point of view of the main principle postulated recently by Chomsky, since the deep structures in the generative transformational grammar are not deep enough to serve as an adequate basis for semantic interpretation. Compare, for instance, the following four paraphrases: (a) Somebody reprimanded John (b) John was reprimanded by somebody (c) Somebody gave John a reprimand (d) John received a reprimand from somebody. The transformational grammar would assign the same deep structure to (a) and (b), while it would assign a different deep structure to (c), and a different one again to (d). In the transformational grammar, the grammatical functions are given directly 6

Noam Chomsky, op. cit. • An exact paraphrase is evidently conceived as an equivalence relation (reflexive, symmetric, and transitive).

170

IRENA BELLERT

by the system for rewriting the rules that generate deep structures.7 Relations such as 'subject of' or 'object of' are defined by the rules of grammar. For instance, the relation 'subject of' would be assigned to somebody in (a), (b), and (c) but to John in (d). Such a treatment has a purely syntactic motivation, but it deprives the notion subject of a universal semantic interpretation. By taking a semantic approach, we might hope to be able to construct a grammar in which the notion subject will have a semantic interpretation which is invariable from utterance to utterance. If we assume that the interpretation of the notion subject is the 'performer of the action', then the grammar should assign the relation "subject of" to somebody in each of these sentences. If, on the other hand, the interpretation of the notion subject is assumed to be "that sentence constituent by means of which we identify the object of the discourse spoken about", 8 then it should be assigned to John in each of these sentences. In no interpretation, however, would the deep structure of (a), (b), (c), and (d) differ. In fact, the changes in the proposed formulations of Chomsky's transformational grammar confirm the necessity of pursuing a semantic approach in grammar construction. Whenever any inadequacies have been detected in the formulations of transformational grammar, it has been precisely because of the incompatibility of the deep structure representation with the notion of the semantic invariants, which should be taken as a basis for grammar construction. Typical examples are the changes proposed for the deep structure representation of interrogatives or imperatives. Klima's work on negation,9 in which he postulates the hypothetical element Neg in the deep structure representation, is a classical instance of the use of this sort of heuristic procedure in arriving at more adequate formulations. In order to proceed further in accordance with the main principle postulated by Chomsky, it is necessary to revise the deep structure representation with the aim of obtaining deep structures which are semantic par excellence and which are the same for any paraphrases, and not just for those referred to as grammatical paraphrases by transformational grammarians. Such a result would be compatible with the empirical data concerning the speaker's competent use of language. The number of paraphrases for longer utterances — especially if those involving changes of lexical items are included — is surprisingly large, often amounting to several thousands.10 In the light of this fact it would seem implausible to assume that a large quantity of sentences which belong to a paraphrastic set and which are recognized as paraphrases with great ease by fluent speakers have a different deep structure 7

See Noam Chomsky, op. cit. This interpretation of the notion subject seems much more plausible. • E. S. Klima, "Negation in English", The Structure of Language, ed. by J. A. Fodor and J. J. Katz (New Jersey, Prentice-Hall, 1964). 10 In the research carried out in the Transformations and Discourse Analysis Project at the University of Pennsylvania under the supervision of Professor Henry Hiz in 1964, we found 41,124 exact paraphrases for a seventeen-word sentence taken at random from a physics textbook. See I. Bellert and H. Hiz, "Paraphrastic Sets and Grammatical Analysis", T.D.A.P. Paper, No. 59 (1965). 8

A SEMANTIC APPROACH TO GRAMMAR CONSTRUCTION

171

representation, whereas only a part thereof (those reckoned as grammatical paraphrases in a transformational grammar) would have the same underlying deep structure, in so far as the ease with which speakers recognize them as equivalent does not seem to be dependent on such a division. Thus it seems strange to suppose that we should equate more easily the utterance John was reprimanded by somebody with its grammatical paraphrase Somebody reprimanded John than with John received a reprimand from somebody. It seems quite clear that no phrases or words and hardly any morphemes of a natural language are understood by speakers otherwise than as abbreviations standing for semantic composite relations. The ability to paraphrase which is evidenced by every fluent speaker, on ability which is not restricted to grammatical paraphrases (in the sense used by the generative transformational grammar) can be taken as evidence that such phrases, words, or morphemes are only the surface representations of certain semantic structures which are understood by speakers and which constitute a basis for identifying utterances as "saying the same". As has already been mentioned, phrases, words, or morphemes of each natural language are specific abbreviations corresponding to certain 'chosen' composite relations which are usually untranslatable into analogical abbreviations in another or in the same language, but are translatable in a larger context. Thus, when trying to explicate a sign, a native speaker or even a semanticist is biassed by his knowledge of other abbreviations, and what he really does is to reformulate the meaning in terms of different abbreviations; even though they may be more elementary than those he wanted to explicate, they are still abbreviations usually standing for complex semantic structures. In general, it seems that at present too little is known about the processes taking place in human cognition and the concept-forming ability of human beings to speak of postulating a set of elementary semantic relations which would underly the understanding of any linguistic sign. However, it may prove possible to dispense with such elementary semantic units in structural descriptions, as they may not be distinguishable linguistically, and to postulate some non-elementary but basic semantic units by means of which all other complex signs could be explicated. In terms of such basic units it would be possible to postulate much deeper structures than those proposed in the transformational grammar, of a kind which would be identical for any exact paraphrases. For example, the two utterances: John has sold a house to Jim and Jim has bought a house from John

172

IRENA BELLERT

should have the same abstract deep structure representation, in spite of the fact that the surface representation of the grammatical relations and the lexical items, which appear in fixed positions in the surface structure, are all different. Similarly, we cannot consider the utterance John is taller than George as being derived from a deep structure representation in which there occur the same phrase markers (or parts of a generalized phrase marker) as those which underly the utterances John is tall and George is tall, as obviously such a representation would not be compatible with the respective semantic interpretations. On the other hand, we may certainly use the utterance "John is taller than George" when speaking of two dwarfs, as this utterance is an exact paraphrase of "George is shorter than John". Much interesting and valuable work has already been done in such a direction by Russian and Polish linguists. However, they do not go so far as to postulate a generative system of rules which would be compatible with the results of their investigations. The results published by Russian linguists11 prove that it is possible to describe syntactical as well as lexical paraphrastic transformations in terms of what they call a 'deep structure paraphraser'. 12 But so far little has been said about the handling of relations other than paraphrastic ones, and it is difficult to see what place their paraphrasing device will have in a generative grammar, and what form of generative grammar is envisaged. Polish linguists13 are concerned with the collection of data concerning the deep semantic structure of natural languages which they assume to be universal, and they have arrived at certain generalizations in this respect. The results of their work seem to be of the most importance in the construction of an adequate generative grammar, but so far no research has been done with the object of postulating a proper form of generative grammar which would be compatible with the results and with the general aim of their inquiry into semantics. In conclusion, I wish to remind the reader that a grammar based on a purely semantic approach that has been advocated in this paper as a necessarily closer approximation to the adequacy of our descriptions of natural languages will constitute 11

A. K. 2olkovskij and J. A. MelSuk, op. tit.; J. A. MelSuk and A. K. 2olkovskij, "An Approach to Semantic Description of Natural Language", paper presented at the Conference on Semiotics, Kazimierz nad Wisl^, 1966; MaSinnyi Perevod i Prikladnaja Lingvistika, MGP IIJ, vypusk 8 (1964). 12 J. A. Meliuk and A. K. Zolkovskij, op. cit. ls A. Boguslawski, Pojgcie Liczebnika ijego morfologia w fezyku rosyjskim (Wroclaw, Ossolineum, 1966); A. Wierzbicka, "O Smyslovyx Ograniceniax dla Pravil Mnizestvennovo Cinteza pri Avtomaticeskom Perevode", Naucno-Texniceskaja Informatsja, No. 5 (Moscow, 1966).

A SEMANTIC APPROACH TO GRAMMAR CONSTRUCTION

173

a development along the lines postulated long ago by Roman Jakobson — the most ardent proponent of the semantic approach to grammar and of the thesis that linguistics is essentially a quest for meaning. DEPARTMENT OF GENERAL LINGUISTICS WARSAW UNIVERSITY

HREINN BENEDIKTSSON

THE PROTO-GERMANIC VOWEL SYSTEM*

E

V E R since the advent of the structural and functional approach in linguistic research it has been clear that the new principles, although they originated and matured in the synchronic field, were bound to become of prime importance for the study of linguistic history as well. These principles not only bring the historical problems into clearer perspective; they have also led to solutions of problems which had appeared to be unsolvable, and they have contributed towards more comprehensive solutions of other problems. Professor Roman Jakobson was, it is well known, a pioneer in this field also. 0.2. The present paper deals with a question that has been treated repeatedly, in prestructural publications as well as from the structural point of view. This is the question of the earliest Proto-Germanic vowel system and of its development in common Germanic times. A certain repetition of well known facts will be unavoidable, since the data are long established.1 But some further specification of the evidence, and an approach which differs to some extent from previous treatments, seem to make a new synthesis worthwhile. 1.1. The dislocation of the Indo-European resonant system — resulting ultimately, on the one hand, in the split of the syllabic nasals and liquids into clusters of vowel + consonant, and on the other hand, in the relegation of the remaining resonants to either the class of vowels or of consonants — no doubt began in pre-Germanic times.

* The writing of this paper was completed while I was at the University of California, Los Angeles, attending the 1966 Linguistic Institute of the Linguistic Society of America with a grant from the American Council of Learned Societies. 1 For the Germanic data referred to, see, e.g., W. Braune, Althochdeutsche Grammatik, 11th ed. (Tübingen, 1963); J. Schatz, Althochdeutsche Grammatik (Göttingen, 1927); J. H. Gallee, Altsächsische Grammatik (Halle-Leiden, 1910); K. Brunner, Altenglische Grammatik nach der angelsächsischen Grammatik von Eduard Sievers (Halle, Saale, 1942); A. Campbell, Old English Grammar (Oxford, 1959); W. Steller, Abriss der altfriesischen Grammatik (Halle, Saale, 1928); W. Braune, Gotische Grammatik, 16th ed. (Tübingen, 1961); A. Noreen, Altisländische und altnorwegische Grammatik, 4th ed. (Halle, Saale, 1923); A. Noreen, Altschwedische Grammatik (Halle, 1904).

175

THE PROTO-GERMANIC VOWEL SYSTEM

With this development the immediate pre-Germanic vowel system may be set up as follows: SHORT (I)

LONG

i

u

I

e

o

5

u

a

5 à

The immanent symmetry of this system, as revealed by the relative functional range of its distinctive features, may be visualized by the following diagram : e

i Long Compact Diffuse Grave-flat

—

—

+

—

o

u —

+

—

+ +

—

a

—

+

ï

è

ü

ö

ä

+

+

—

—

+

+

—

—

+ +

+

—

—

—

+ +

—

+

1.2. In each subsystem, short and long, a merger took place at the close of the preGermanic stage, viz. of short a + o > a and of long ä + ö > ö. The result was a simpler system, of equal functional balance, but with a distinctly different pattern of redundant features : SHORT

(II)

LONG

i

U

e

a

I è

Ü

ö

or, in terms of its distinctive features : i Long Diffuse Acute

—

+ +

e

u

a

—

—

—

—

+

—

—

—

+

î

+ + +

ë

ü

ö

+

+ +

+

—

—

—

+

—

Whether this system in fact ever existed, at any stage of pre- or Proto-Germanic development, will necessarily remain a matter of conjecture. But be that as it may, positing this stage provides a convenient means of elucidating the potential changes in distinctive feature structure from pre- to Proto-Germanic. 2.1. The next step was the augmentation of the long subsystem by the unit usually designated as e2. The origin of this unit is disputed, and not of immediate concern here. What matters in the present connection is that, in the earliest Germanic system that is attainable by Germanic comparative evidence alone (i.e., without Indo-European evidence), this unit was an integral part of the phonemic inventory, at least in the greatest part of the Germanic area, viz. in those varieties which correspond to the later West and North Germanic dialects; whether, on the other hand, this phoneme existed

176

HREINN BENEDIKTSSON

in pre-Gothic may be questioned.2 The introduction of e 2 had considerable effect on the distinctive feature structure not only of the long, but also of the short subsystem, as follows: SHORT

(III)

i e

LONG

u

i e2

Ü ö

a or, in terms of the distinctive features:

Long Compact Diffuse Grave-flat

i

1

u

«

—

—

—

—

—

+ -

(+) +

( - )

a

ï

ë2

ü

ö

êi

—

+

+

+

+

—

—

—

+

+ +

—

—

—

+

—

+ +

—

+

2.2. Comparing the distinctive features of this and the preceding stage II, we note that in stage II there is only one sonority feature, and that this is diffuseness: for one must assume that the actual degree of compactness in short a and long o was different, so that what they had in common was their nondiffuseness as opposed to u and u. But in the later stage the sonority features have become two (i.e., implying a three-way opposition of tongue height). And in this feature constellation, compactness must be the primary feature in relation to diffuseness, as shown by the redundancy pattern. 3 This relationship gradually manifested itself in the allophonic variation described in §3.2.

Similarly, the single tonality feature in stage II must have been acuteness, since the actual degree of gravity in short a and long o was probably different. But in the later stage we must assume the optimal feature (grave-flat), since there are no grounds for dividing them or preferring one to the other. The result of these changes in distinctive feature structure was that short e and long ' See, e.g., J. Kuryfowicz, "The Germanic Vowel System", Biuletyn polskiego towarzystwa j?zykoznawczego, XI (1952), 50-54. * The inverse order of the features would fail to bring out the redundancy pattern: Long Diffuse Compact Grave-flat

i

e

u

_ +

_ -

_ +

—

-

(-)

+

a —

—

+

I

ë*

ü

ö

ëi

+ +

+

+ +

+

+

— —

—

—

+

—

—

-

+

+

In this diagram plus-diffuse in u necessarily implies minus-compact, which need therefore not be marked. In the above diagram, on the other hand, minus-compact in u implied neither plus- nor minus-diffuse, and this feature must therefore be marked (as redundant).

THE PROTO-GERMANIC VOWEL SYSTEM

177

e2 became minimal correspondents within the quantity correlation instead of e : ex\ and so did a : ex instead of a : o. However, a further specification, in terms of the distinctive feature structure, is necessary: i and i may be termed DIRECT minimal correspondents within the quantity correlation, since they had all distinctive features in common except the one by which they were kept apart. On the other hand, u : u and e : e2 may be called INDIRECT minimal correspondents, since one feature common to each pair was distinctive only in one member of the pair, but redundant in the other (the short member). And a : ex were also indirect minimal correspondents, since one member of this pair ( e j had a redundant feature which the other lacked. In stage II, a : o were of course indirect minimal correspondents, but this does not appear in the diagram for the reason that the two features by which they probably differed (in addition to length), viz. compactness and flatness, were distinctive nowhere in the system. 3.1. As far as the long subsystem is concerned, the introduction of e2 and the consequent development of the vowel distinctive feature structure, as outlined above, apparently opened up two possibilities of further change: (1) Compact e 1 loses its redundant acuteness (and naturalness) and becomes neutral (with respect to tongue and lip position), viz. a. This happened in North and West Germanic. (2) The compactness feature disappears, and e 1 merges with its nearest neighbor, viz. e2 — which virtually means reverting to stage II. However, whether this was in reality a potential alternative may be questioned. The merger is usually assumed to have taken place in Gothic, but the result there, at any rate, was not identical with stage II. And the existence of e2 as a separate unit in the phonemic inventory of preGothic may indeed be questioned (§ 2.1). 3.2. In the short subsystem, as remarked above (§ 2.2), the rearrangement of the distinctive feature structure brought about a pattern of redundancy which manifested itself in two processes of development, or rather as a process with two aspects: (1) The diffuseness feature being redundant for u, a nondiffuse allophone ([o]) developed under certain conditions, the allophonic variation [u]: [o] corresponding to the phonemic distinction u : o within the quantity correlation. (2) The phonemic distinction of short i: e tended to develop into one of allophonic variation, the proportion [i]: [e] corresponding to long I: e2 as [u]: [o] to u : d. In terms of their distinctive feature structure this development was as follows: Let X be the conditions under which the allophone [o] appeared, and Y its complement, i.e., the conditions under which [u] remained. Leaving a out of consideration, we then find the three-way oppositions [i] Y: [e] Y: [u] Y and [i]X: [e]A": [oJA'. In the former case acuteness-naturalness (i.e., minus-grave-flat) is redundant for e. In the latter case, on the contrary, this feature is distinctive for e (opposing it to [o]), while it has become redundant for i. This may be shown by the following two diagrams, the former of which is identical with the short-vowel part of diagram III:

178

HREINN BENEDIKTSSON

i Compact (IVa) Diffuse Grave-flat

e

—

u

-

a

—

i Compact (IVb) Diffuse Grave-flat

+

+ - (+) -

(

-

)

+

e

-

-

o

a

-

+

+ - (-) (

-

)

-

+

In other words, these two diagrams are complementary: acuteness-naturalness is alternately redundant — and, conversely, alternately distinctive — in i and e. This may be illustrated by the following diagram :

Compact (IVc) Diffuse Grave-flat

i

e

+

-

-

-

u ~ o

a +

(+ ~ - )

(-) ~ (-)

+

At this stage the condition governing the alternate distinctiveness (or the alternate redundancy) of the acuteness-naturalness feature is diffuseness (plus or minus), and, conversely, the condition governing the redundancy of diffuseness (in [u ~ o]) is gravity-flatness. In other words, from the point of view of the total range of distinctive opposition in the noncompact vowels, diffuseness and gravity-flatness are partially complementary. In accordance with the principle of structural simplicity, the result was almost bound to be a leveling of the mutual distinctive function of these two features, so that the redundancy of the acuteness-naturalness feature was transferred to the diffuseness feature. The potential outcome of this transfer would be the reduction of the distinctive opposition i: e to one of subphonemic alternation, parallel to [u] ~ [o], the three-way oppositions mentioned above then being reduced to twoway oppositions, [i] Y : [u] Y and [e]X: [o]Z. The total result, including the long subsystem under alternative (1) mentioned in §3.1, may be diagrammed as follows: SHORT

LONG

1 ë

(V)

Ü Ö ä

or, in terms of its distinctive features: i ~ e Long Compact Diffuse Grave-flat

u ~ o

a

—

—

—

—

—

+

(+ ~ - )

(+ ~ - ) +

I

ë

û

ö

à

+ +

+

+

+

+

—

—

—

—

+

—

—

—

+ +

—

+

4.1. The development sketched in the preceding section was only a potential one.

179

THE PROTO-GERMANIC VOWEL SYSTEM

The extent to which it manifested itself, or whether its final stage was reached in any variety of Proto-Germanic, has been a matter of debate. And at any rate, the course by which it proceeded in its details was different in the different varieties of ProtoGermanic. On the available comparative evidence, this development appears to have led to a bifurcation of the Proto-Germanic area: on the one hand, pre-Gothic; on the other, pre-West and pre-North Germanic. The difference between the two may have been in the extent to which the above development took place, but it was mainly in the course by which it proceeded. The Gothic situation is too well known to need to be outlined here. The development in pre-West and in pre-North Germanic may, in spite of important differences of detail, be subsumed into a single process, both because of its intrinsic similarity and also because of its common differences from the Gothic development. The West and North Germanic development is generally taken to represent the Proto-Germanic changes proper. But, in fact, there is very little or no evidence to suggest that this development took place in pre-Gothic. The principle of simplicity therefore requires it to be considered a characteristic of the West and the North; it will be examined in § 5. 4.2. The dominating factor of the potential development of the vowel system sketched in § 3.2 was the interrelationship of its distinctive features, including the difference in their relative functional range. In this sense, therefore, the development may be described as a tendency towards functional balance, producing a fully symmetrical short subsystem and a balanced correspondence between the two subsystems within the quantity correlation (stage V). The same degree of functional balance and a more perfect (viz. a one-to-one) correspondence between the short and long subsystems might have resulted from a different development of the short vowel structures in stage IV (see especially IVc). The direction of the leveling of the mutual distinctive function of the acutenessnaturalness and the diffuseness features between stages IV and V was towards the transfer of the redundancy of the former to the latter. The opposite direction of the leveling, viz. towards the transfer of the distinctive function of acuteness-naturalness to diffuseness, would have resulted in the following system: SHORT

(VI)

LONG

i

X ë

U

e

o a

Ü ö ä

or, in terms of its distinctive features: i Long Compact Diffuse Flat

e

u

o

—

—

—

—

—

—

—

—

+ —

— —

+ +

—

+

a —

+

ï

ë

ü

ö

+

+

+

+

—

—

—

—

+ —

+ +

—

+

a

+ +

180

HREINN BENEDIKTSSON

However, this development, though perfectly consistent with the distinctive feature structures of the preceding stages, would have meant a phonemic split of u into u and o, a split which was unmotivated at this stage. 4.3. But later, it is well known, such a split did take place, resulting not only in the phonemicization of u : o, but also in the rephonemicization of i: e (or in the reinforcement of this opposition, if it was never lost entirely). The ultimate motivation for this split was the coalescence of vowels in unstressed syllables, or, in terms of the formulas in § 3.2: [ i ] 7 : [e]X and [ u ] 7 : [o]X became, by the merger of X + Y> Z, iZ : eZ and uZ : oZ; compared with X and Y, Z stands for [a] or zero. Stage VI, which may be called Common West and North Germanic in the sense that it was the starting-point for the further development in both branches (umlaut, etc.), is practically identical with the immediate pre-Germanic vowel system set up in I above. However, the fact that the two pairs of vowel diagrams (in I and VI) are the same must not be permitted to mislead. The distinctive tonality feature is decidedly different in the two systems; grave-flat in I, only flat in VI (with gravity redundant). This difference manifested itself in the allophonic variation which was the essence of the changes commonly referred to as umlaut, in particular the development of acute allophones of u, o, a, short and long (/-umlaut). (Besides, of course, the historical correspondence between the two systems was not one-to-one.) 5.1. The development outlined in § 3.2, from the system represented by diagram III, through IV, towards the potential system shown in diagram V, may be said to have been reversed, in pre-West and pre-North Germanic, by the merger of unstressed vowels referred to in § 4.3 — or rather by the consequences of this merger for the reinterpretation of phonemic relationships in stressed syllables. As outlined in diagram V, the outcome of this development would have been the reduction of the distinctive opposition i: e to one of subphonemic variation. Whether, however, this final stage had been reached when the reversal of the development, which led to the system in diagram VI, set in, has been a matter of debate. On the one hand, it has been maintained that [i] and [e] became members of one and the same phoneme, and that their relationship was parallel to that between [u] and [o]; 4 on the other hand, it has been claimed that even though these two phonemes merged in a number of positions, their distinctive opposition was maintained in some environments.5 In this section the data will be summarized and clarified, and some additional evidence brought forward. 5.2. The reduction of the distinctive opposition i: e to one of subphonemic variation, parallel to [u]: [o], would imply a change of e > i (i.e., a merger e + i > i) under conditions Y, and of i > e (i.e., e + i > e) under conditions A'(see § 3.2). There is no

4

See J. W. Marchand, "Germanic Short *i and *e: Two Phonemes or One?", Language, XXXIII (1957), 346-354. ® See, especially, W. F. Twaddell, "The Prehistoric Germanic Short Syllabics", Language, XXIV (1948), 139-151. See also Kurylowicz, Biuletyn, XI, 50-54.

THE PROTO-GERMANIC VOWEL SYSTEM

181

doubt that these two changes took place. What has been a matter of debate is only whether they took place in all T's and X's respectively, or only in some, i.e., whether they together covered the total set of relevant environments, or only a proper subset of this total set. The relevant conditions, Y and X, are, it is well known, generally considered to have been of two kinds: (A) The vowel of the following (unstressed) syllable, and (B) the consonant(s) immediately following the stressed vowel. The unstressed vowel system implied in (A) may be regarded as identical with the system represented by diagram II above, with one important difference, viz. the merger of

e+ i> i which is commonly regarded as general and unconditional in this position (a rule which, however, requires some further specification; see § 5.7). To this are added the three diphthongs iu, ai, au, whose historical status will be examined in § 6.1. The total set of postvocalic consonants and consonant clusters, (B), may be divided into two, M and N, viz. those which do or do not, respectively, form part of the relevant environment of the stressed vowel changes involved. The maximum set of combinations may then be represented as follows (on the parentheses, see § 5.7): l

I iu u M (VII)

+

Ü

+ N

ë

ai a ö

au 5.3. Diagram VII permits the formulation of the three principal and well-known conditions of merger of e + i > i, viz. (a) before a second syllable with i, f, or iu, (b) before clusters of consonant + _/,6 or (c) before clusters of nasal + consonant ((b) and (c) thus form a subset of M), 7 • Morphophonemic clusters of more than one cons. + j were realized as combinations in (a) according to the well-known rule of alternation i : j dependent upon the length of the preceding syllable (type Goth, hairdeis). Therefore, in Figure 1 below, (b) and (c) do not intersect; *-ndja-y *-ndjiwere realized as *-ndia-, *-ndi-.

182

HREINN BENEDIKTSSON

and the part of the diagram that is included by the outer parentheses contains the principal condition of u > [o], viz. (d) before a second syllable with a, d, or au. It will be noticed immediately that the three sets of combinations implicit in (a), (b), and (c) respectively — all proper subsets of (VII) — are mutually noncomparable, i.e., each contains combinations which do not belong to the others. 8 But at the same time, the first set, (a), intersects with each of the other two, there being combinations containing elements from both (a) and (b) or (c), e.g., 3rd sg. *bedjip 'bids', *bendip 'binds'. What is noteworthy about the three sets of combinations involved in the merger of e + i > i — precisely because they are noncomparable — is that when we compare them to the conditions for the allophonic variation [ u ] : [o], we observe that they form a proper subset of the complement of the conditions of u > [o]; that is, they are AMONG 7

the environments in which u remained, 9 as follows:

Twaddell doubts that these changes were in fact mergers (of e with i), since, in his opinion, i did not occur in these positions; he says (p. 141): "Distributionally, they [i.e., i and e] were partially in direct contrast (both occurred before obstruents), partially in complementary distribution (/e/ occurred before resonants, /i/ did not). In one other respect, /i/ and /e/ were in complementary distribution: historically, no words in late Indo-European had /i/ in two succeeding syllables, nor did /i/ occur before /j/ of a following syllable." Therefore, in his opinion, the two changes rather consisted in the development of high allophones of e and of a subsequent redistribution of these allophones from e to i. This state of affairs cannot, however, have existed in Proto-Germanic. As regards change (a) we must note that it took place also before an unstressed i that derived from PIE e by the merger described above, cf., e.g., PIE *bheresi > OHG biris, etc.; PG *mekela- > ON mikell, OHG michil, etc. Therefore, even if i had not occurred before an unstressed / in pre-Germanic, the merger of unstressed e with i would have introduced the opposition i: e into this position, cf. on the one hand, verbs of the type Skr. vidati, Goth, digan, and on the other, thematic e-grade verbs; e.g., 3rd sg. *wigip 'kills' vs. *wegip 'carries'. Besides, /-stems like ON skripr, OHG scrit, OS skridi, OE scride; ON gripr, OHG grif, OE gripe \ Goth, wlits, OS wliti, OE wlite (but ON litr, «-stem); OHG biz, OS biti, OE bite (but ON bit, a-stem), or a ya-stem like ON gil — all of which belong to the first ablaut series — show the occurrence in early Germanic of original i before i and j. And even for 'late Indo-European', in view of such forms as Skr. loc. dim, Gr. 8ii, or Lat. sitis, Skr. ksiti-, Gr. tpOicng, the above assumption is not tenable. Before a single resonant we also find examples like ON OHG gil, ON OE gin, ON skin, all belonging to the first ablaut series. And in verbs of the type Skr. bhinadmi, bhindanti, i would have occurred before a nasal + consonant, provided that any such verb was preserved in Proto-Germanic. One of the possible examples is Goth, sigqan, ON sokkua, if this verb is related to Skr. siiicdti; this etymology has, however, been challenged, and the same is true for the other potential cases of this type. At the same time it may quite well be true that i was relatively less frequent than e in these positions, and before r or / + consonant (type *werpan) it may not have occurred at all. But if so, from the Germanic point of view, this was no doubt an accidental distributional gap, and did not represent the systemic feature of neutralization of the opposition i : e. * The same was true of the sets of combinations implicit in (A) and (B) above, at least if, in diagram VII, we replace stressed e j by V (V = the set of all stressed vowels). On the one hand, there were such combinations as pret. *band 'bound' not belonging to (A); on the other hand, such combinations as *buan not belonging to (B). 9 This was pointed out by Kurylowicz, who says (p. 52): "As regards d, it originates from Germanic u under conditions representing just the negative counterpart of [the conditions of e > /]." More precisely, the conditions of u > [o] represent only PART of the 'negative counterpart' of the conditions of e > i.

THE PROTO-GERMANIC VOWEL SYSTEM

183

Fig. 1

Since the sets of combinations in (b) and (c) interersected with (a), and since (b) and (c) as well as (a) were proper subsets of (VII), (b) and (c) intersected also with the complement of (a) ( = (a)'). In particular, it follows from (VII) that (b) and (c) intersected with the proper subset of (a)' containing the conditions for u > [o] (viz. (d)), as shown by Figure 1 (in which the vowel notations are slightly abbreviated compared with the formulation in (a) - (d) above; Cx = one cons.; C" = one or more conss.). This implies that not only does u remain under conditions (a), but its change to [o] was also prevented in the intersection of (b) or (c) and (d), cf., e.g., OHG joh : hund < *juka : *htinda, ON lofa : dylia (with y < u) < *lobon : *duljan. It can of course hardly be a coincidence that the same three noncomparable sets of environments are among the conditioning factors of change for e and u, in a positive and negative way respectively. On the contrary, this circumstance shows quite clearly the mutual dependency of the merger of e with i and of the development of the allophonic variation of u, and thus provides the tangible evidence for the structural tendency outlined in § 3.2. 5.4. The dephonemicization of the opposition i: e would, in accordance with the change u > [o] under (d) above, presuppose a change i > e (i.e., a merger e + i > e)

184

HREINN BENEDIKTSSON

under the same conditions. There are, it is well known, distinct traces of this change, but only in a very few words is it general in the entire Germanic area (*wera- < PIE *wiro-, *nesta- < PIE *nizdo-). Otherwise, it occurs sporadically, though quite distinctly. In some cases the change is to be found perhaps only in one dialect, most frequently in Old High German, e.g., OHG quec, but OE cwic, ON kuikr; OHG leben, but OE libban, lifgan, ON Ufa. In other cases doublets exist, e.g., OHG skif: skef, OE spic : spec, ON stige : stege, OSw. slipi: slxpi. Finally, in some cases, the change is absent in all dialects, e.g. OHG OE fisc, ON fiskr; OHG biben, OE bifian, ON bifa. The most important group of forms of the last kind is the pret. part, of 1st class strong verbs, as opposed to 2nd class part, with o < u (e.g., OHG gistigan: gibogan, ON stigenn: bogenri). The commonly accepted explanation of this vacillation is that the change i > e led to an alternation between i and e in each paradigm, e.g., nom.-acc. sg. *skepa, but dat. *skipoi, or nom. sg. * stige, but gen. *steganiz, and that later (i.e., after the rephonemicization of i: e, if this opposition ever disappeared) leveling took place in each paradigm in favor of one or the other of the two stem forms. The direction of the leveling was often different in the various Germanic dialects, and occasionally it led to the creation of doublets. As a matter of fact, the conditions for an alternation of this kind were probably present in most, or even all, of the paradigms concerned. But, of course, the fact alone that these conditions were present does not prove that the alternation arose nor that leveling took place; it at most makes the assumption permissible. The difficulty is that the intermediate stage, with the phonologically regular alternation, is attested nowhere. The corresponding change u > [o] is much more regular. Examples like *joka( < PIE *yugo-), with o in the entire area (OHG joh, OE geoc, ON ok), are much more numerous than those with e < i. Nevertheless, the change is far from being universal in the Germanic dialects. Here also we find a difference from one dialect to another, e.g., OHG fol, but OE full, ON fullr (OSw .folder beside fulder); OHG wolf, but OE wulf ON ulfr (-olfr in compound names, e.g., Piopolfr); and doublets, e.g., OHG furhten \forhten, pret. wurhta : worhta, plur. lucher (luhhir) : lochir, OE spura : spora, ON sunr : sonr, gull: goll, stufa : stofa, OSw. brut: brot, kuna : kona. The change is even practically absent in all dialects in at least one word, viz. OHG ON sumar, OE sumor (but OSw. somar beside sumar). It is clear that in view of the overwhelming evidence there are no grounds for denying the occurrence of the change u > [o], as has been done in the case of i > e.10 The observable deviations — whether they appear in the form of doublets or as cases of apparent nonoccurrence of the change — must then be accounted for by the morphological environment, i.e., as cases of analogical leveling. The difference here is that traces of the intermediate stage, with the phonologically regular alternation in each paradigm, are probably preserved, at least in early (Runic) Proto-Nordic, cf. acc.sg. horna (Gallehus), but dat.sg. -kurne (Tjurko); 1st sg.pret. worahto (Tune), but 10

See O. Bremer, "Die germanische 'Brechung'", Indogermanische Forschungen, XXVI (1909), 149.

THE PROTO-GERMANIC VOWEL SYSTEM

185

3rd sg. wurte (Tjurkö). Possibly, traces are to be found even in 13th-14th-century Old Swedish, in forms like nom.-acc.sg. lof, but dat. luui; nom. sg. kona, but obi. kunu, which are regular in one manuscript especially (Cod. Holm. B 50 of the Östgöta-Law).11 In view of the extensive parallelism between i: e and u : o, the fact that the observable alternation of u : o must in part be related to the morphological environment is of course a strong argument in favor of explaining the distribution of i:e in the same way. Therefore, even though i has remained more frequently than u, and though no traces are preserved of the intermediate stage of the alternation i : e as of u: o, analogy is the most likely explanation for i: e, too. This must be true of the 1st class pret. part, of the type of OHG gistigan as well, for the following two reasons. On the one hand, there are forms with e, viz. OHG wesan 'rotten' (as opposed to ON uisenn; cf. also OHG wesanen, ON uisna), and ON bepenn (from bipa, OHG bitan, etc.). Indirect evidence of the same e is furnished by the 4th and 5th class verbs which were originally 1st class verbs with zero-grade present, e.g. OHG stehhan, OS stekan < *stik-, cf. Lat. instigö; ON uega 'kill' (ONorw. also uiga), pret. ud, part. uegenn, OHG -wehan (pres. only), as opposed to Goth, weihan, OE wigan, OHG pret. part, -wigan.1* On the other hand, although o is prevalent in the pret. part, of 2nd, 3rd, and 4th class verbs (except before nasal + consonant), significant traces of u are to be found, especially in North Germanic. In the earliest East Nordic (Swedish and Danish) records there is a fluctuation between u and o in a great many of these forms, and in several cases, in modern times, the form with u has prevailed. The old fluctuation was, in part at least, dialectal; broadly speaking, while the forms with u are the more general, especially in Old Swedish, the forms with o seem rather to have been characteristic of the more westerly dialects of East Nordic. Thus, for instance, the abovementioned mid-14th-century Cod. Holm. B 50 of the Östgöta-Law has skurin, stulin, burin, rutna, and the early-14th-century Cod. Holm. B 53 of the Södermanna-Law has stulin, burin, rutna, but skorin. In contradistinction to these, the late-13th-century Cod. Holm. B 59 of the Older Västgöta-Law, the mid-14th-century Cod. Holm. B 58 of the Younger Västgöta-Law, and the early-14th-century Cod. Ups. L 12 of the Upplands-Law have stolin (also stulin in B 58), skorin, borin, rotin, All these five manuscripts have invariably u, for instance, in bupin, brutin, lukin, lutit, guldin, hulpit, burghit, hurvit, vurpit (none of the manuscripts has examples of all these forms). Further, the early-14th-century Cod. Holm. B 57 of the Västmanna-Law and the mid-14th-century Cod. Holm. B 54 of the Dala-Law have bopin, brotit, lokit, stolin, borin, as well as bupin, brutin, skurin, stulit, but only, for instance, guldin, hulpit, 11 True, it is not certain that this alternation is a remnant of the early Germanic change; it is conceivably the result of a later Swedish development; see E. Wessen, Svensk spräkhistoria, I: Ljudlära och ordböjningslära, 3rd ed. (Stockholm, 1951), 8 and 57-58. Besides, in Cod. Holm. B 50, this alternation is absent from many paradigms in which it must have arisen; thus, this manuscript has invariably, for instance, bup, brut, bor p. 12 W. Streitberg, Die Erforschung der indogermanischen Sprachen, II: Germanisch (= Grundriss der indogermanischen Sprach- und Altertumskunde, II) (Berlin-Leipzig, 1936), 363.

186

HREINN BENEDIKTSSON

burghit. In Old Danish we find, for example, bupit, brutin in the Scanic Law, both buthzn and bothzn in the Zealandic Laws, but only bothaen, brotxn in the Jutlandic Law. All three, on the other hand, have stolxn, bormn, skorxn, etc., and all three (in so far as examples occur) have (v)urpmn, guldset, burghxt.l3 Modern Standard Swedish has regularly, for instance, b(j)uden, brüten, skjuten, buren, skuren, stulen, and Modern Standard Danish has budt, brudt, skudt, but bäret, skäret, stjälet. The probable explanation is that of analogical leveling of the morphophonemic alternation between the pret. plur. stem (e.g., hulpum) and thepret. part. (holpin), on the basis, on the one hand, of stigum : stigin, etc., and on the other, of sunkum : sunkin, etc. From the 3rd class participles the resulting fluctuation between o and u spread to 2nd class part, like bopin, brotin, to 4th class part, like borin, stolin, and to isolated adjectives (original participles) like rotin, lopin. This leveling started in the East in prehistoric times and spread westward with gradually decreasing force. On the basis of the above sources, its center was in Östergötland and Södermanland, where it affected practically every participle. Further to the North and West, in Uppland, Västergötland, and Scania, it was virtually restricted to 2nd and 3rd class participles, while 4th class forms {stolin, etc.) and isolated words {rotin, etc.) mostly retained the o. The manuscripts of the Vastmanna- and Dala-Laws show slight deviations from this pattern (with fluctuation both in 2nd and 4th class forms). Further to the West, in Zealand, the leveling is regular only in 3rd class verbs, while there is fluctuation in the 2nd class. And, finally, in the farthest West, in Jutland, it did not extend beyond 3rd class verbs.14 This agrees well with our general view of the Nordic speaking area, with the East as the center of innovations.15 Cases of fluctuation between u and o extending beyond East Nordic are not numerous. The best known are the verbs ON nema, koma. In Old Swedish we find both numin, kumin and nomin, komin; in Old Danish numin, kumin; in Old Norse komenn, but numenn (ONorw. also nomenn); OHG ginoman and quoman (and, sporadically, kuman) beside the regular queman; OS ginoman (and also ginuman), but gikuman; OE numen, cumen. All these examples show that the pret. participles were liable to change by analogy with other verbal forms,16 and this circumstance makes the assumption quite 13

See, in addition to the reference in fn. 14, W. Spickermann, Das Verbum in den seeländischen Gesetzen (Weida (Thür.), 1912), 21-22. 14 See, especially, A. Kock, "Der a-umlaut und der Wechsel der endvocale a : i (e) in den altnordischen sprachen", Beiträge zur geschichte der deutschen spräche und literatur, XXIII (1898), 484-554, in particular 503-506. The principal objection (by Noreen) against this explanation is the apparent increase of participles with o in Middle Swedish. At the same time, however, Noreen suggested that these o-forms were due to Middle Low German influence; see A. Noreen, Altschwedische Grammatik (Halle, 1904), 430. 16 See, for instance, the present writer, "Some Aspects of Nordic Umlaut and Breaking", Language, XXXIX (1963), 425-426. 16 Later, we find more cases of analogical change in the pret.part. Thus, in Swedish, the part, was in many cases changed on the analogy of the pres. stem. This appears very clearly already in Old Swedish in 5th class verbs. In verbs like Ixsa, which had only one stem form in the pres., the part.

THE PROTO-GERMANIC VOWEL SYSTEM

187

plausible that such analogical leveling took place in the 1st class participles as well, even though the process cannot be followed in any detail. It is especially noteworthy that the forms in which the e is retained are similar to those most resistant to the analogical replacement of o by u in East Nordic: on the one hand, an isolated form, OHG wesan; on the other, a form, ON bepenn, homonymous with, and therefore subject to the retentive influence of, another participle, bepenn from bipia.17 But, at the same time, the question of course remains why the analogical leveling was so much earlier, and so much more thorough, in the 1st than in the 2nd class participles. In view of our scanty knowledge of the nature and limitation of analogy in general, this question is likely to remain unanswered. But it seems reasonable to connect this difference with the fact that, in the 1st class, the morphophonemic alternation between the pret. plur. stem and the pret. part, stem was, as will appear below (§ 5.8), one of two distinct phonemes at a time when, in the 2nd class, it was still one of two positionally conditioned variants only. 5.5.1. Before a second, unstressed, syllable with u or u, both i and u were preserved, as appears, for example, in the pret. plur. of 1st and 2nd class verbs, OHG stigum, lugum, etc. The dephonemicization of the opposition i: e would therefore presuppose the change e > i (i.e., the merger of e + i > i) before u. Whether this change in fact occurred is a matter of dispute. On the one hand, there are clear examples of its nonoccurrence, especially in w-stem nouns. In some, the change is absent throughout the entire Germanic area, e.g., ON migpr (< *medu-), OHG meto (earlier, presumably, *metu), OE medu. In others, the change is present in a part of the area, mainly in Old High German (and, in some cases, in Old Saxon), e.g., ON figl- (< *felu-), OE feolu, but OHG OSfilu\ ON higrtr (< *herut-\ OE heorot, but OHG hiruz, OS hirof, ON fe (< *fehu), OE feoh (a-stem), OS fehu (also fihu), but OHG fihu (later feho, with little doubt due to the analogy of gen. fehes).1" On the other hand, there are cases in which the change has been posited for the entire, or at least most of the Germanic area. The most commonly quoted example is ON sipr, Goth, sidus, OE OS sidu, OHG situ < *sedu-, cf. Gr. '¿Qoq. This etymology has, however, been challenged. Further examples are: *sibun 'seven' (OHG sibun, OS sibun, OE siofon (early sifuri); ON siau is problematic, and Goth, sibun is, of course, ambiguous) < *sebun, Lat. septem, etc.; *miluk- 'milk' (OHG miluh, OS miluk, OE mioloc; Goth, miluks is

has only one form, Isesin. But in verbs which had more than one stem form in the pres., like gxta, gizta, gita (the last due to the influence of pres.sg. giter), we also find more than one part, stem, gstin, gixtin, gitin. In Modern Swedish we find, for instance, bjuden, skjuten (for buden, skuten) on the analogy of inf. bjuda, skjuta. 17 In Old Swedish, on the other hand, both these participles have i (bipin), one probably under the influence of the pres. stem {bipia), like, for instance, sitin ( O N setenn) from sitia (see the preceding footnote); the other on the analogy of other 1st class verbs. 18 Long u was no doubt rare in this position. The principal example — which, however, is present only in North Germanic and German — is in the sing. obi. cases of fem. «-stems (OHG zungun, OS tungun, O N tungo), whatever the origin of it may be here. In Nordic, this u accounts for such doublets as stufa : stofa, kuna : kona mentioned above (§ 5. 4). Cf. also *swegruz below.

188

HREINN BENEDIKTSSON

ambiguous, while O N miolkmay contain a different stem form, *melk-, cf. Gr. djie^yco, O H G melkan, etc., or, together with OFr. melok, derive from *meluk-, without the change e > i); *swigrü- 'mother-in-law' (OHG swigar; OE sweger has e, while O N suqra, and of course Goth, swaihro, are ambiguous) < *swegrü-, cf. Skr. svasrtis.19 On the assumption of the change e > i before u, the numerous exceptions must be explained as the result of analogical leveling. But, although the prerequisites for analogical leveling may have been present in each case, an explanation along these lines is beset with considerable difficulties no less than in the cases of i > e and u > [o]. The following points may be advanced in favor of analogy. In Old High German, Otfrid has regularly nom.-acc.sg./z/iw, but gen.fehes, while Notker has feho, fehes, representing analogical leveling. The «-stem *fripu- 'peace' (OHG fridu, OS frithu, O E f r i d u , O N f r i p r ) , with original /(PIE *pri-tu-), has e in O F r . f r e t h u , no doubt due to the analogy of gen.sg. *frepauz (with e in accordance with § 5.4). Similarly, Old Norse has -frepr in compounds (e.g., Hallfrepr); -frepr: fripr is parallel to -olfr:ulfr (§ 5.4). Finally, as regards the «-stems in particular, it is worth observing that, in Old Norse, when analogical leveling took place (with or without transfer to other stem classes), it was the gen.sg. stem form that prevailed. Thus, hgttr, hgtt, hette, hattar was replaced by hattr, hatts (a-stem); nom.sg. togr and tegr coexist, the latter due to the analogy of gen. tegar. The principal point against analogy, in the «-stems, is the vowel alternation in Old Norse. If, in a paradigm like verpr, verp, virpe, verpar 'meal', the e of the nom. and acc. verpr, verp had replaced the regular i already in ProtoNordic, under the influence of gen. verpar, one would have expected the same replacement to have taken place at the same time in the dative. But, in fact, virpe was not replaced by verpe until much later (it is attested in the Edda as well as in the Grägäs); similarly, Nigrpr, Nigrp, Nirpe, Niarpar, and others, in which the analogical leveling would have to have preceded the 'breaking' of e > ig or ia. Besides, the practically complete absence of doublets in Nordic (with the exception of -frepr : fripr) is surprising if the attested paradigms are the result of analogical leveling; no doublets like figrpr : *firpr (parallel to sonr : sunr) or reka : *rika, iata: *ita (parallel to koria : kuna) exist. 5.5.2. In some of the principal examples, an additional complication is involved. The common formulation of all the above changes is tacitly adapted to bisyllabic words only, the implication apparently being that, in forms of more than two syllables, only the syllable immediately following the stressed syllable was part of the relevant environment. This was in fact also done in § 5.2 above. But, for instance, the later development of forms like O H G menigi, as compared with manag, or O N ople < *apulia, in which the vowel of the third syllable was decisive, indicates that the third syllable

19

Whether *silubra 'silver' belongs here ( < *selubrd) is less certain; see, e.g., H. Weyhe, "Beiträge zur westgermanischen grammatik", Beiträge zur geschickte der deutschen spräche und literatur, XXXI (1906), 64-66; S. Feist, Vergleichendes Wörterbuch der gotischen Sprache, 3rd ed. (Leiden, 1939), 421. See also below.

THE PROTO-GERMANIC VOWEL SYSTEM

189

may have been part of the relevant environment of the above Proto-Germanic changes as well. Therefore, Kögel assumed, for Old High German at least, that in the combination eNuCA ( A = any of the vowels conditioning the changes u > [o] and / > e;C = any consonant(s); for N, see § 5.2) the e remained, while it became i before any other vowel in the third syllable.20 In accordance with change (d) in § 5.2, Marchand changed this formula to eNoCA.21 Conversely, Weyhe posited the change e > i in the combination eNuCi (but at the same time he maintained that bisyllabic eNu remained). 22 The difficulty here is that the relevant examples are too few. The theoretically possible combinations may be deduced from diagram VII by simply repeating what is included in the outer parentheses (while leaving out the inner parentheses). But only a very small part of the theoretically deduceable combinations can be exemplified. Nevertheless, Kögel's rule might account for some of the exceptions. Among these are OHG swehur 'father-in-law', OE swehor i before i.2i Thus, PN (Runic) erilaR (Kragehul, Lindholm, etc.) would be the phonologically regular form, while the later irilaR (Veblungsnes, By; provided that this is a genuine form 25 ) would be due to the analogy of dat. */n7e;26 similarly, O N gen.pl. Uenpa (beside Uinpa) < *weniöö as opposed to nom.pl. Uinpr< *wenidiz (cf. Lat. Uenedi, Gr. OueveSoi); OSw. skialnaper, ODan. skialnath 'difference' < *skelinööu-, beside ON skilnapr (due to the analogy of skilia).27 But other examples offer greater difficulties. Thus, Marchand explains the OE (North.) by-form seofo 'seven' (in the Lindisfarne Gospels, = L), compared with siofu in the Rushworth Manuscript ( = R2) and sifun, siofon in other dialects, as due to the influence of inflected forms with a in the third syllable, *sebona- (L seofona, etc.; R 2 siofunse, siofune), as opposed to the usual /-declension. 20 R. Kögel, "Zu den reduplicierten praeterita", Beiträge zur geschickte der deutschen spräche und literatur, XVI (1892), 501. 21 Marchand, Language, XXXIII, 351. 22 Weyhe, Beiträge, XXXI, 43-78. These formulas represent an interpretation by the present writer of Kögel's and Weyhe's statements and of Marchand's formula. 23 OE sweger would then either have to be derived from *swegurd, i.e., be related to OHG swigar as Gr. ¿Kupa to Skr. svasru-, or, which is more likely, owe its e to the influence of swehor. 24 See A. Kock, Umlaut und Brechung im Altschwedischen (Lund-Leipzig, 1911-1916), 55-56. 25 See, e.g., W. Krause, Runeninschriften im älteren Futhark (= Schriften der Königsberger Gelehrten Gesellschaft, 13:4) (Halle, Saale, 1937), 67 and 136-137. 26 In the same way, for instance, nom. *mikilaz (OHG michil, etc.) would have to be explained as due to analogy (instead of *mekilaz). 27 Whether O N pret. hlepa, sepa (to hlyia, *syia, cf. Goth, siujan) < *hlewidö, *sewidö, as opposed to skilpa < *skilidö < *skelidö, belong here is less certain. If they do, the e of *hlewidö would be regular in the 1st sg., while the i of *skilidö would be due to the influence of the other forms (3rd sg. *skilide, etc.). But it is perhaps also conceivable that the e represents the regular lengthening of i, with concomitant lowering conditioned by the w (as, for instance, before h). At any rate, in spite of OHG soum, OE seam, O N saumr, it is by no means certain that *syia had Proto-Germanic e, and not i, cf. Skr. stvyati, syütä-.

190

HREINN BENEDIKTSSON

But Marchand's objection to Weyhe's explanation (by an alternation between *sebun and *sibuni) — viz. that a form *sibuni cannot be posited for Proto-Germanic, since the /-declension of the numerals is only West Germanic and, in part, Gothic 28 — may in fact be turned against his own: there are no grounds either for assuming an fir-declension for Proto-Germanic, since the North Germanic as well as the preGermanic evidence indicates that the numerals 5-19 were indeclinable in ProtoGermanic. Further, the consonant stem inflection oi*miluk- (cf. ON gen.sg. miolkr), nom. *miluks, acc. *miluku, dat. *miluki, gen. *milukiz, leaves no room for a by-form *meluk-; for, in spite of OE dat. meolcum, OFr. melokon (which is probably an old instr. *melukum, sometimes treated as a dat.pl. in Old English), it is quite unlikely that this stem ever had pluralforms. And the stemform *meloka- posited by Marchand is without foundation. Similarly, provided that j between the second and third vowel had the same effect as i in the third syllable, *fergunjo-, *fergunja- (Celt. *(p)erkunia, Goth, fairguni, cf. Lith. perktinas, etc.) should have left no room for the e underlying ON Figrgyn, but only for the i in OE firgin-, OHG Firgunnea. It should be observed that although Weyhe's formula accounts for some of these forms (e.g., *meluk-, which would be regular in the nom. and acc., while *miluk- would have arisen in the dat. and gen.), it does not account for Figrgyn. A crucial test of Kogel's formula, as far as the possible dephonemicization of i : e is concerned, would be provided by an example of the type iNuCA, with original i in the first syllable, which, if Kogel's formula is valid, must have become e in order for the opposition / : e to be dephonemicized. A possible case in point is the doublet *silubra : *selubra 'silver', provided that the first is the basic form (cf. Lith. sidabras, OPr. siraplis, OChSl. sirebro). But since this is a loanword of obscure origin, too much importance should not be attached to it.29 5.5.3. The conclusion must be that the phonemic status of the Proto-Germanic opposition i: e is more uncertain, or more difficult to determine, before a following syllable with u than in the positions treated in the preceding paragraphs. On the one hand, there can be little doubt that the change e > i before u, which appears in West Germanic, is a manifestation of the same structural tendency, outlined in § 3.2, as the changes discussed in the preceding paragraphs. On the other hand, the fact remains that in North Germanic there is not a single fully certain example of this change. On the contrary, there are many examples with e preserved before an u, and although these cases are conceivably due to analogy, it is still true that only the uncertain sipr can be adduced as positive evidence. Therefore the possibility must be taken into consideration that the above-mentioned structural tendency manifested itself with varying force in different parts of the Germanic-speaking area. Unless sipr is a conclusive example, this tendency went one step further in pre-West than in pre-North Marchand, Language, XXXIII, 352-353. PG *widuwo (Goth, widuwo, OHG wituwa, OS widowa, OE wide we, with transfer to the «-stems) < PIE *widhewa (Skr. vidhava, Lat. uidua, etc.) is too special to be regarded as a counter-example.

>B 28

THE PROTO-GERMANIC VOWEL SYSTEM

191

Germanic, and already at this stage of the presentation, we can then conclude that the opposition i: e was at least not reduced to one of subphonemic variation in the entire Germanic area. A nonminimal pair in pre-North Germanic would be *lipuz 'a joint' vs. *meduz 'mead'. 30 5.6. Of the combinations implicit in diagram VII there still remain two, viz. / : e + cons. + {e, ai}. The dephonemicization of i: e would presuppose a change of either / > e or e > / (i.e., a merger e + i > e, or e + i > i) in these combinations. The current formulation of 'a-umlaut' includes e as one of the unstressed vowels (besides a and o) conditioning the changes u > [o] and i > e, while authors are mostly silent on the role of ai. Possibly this silence is to imply that since ai is a diphthong whose first member is a, it must have played the same part as a in changing i and u of the preceding syllable.31 Possibly, also, the relative rarity of relevant examples is to blame. In the formulation of rule (d) in § 5.3 not only ai, but also e, was omitted. This was done because examples substantiating the changes u > [o] and i > e before a, o, or au are much easier to come by, and also because some of the principal examples of unstressed e and ai indicate that these changes did not take place in these positions. These are the earliest Proto-Nordic (Runic) examples. Thus, the 3rd sg. weak pret. wurte (Tjurko), with original long e, as opposed to 1st sg. worahto (Tune), shows that u remained before e; the later form orte (By), O N orte, is due to the analogy of the 1st sg. In the same way, on the same inscription (Tjurko), the ¡z-stem dat.sg. -kurne shows the retention of u before ai; the ending, it is generally agreed, represents the PIE dat. ending *-oi, possibly preserved as -ai in the very earliest Runic inscriptions, e.g. hahai (Mojebro); see §§6.1 and 5.4. More indirect evidence is furnished by such weak masc. doublets referred to above (§ 5.4) as O N stige : siege, OSw. bughi: boghi, which can hardly be accounted for except on the assumption that i and u remained in the nom. sg. (before *-e). These examples suggest the conclusion that i and u remained before e and ai. Of the examples pointing in the opposite direction especially two merit some attention. One is OHG tohter, OS dohter, OE dohtor, O N dotter < PG *duxter, cf. Gr. Guya-crip, Lith. duktS, etc., where o seems to derive from u before e. There is, however, no necessity for the o to have arisen in the nom. sg. It is clear that the o, which is universal in this word in the Germanic languages, must in any case have been extended to all forms by analogical leveling. For, in some case forms the u should have remained, e.g., nom.pl. *duxtriz < PIE *dhuktres (with analogical zero grade of the

30

The change of e > i before secondary u (< o), e.g., 1st sg. OHG OS hilpu to helpan, dat. gibu (sporadic) to geba, OHG hilfa (due to the analogy of dat. hilfu) beside helfa, is clearly to be separated from the above earlier change. For it was a specifically German change, which is not present in Old English, cf. Angl. beoru, bero. Whether Runic gibu (Zealand) to *geban is evidence of the same change in Proto-Nordic is questionable. More probably, gibu is due to the analogy of 2nd sg. *gibiR, 3rd sg. *gibip. • 1 Marchand (Language, XXXIII, 348-349) is one of the few authors, if not the only one, who states explicitly that i became e before ai.

192

HREINN BENEDIKTSSON

suffix); PN (Runic) dohtriR (Tune), O N dotr, must be analogical. The [o] would be regular in the gen. pi. *doxtrd, and in the gen. sg. *doxtraz < PIE *dhuktros, cf. Gr.

Ouyaxpog, provided that this variant of the gen.sg. existed in Proto-Germanic. From these forms the o was extended to the whole paradigm, including the nom.sg. The other is the weak e-verbs, which, as shown above (§ 5.4), have examples of the changes i > e and u > [o], especially in Old High German (e.g., leben, dolen). At first glance this seems to be a quite straightforward case; the Old High German paradigm has e in all forms, and therefore stressed e and o apparently can be due to nothing but the e, and should accordingly appear everywhere. But on closer examination, the problem appears more intricate. On the one hand, not all e-verbs show these changes. Old High German has, for example, zilen, hlinen, biben vs. leben, but invariably o, c.g.,folgen, dolen, wonen. The remnants of this class in Old English and Old Saxon show invariably i (OE libban, OS libbian). In Nordic we find invariably i ( O N lifa, OSw. liva), but fluctuation between u and o, e.g., O N duga, una vs. pola,

pora; OSw. both pula, pura and pola, pora. On the other hand, the morphological structure of these verbs and its development is highly problematic in its details.32 Thus it is quite likely that the Old High German paradigm, with e in all forms, is due to secondary analogical leveling. Variant forms, probably representing earlier types of formation, are preserved; on the one hand, forms with a instead of e; on the other, a pret. without link vowel, hapta, corresponding to OS habda, OE hsefde (1st sg. *libdo, e.g., would regularly have yielded O H G *lepta, cf. especially pret. hocta to huggen, which was originally an e-verb, cf., i.a., O N pret. part, hugat beside hugt). The e and o for earlier i and u in these verbs may therefore be derived from rule (d) in § 5.3, and need not have been conditioned by the e of the following syllable. Thus there are no compelling reasons for assuming the changes i> e and u > [o] in these two positions, and the opposition i: e may have maintained its distinctive function in Proto-Germanic before e and, at any rate, before ai. A minimal pair would be 2nd sg.pres.subj. *wigaiz 'you kill' vs. *wegaiz 'you carry'. 5.7. In the current discussions of the phonemic status of the opposition i: e in ProtoGermanic, the combinations dealt with in §§ 5.3-6 appear to be regarded as exhausting the possible positions of i: e. But the evidence is that this was not the case. Diagram VII in § 5.2, which was constructed primarily to show all possible bisyllabic combinations, but may also, with the remarks in § 5.52, be taken to imply combinations of more than two syllables, includes two parentheses, which indicate that the elements they include are optional parts of the combinations. Thus the outer and inner parentheses signify that there were monosyllabic combinations ending in a vowel (i or e) or a consonant (or consonants), respectively. The phonemic status of i: e in these combinations will be examined in this paragraph. The first type of combination, ending in a vowel, i or e, was no doubt quite rare.

82

See, e.g., W. Wilmanns, Deutsche Grammatik, III, Ist and 2nd ed. (Strassburg, 1906), 87-90.

THE PROTO-GERMANIC VOWEL SYSTEM

193

The principal example is the negative particle PIE *ne (Skr. na, Lat. né-, etc.), Goth. ni, OHG OS ne, ni, OE ne, PN ni, ON né (in part probably).33 The second type was much more frequent. From the historical point of view this type represents combinations of two kinds : (a) Forms that were monosyllabic already in pre-Germanic times. The most certain examples are PIE *seRs 'six' (Skr. sa(, Lat. sex, etc.), PG *sexs, Goth, saihs, OHG OS sehs, OE siex, ON sex, and PIE pron. *is (Lat. is, etc.), Goth, is, OHG er, ir.3i (b) Forms that were bisyllabic in Proto-Indo-European, but became monosyllabic in early Proto-Germanic through the apocope of a final short vowel (of which there is no trace in any of the Germanic dialects). E.g., PIE *esti 'is' (Skr. àsti, etc.), PG *ist, Goth. OHG OS ist; PIE perf. 1st and 3rd sg. *woida, *woide (Skr. véda, etc.), PG *wait, Goth, wait, OHG weiz, OS wét, OE wat, ON ueit; them, imper. *-e (e.g. Skr. bhdra, Gr. cpépe, etc.), PG *-0, Goth, hair, OHG bir, OE ON ber. Potential pairs of i : e may be sought, for instance, in imper. forms of zero-grade 1st class presents vs. 4th and 5th class verbs. Thus, Old Norse has uit (from uita) vs. ber, gef (from bera, gefa), with i and e in nonminimal opposition. But in West Germanic the imper. of the first verb has been replaced by the subj. form (OHG wizzls, OE wité), and Old High German has bir, gib. The evidence is therefore not entirely unambiguous; the question is whether OHG bir is due to the analogy of pres. biru, biris, birit, or whether ON ber is due to the analogy of inf. bera, pres. ber, berr. The Old English evidence rather favors the former alternative; as in Old Norse, we find (North., sporadically) ge-wit (in L), but ber. If the pre-Old English form had been *bir, we should not, in view of 2nd sg.ind. bires(t), have expected analogical replacement of i by e. Old Saxon fluctuates between e and i (ber or bir). The opposition i : e therefore maintained its distinctive function in monosyllabic forms during the entire Proto-Germanic period; a minimal pair would be imper. *wig 'kill' vs. *weg 'carry'. The same opposition must also have been produced by a phonemic split of e through the early loss of final short vowels. Thus pre-Germ. *esti vs. imper. *bhreste resulted in the nonminimal pair PG *ist vs. *brest ; similarly *ist vs. *sexs, where there is an original monosyllable. In the above examples it has been taken for granted that final e did not become i, thus causing the change e > i, before it was apocopated. The evidence for this view is in fact strong, cf. especially the thematic imper. in *-e. In addition, the retention in West Germanic of PIE final i after a short syllable (OHG OS meri, OE mere < *mari) precludes the merger of final e and i. This permits the establishment of a relative chronology: (1) Apocope of final short e and a; (2) merger of e + i > i in unstressed syllables (§ 5.2); (3) stressed e + i > i before i (§ 5.3, a); 33 Another example is PIE *kv;e 'and' (Skr. ca, Lat. que, etc.), Goth. (u)h, O N né (< But since this was an enclitic, it is of no interest here. 34 To what extent O N rei. es belongs here is uncertain.

*ne-kwe).

194

HREINN BENEDIKTSSON

(4) apocope of final i and u after a long syllable (but retention, also of u, e.g., *felu, *fexu (§ 5.51), after a short syllable). The apparent exceptions have been explained convincingly, viz. as unstressed variants. E.g., OS ON mik, OHG mih, as opposed to OE mec (provided that this corresponds to Gr. ¿jxeye). Other unstressed variants of the same kind, some of them no doubt original monosyllables, are OHG OS PN ni (beside ne); 1st sg.pron. OHG ih, OE ic, OS PN ik, as opposed to ON ek; 1st dual pron. OS OE wit, ON uit < PIE *wed(-). (Besides, of course, forms of the type eMe show e > i, e.g., PIE *pinkwe (Gr. Tcevte, etc.), O H G f i n f , OS O E f i f , O N f i m m . ) 5.8. The general conclusion is that the structural tendency discussed in § 3.2, which would have resulted in the system in diagram V, never reached this final stage. The opposition i : e retained its distinctive function, throughout the entire Proto-Germanic period, at least in monosyllables and before ai and e, and probably also before u in part of the area. No doubt, though, this distinction had a narrower functional range than the other vowel distinctions until its reinforcement, discussed in §§ 4.2-3, led to the Common West and North Germanic system represented by diagram VI. 6.1. Finally, as an excursus, we shall examine in this section, the development of Proto-Germanic diphthongs in unstressed syllables. The three diphthongs, iu, ai, and au, included in the list of unstressed vowels in § 5.2, represent what may be called the pre-West and pre-North Germanic stage, which differed in part from the pre-Gothic. On the basis of Germanic and Indo-European comparative evidence, as expressed in the system in diagram II above, one may posit eight diphthongs, short and long, for the earliest Proto-Germanic. These diphthongs are as follows, arranged in four columns, with one or more examples of each: LONG

SHORT

I ei (+ ï > i) Goth. Nom.Pl. ansteis ON béner OHG ensti OS ensti OE wini (later -e) mxhti (later -e, non-WS)

II

III

êi Goth. Dat.Sg. Goth. anstai ON ON brüpe (?) OE OHG ensti Goth. Pres.Subj. bairais,bairai OS ansti, giweldi ON ON berer, -e OHG OHG berês, -e OS OS beres, -e OE OE bere ai Goth. Gen.Sg. anstais

OHG Gen.Sg. ensti Goth. Nom. Pl.M. OS ansti blindai

IV 01

Dat. SG. dege tage dage dwge (-a?)

195

THE PROTO-GERMANIC VOWEL SYSTEM

OE

uyrdi (-e)

OE

Dat.Sg. bergi (-e)

ON OHG OS OE

eu

blinder blinte blinde blinde (earlier -a?) eu

au Goth. Gen.Sg. sunaus ON sonar OHG fridö OS suno OE suna

Goth. Nom.Pl. sunjus ON syner OHG suni OS suni, megi

OE

Goth. Dat.Sg. sunau Runic magiu ON syne OHG suniu (suni)

Nom.Pl. suna (?)

ou Goth, ON OHG OS OE

ahtau atta ahto ahto eahta

OS OE

Dat.Sg.iWrto(?) suna (?)

In all the individual Germanic dialects these four columns were reduced to two. This led to a bifurcation of the Germanic area, of the same kind as in the development of the stressed vowel system referred to above (§ 4.1): Gothic on the one hand, and West and North Germanic on the other. In Gothic, cols. II, III, and IV merged, while in the West and North it was cols. I and III, and II and IV that merged, pairwise, as follows: I

II

III

IV

ei Ju

ai au

ai au

ai au

GOTHIC

NORTH GERMANIC

e ( + umlaut) e

( + umlaut) e ( + umlaut) (Runic iti)

( + umlaut) WEST GERMANIC

OHG

i

OS OE OHG

i

OS OE

i

-e-e e e ö (old) o a

e e IU (old)

(*) o

o a

196

HREINN BENEDIKTSSON

The outcome of this reduction was, for West as well as North Germanic, a set of three diphthongs, viz. iu (< eu, eu), ai (< ai, oi), and au (wfc-dialect was more or less willingly understood by a cux of Nunivak Island but completely incomprehensible to a Suk, suk of Kodiak or the Peninsula. The linguistic border between the latter and the Aleuts is so sharp that Ioann Veniaminov, who in the 1820-40's initiated the scientific study of both of them, could see little more than structural similarities and a number of common suffixes. Among the common words, found all the way to Attu, a distance from Port Moller of some 1200 miles, apparently is: WE. (1) yunraq 'young man' (Barnum), 'child' (Rasmussen), (2) cunraq, (4) 1848 sunáaq [suná[r]aq ?] 'young man' and Aleut suyanr-ir, with a 3rd p.sg. possessive suffix suyanr-a 'young person, young man'. The Eskimo word is derived from the shibboleth with a complex suffix -nr-aq and loss of the stem consonant -k before it, cf. Greenl. mánik = marrnik 'egg', mánernaq 'dwarf egg'; EE. pauq 'soot, lamp-black', Chapl. payuq '[black] ripple in the sea' (payula 'dust'), Greenl. paornaq, western EE. and Nunivak paunraq, Chapl. payunraq 'crowberry, Empetrum nigrum'. Formerly there was much traffic along the Aleutian chain and the Pacific coast, so the Aleuts could possibly have borrowed the word suyanr- from Eskimo at some early date. But how can we know that it is a loanword rather than an old cognate or an Aleut derivative with the suffix -nr- of a former Aleut word? The BORDERS between Western Eskimo and Eastern Eskimo are indubitable, but the unity of the former is less evident than the uniformity of the latter. According to Hammerich "the languages of Unalakleet and St. Michael are mutually incomprehensible, but many persons are bilingual",14 whereas E. W. Nelson stated his experience of the 1880's as follows: "People speaking these tongues do not readily communicate at once, although it takes but a short time for them to learn to talk with one another." The yupik interpreter from St. Michael who in 1881 accompanied Nelson to the Asiatic side, however, could understand "a considerable portion" of what the yupiyit said at Plover Bay (Bukhta Provideniya), but practically nothing of the same dialect as spoken on St. Lawrence Island.15 In 1895 N. L. Gondatti found that the jw/i-dialect of East Cape (Naukan) and the /««^-dialect of the Diomede Islands and Cape Prince of Wales on the American side of the Bering Strait were so different that the people had to use interpreters — even though the islanders regularly 11

Jenness, o.c., 3. Kaj Birket-Smith, The Chugach Eskimo (1953), 19. 13 Ivan Veniaminov, Zamecanija o koloSenskom i kad'jakskom jazykax (1846); I. Tyinov, Matfea svatoe Blagovestvovanie (1848). 14 O.C., 632. " E. W. Nelson, The Eskimo about Bering Strait (1899), p. 25. 12

THE ESKIMO SHIBBOLETH "INUK/YUK"

207

crossed the strait, in the summer in their skin boats and in the winter even on foot, 16 whereas Bogoraz observed five years later that the dialect of Naukan forms the transition to the dialects on the American side. According to Ronald W. Senungetuk, native of Cape Prince of Wales and now professor at the University of Alaska, there was also a regular intermarriage across the Bering Strait, so the difficulties could hardly have been more serious than in Norton Sound, even though there may be a regular intermarriage also between people speaking entirely different tongues. 17 As the example of the inuit and e'win of Greenland shows, however, the degree of mutual comprehensibility, which depends upon all sorts of imponderable factors, is a poor measure of genetic affiliation, so from Nelson's statements one could not infer that, diachronically speaking, the two dialects in Norton Sound were more closely related than the two jHpi'fc-dialects separated by the Bering Sea. But it remains to be shown that all those dialects which are NOT 'Inupik' were 'Yupik' in a diachronical sense, i.e., that the loss of the nasal in the shibboleth — apart from Eastern Greenlandic e~ < inuk — could be correlated with innovations shown to date from a 'Proto-Yupik'. In North Alaska the shibboleth has a strongly palatalized nasal: [inuk], i.e., an initial vowel phoneme i rather than an i, after which there is no palatalization, e.g., in (1) inur-, Kusk. Hinz nora = nur-a-a 'does not reach it', (2) ini 'village etc.', inaa 'his place, house', Greenl. ine = ini, ina = inaa, EGr. ea\ Nauk. ini, Chapl. ni, no., pi. nit, Sir. ini, ina, Kusk. ina or lna = (i)na, ine — inii, Chug. 3na, wa 'log house', Aleut nay-a 'its interior'. The palatalized nasal is found also, for example, in [kina] 'who', in all the other dialects kina (except EGr. kia), and after an initial i in (a) iniq- 'finish', inaaq- 'be ready', Greenl. inerpoq, EGr. e'rporj etc., Kusk. inerioq = in(i)r- 'is fully developed or grown', Aleut ina- 'come to an end'; (b) iniqtir- 'forbid', Kusk. inarqora id., Chapl. iniRt- 'calm down, warn', Aleut inayu- 'be forbidding, "do not want people around him"; persecute (mission.)'; (c) ini- 'hang up (e.g., meat for drying)', innisaq 'rack' < *ini-utaq 'means for hanging up' (on the gemination cf. below), Kusk. ini-ga 'hangs it up', Chapl. nin 'rack' < *innin < *ini-un, Aleut in-ka- 'hang (e.g., fish for drying); sky, air', ini- 'sky, air' (reference to clouds: inkamvru-), pi. ni-n, Atka -s 'flake for drying fish; storing platform'. In North Alaska, the phoneme i, opposed to i before all consonants and in final position, conditions also a strong palatalization of a following / and the assibilation of a t to s (merger with s), e.g., (d) [Warja] 'his comrade, relative; part of it', ila[rj]at 'one of them', Chapl. ilaqa, Sir. illya, Kusk. He = ilii (stem ila-), Aleut ila-, rel. ila-ya-n, vs. (3) [ila'] ilaa 'he himself', ili-yit 'they themselves', etc., Nauk. iL-rja, Chapl. 16 V. Miller, "Ob Eskimosskix narecijax Anadyrskago okruga", ¿ivaja Starina, vyp. II, god 7-i (1897), 134f. " So, for example, certain Ostyak clans belong to the one of the two marriage groups ('moieties') into which the Yurak Samoyed clans are divided, so that all these Ostyaks have Samoyed spouses.

208

KNUT BERGSLAND

(l)Ltja, Sir. la-ya, Kusk. Hinz itle, Barnum Le = (¿')£;[i], pi. (i)Lait, Aleut ya'n 'for him', yi'n 'for them', presumably < *al-rja-n, *al-yi-n, cf. imin, Atka imis 'for you' < *al-mit, Alaskan WE. (i)Lpit, EE. ilvit etc., Nauk. ILpik, Chapl. (i)Lpik, Sir. (i)Lpi < -piy < *-pit or *-plt as on the American side; (e) [iyFoa] iylua 'his match; the one, the other (of two)', Chapl. iyluya, Kusk. iylua, vs. (4) i'ylu 'house', Nauk. inlu, Chapl. riiylu, Sir. lu 'house', Kusk. rj'ilu 'beaver's house', Chug, 'rjilo 'cave', Kod. yluq 'house'; (f) isiq-tuq 'enters', Nauk., Chapl., Kusk. it'ir-, vs. (5) it'iq-tuq 'awakes', Chapl. taR-tur- < *tiR-, (6) tt'iq 'anus', Nauk. itiq, Chapl. ti'q, Sir., Nun. tl'R, Aleut iti-r. The assibilation is found south to Unalakleet, but not at the Bering Strait, nor in Canada (apart from the recent immigrants from North Alaska): Wales, King I. and Mackenzie, etc., itiq-toq 'enters' (Jenness), but then again in Greenlandic: iserpoq 'enters' vs. iterpoq 'awakes' (also in EGr.). To judge from certain spellings in the old documents, the phonemic opposition i vs. i may have occurred in Greenlandic as late as in the 18th century, so the assibilations in Greenlandic and in North Alaska, being conditioned by the vowel phoneme i, are most reasonably explained as instances of parallel development. In Western Eskimo, an initial i, as in the examples (1-6) above, has largely been lost, but not in the same way in all the dialects and not at all in the dialect of Naukan. In this respect the dialect of Chaplino comes very close to Western Eskimo of Alaska but there is no full identity and the dialect also recalls Chukchee; in Chukchee the initial 3 is very short and apt to disappear, and through the loss of i in the first syllable the dialect of Chaplino has got a pattern of initial consonants very much like that of Chukchee, viz. oral continuants and certain clusters in addition to the initial consonants found in Eastern Eskimo: p, t, s, k, q, m, n and, in one bird's name in the west, /. On the Alaskan side there is a rather close agreement, but no full identity, between Western Eskimo and Aleut, so it seems reasonable to assume some diffusion across the border on Alaska Peninsula, i.e., along the coasts of the Pacific and of the Bering Sea. Anyway, the various dialectal innovations in the vocalism do not correlate with the Inupik : Yupik dichotomy. After the initial /, a velar nasal has been lost in the western dialects at least in the following words: (1) Labrador all. iy-uya 'over there (behind, in the north)', N. Alaska (Nunam.) ta-inuya, loc. [wia-wi] 'over there, yonder', ta-inani-kpa-ni 'at a distance', Nauk. Jenness yedvan-toq 'it is far away', Kusk. Hinz ya, yane 'yonder', yaktak 'far', yaliagane 'on the day before yesterday etc.', yan, yate 'the space on the other side of it', Nunivak ya~ni 'over there, far away', Chug. loc. ya-qRua-ni, Kod. abl. ya-kwa-nik 'far', Aleut ya-m 'yesterday', ya-ta- 'be at such or such {-'ta-) distance', Attu yavyuyara 'aside' < *ya-S-muS-ar-an — cf. sa- 'front', sa-8-a-n 'at the front side of it' (EE. saa-ya-ni etc.) and saSmudaran 'forwards' — but iya-n, rel. iya'n, pi. iya-ku-n, Atka -ku-s 'that one sitting over there (not far)' with a preserved nasal as in Chapl. iy-na, iy-um,

THE ESKIMO SHIBBOLETH "INUK/YUK"

209

iy-ku-t, Sir. iy-na, iy-cïm-a, iy-kï-r-a < *iy-rum-a, *iy-ku-r-a 'that one over there (not far)', perhaps = Labrador iy-na, ik-sum-a (< -rum-) if this is not the pronoun ik-, Aleut ik-a-n 'that one (object) over there (e.g., on the other side of the bay)'; (2) Greenl. ingag-poq 'goes too far', ingassag-poq 'has exaggerated', Sir. yayïyït'go too far', maybe < *ya-ray-ït- ; (3) Greenl. ingalag-pâ 'avoids it', Chapl. yakur-naq 'danger', 'something to be avoided', Kusk. Hinz yag-toq 'abstains from certain things; observes superstition', perhaps Aleut yay- 'be kind (mission.)' ; (4) EE. iyiulik 'wave, swell', Chapl. vluk 'wave', presumably < *iwluk < *iwulik < *yuyulik < *iyu-yu-lik, cf. Chapl. kiwïk 'river' < *kuyïk, Nauk. kuyi'k, Sir. kucïx < *kurîk, Kusk. kuik, kwik, Nun. kwi'k, EE. kuuk (regular loss of r after vowels other than ï) ; (5) Greenl. ingûpâ, igûpâ, Wales Jenness iyu-t-ka[-a] 'he speaks ill of him', Kusk. Hinz. yu-lior-toq 'slanders', Sir. yu[-]m- 'scold'. The velar nasal in Greenl. ingu-voq 'the ice screws' seems to reflect a -w-, if not -y-: Canada-Alaska Jenness ivu- id., St. Michael Swadesh ivvuu-nïq 'piled-up ice'. After a and u the velar nasal, like the spirants, is largely lost in Eastern Eskimo, e.g. Greenl. nâ-voq etc., Chapl. nayï-, Kusk. nang-oq 'is consumed, concluded', cf. (-r-) EE. nai-, Chapl. nar[i]-, Kusk. nar-oq 'smells something', (-y-) EE. naa-lak-tuq, Chapl. naya-qur-aquq 'listens, obeys', (-r-) EE. naaq '(skin of) abdomen', Greenl. 1750 du. nersek = nassak < *nara(r)-ak or *-ak with an ancient contraction and a subsequent gemination of the preceding consonant; 18 after u: EE. uu-ma-, Chapl. uyu-va-, Kusk. Hinz ungû-va-uq 'is alive', cf. EE. and Kusk. uu-, Chapl. uyu- 'be burnt etc.' and EE. uuyaq, Nauk., Chapl. uukaq, Sir. ucïkïR 'tomcod', apparently a participle -kaq of a verb *uru-, perhaps Aleut huôa'to dry fish; dried fish' (the h-, usually = Esk. p- or m-, remains obscure). After i", on the contrary, these continuants are regularly preserved in Eastern Eskimo, e.g., (-//-) nïyauk, Chapl. niya[a] 'son- or brother-in-law', Aleut yarn- < *nïyaruk; (-7-) EE., Nauk., Sir. ïya-, Chapl. ya-, Nun. xya- 'cook', EE. ïyyan, Nun. iyan, Kusk. Hinz gan, pl. gâtit 'cooking-pot' < *ïya-un 'means for cooking', EE. ïyalïq 'kitchen (Greenl.), window', Kusk. Hinz gâleq, pl. gâlerit = yal(ï)rït 'window', Aleut yaly'entrance hole in the roof of the hut'; (-r-) EE. Nunam. ïraruq-tuu-ruaq, Wales izaroutuuzoaq 'coarse', Sir. ïrax-tï-valâR, Chapl. yax-tu-lri 'large, KpynHbiii'; EE. qïruk, Canada qiyuk, Greenl. qissuk = qisuk, Sir. qïcïx, qïrux-, Chapl. quuk 'wood etc.'. The treatment of -y- in Eastern Eskimo is not regular in a neogrammarian sense, cf. nuyu- 'be finished, used up', etc. The -y- is lost in Alaskan Western Eskimo almost to the same extent as in Eastern Eskimo but is generally preserved in the Asiatic dialects. The -r- is usually preserved, as such or as c, in Sirenik but in the other western dialects is regularly lost before a and u (before z mostly preserved as -y-), in Eastern Eskimo after these phonemes, and also after i. After 1, all the dialects 18 Cf. my paper "The Eskimo-Uralic Hypothesis", Journal de la Société Finno-Ougrienne, LXII, 2 (1959), 8ff.

210

KNUT BERGSLAND

seem to have an -tj- in words such as kiiju- 'back, hinder part', Aleut kit]- 'subsequent time; younger sibling'. But the -IJ- seems to have passed to -n- in EE. sinik-tuq 'sleeps', cf. Aleut syay- 'dream', and in Greenl., Canada inalu vs. N. Alaska iyalu(k) 'intestine' (Aleut irjlu- 'leader, hook line'?). Perhaps, then, the sibboleth reflects *ir/uk rather than *inuk. At least in Aleut there is a different word inu- 'piece of food; put (food) into one's mouth'. The velar nasal apparently has become dental after an initial i- also in EE. innaqtuq, -puq 'goes to lie down, goes to bed', Kusk. Hinz. inar-toq 'lies down' vs. Kod. 1848 iijagna- = iqar-na- 'lie (a sick person)', Chapl. irjaRtaquq 'lies down' etc., a contract derivative of the type EE. aullaq-tuq, Chapl. arulaq- 'has gone away', from aula-ruq, Kusk. Hinz arula-uq 'moves', Chapl. arula 'dance'. If Chug. Suk reflects *irjuk, the word underlying EE. innaq-tuq etc. might conceivably be contained also in Chug, Saruq 'sleeps', Aleut sara-kur id. Anyway, in Western Eskimo the geminates due to the ancient contraction have in general merged with the corresponding single consonants. On the Alaskan side, however, there are certain traces of the geminates, e.g., Nun. iyan vs. xya- above (unless the former is just the informant's EE. form) and Kusk. Hinz agger-toq, EE. ayyiq- 'is on the way, is coming', derived from EE. ai-, Chapl. ay[i]- 'go home etc.', Kod. ayi- 'go', Aleut ay- 'pass', cf. the loss of the single -y- in Kusk., Kod. tai-, Chapl. tayi- 'come (not from far)', EE. tai- 'call on; call, name', Greenl. taggit < *tayi-un 'name, term; that which one asks someone to do' (Aleut taya- 'alight'?). In the yupik dialect the abbreviation of the geminates is clearly later than the loss of the single intervocalic -y- and farther south it may be connected with Aleut, where there are no phonemic geminates but at least one fairly certain example of the ancient contraction: iyasi- 'wing' < *iya-usi-, iya- 'fly', -usi- 'means', Esk. -un, -ut'i-, cf. (1) the apparently later contract type sayu'si- — sayuusi- 'rope', sayu-, Nun. cayux-, EE. sayuk- 'pull'; EE. aylaun 'pencil', aylak-tuq 'writes'; Chapl. simi[i]n < *simiun, Sir. simiyita < *simiy-uti- 'substitute', and (2) the preservation of the stem consonant after i, later syncopated, in anr-usi- 'nose', anr-, Esk. anir- 'breathe'; EE. qilir-un 'cord', qiliq- 'tie' etc. The abbrevation of the geminates, which has perhaps spread to the Yukon area from the south, may have helped push the single -y- to -y- as in Kusk. yane etc. < *iya-ni. In Pacific Eskimo, where to the Eastern Eskimo and Asiatic Eskimo initial scorresponds c-, as also in the rest of Alaskan Western Eskimo, the secondary initial consonant of the shibboleth suk has merged with the reflex of *iy-, in EE. regularly is-, as in Kodiak 1848 Sax-t- 'stretch out (his arm)', Chug, saqa 'wing', Kusk. Hinz yagta, yakoq = Chapl. yaquq, Nunivak Jenness yaRtoR [yax-tuR], yaq'o-a, EE. isaktuq 'stretches out his arms or wings', isaruq, pi. isaqqut 'wing', not to be confused with Kusk. Hinz shagta 'spreads it out', Chapl. sa'x-t- id., Sir. sirax-t- 'smoothen', EE. siak-kai 'spreads them out, arranges them in a row', Greenl. siak-tu-t pi. 'the belt of Orion', Aleut sda- 'star'. The former word apparently means primarily 'to stretch out both wings or arms' (otherwise the bird would drop to the ground) so one may reasonably compare Aleut sa- 'bird' — no doubt different from sa- 'front' —

THE ESKIMO SHIBBOLETH "INUK/YUK"

211

rather than E. Aleut ya- 'the distance from the chin to the end of the middle finger of the right hand stretched downward' (Jochelson), which could hardly be separated from the above mentioned ya-m, ya-ta-, *ya-5-mud-ar-an, Kodiak ya-kwa-nik 'from far' etc. < *it]a-. In the dialect of Nunivak the secondary initial consonant of the shibboleth cux has merged (1) with the initial affricate as in culu, EE. suluk '(wing) feather', Chapl. suluk 'sound of the wind', Aleut suluy- 'noise of wind, thunder etc., echo (as when one hollers towards a cliff in the fog, to orientate oneself)', and (2) with the sequence *iy-, EE. is-: Hammerich cuvaRaRa, Kusk. Hinz yuverera, Chapl. yuvrir- 'examines it', EE. Barrow, Copper Jenness isivrioqtoq 'takes hold of, examines something', cf. Chapl. yuyaraq, EE. iswjrjaq 'jaeger, Stercorarius (parasiticus etc.)', derived from *iyu, EE. isu 'end', maybe the cognate of Aleut su- 'take, grasp, gather etc.'. The Nunivak form yuyutaR 'perspiration', which Jenness obtained from his inupiaq informant, seems to have the velar spirant from EE. isuyutaq 'moisture, dew' (cf. below) and could have the initial y- from the yupik dialect: Barnum yu ku' tak = yukwtaq 'dampness', but actually it goes with Aleut: yuy 'moisture; be damp'. If the form is genuine, the different treatment of -y- in *iyuvrir- and *iyukutaq could have been conditioned by rhythmical factors, for example by a heavier stress on the closed second syllable of the former form than on the open one of the latter. The same explanation could be applied to the different treatment of the nasal of the shibboleth in the yupik dialect: yuk < *irjuk, yunraq 'young person' < *iyunraq vs. Barnum in nu' gwo &k = inuyuaq 'doll' (Hinz inugauq), Greenl. inussaq = inuusaq, with the very common suffix -uaq, EE. -uraq 'resembling, etc.', thus *iyuy-uraq or *iyuyuraq, Nunivak Lantis cucu'yoax, perhaps = curjuaR. Hinz gives also the variant yuguaq 'doll', and Barnum's form could possibly have been borrowed from the north, even though from North Alaska (Nunam.) to Labrador the word for 'doll' has a different suffix: inurjtjuaq (Greenl. inunguaq 'small person'). But Hinz's form could also be a later derivative of yuk, or else it could represent a more southern step in the process which yielded yuk, yunraq, cux, cunraq, suk, sunaaq and, possibly, Aleut suyanr- 'young person'. Then the dental nasal in inuyuaq would simply go with that in Kusk. Hinz inar-toq, EE. innaq-tuq vs. Chapl. iijar- and, possibly, Chug. saruq, Aleut sara- 'sleep'. In the disyllabic form Nun. ya'ni, Kusk. yane < *irja-ni, as well as in Kusk. galeq < iyaliq 'window', gatit < iyyatit 'cooking pots', there is the same long vowel as in the trisyllabic form Nun. awa'ni, Kusk. avane, EE. awa-ni 'there far away' vs. Kusk. kakmani = EE. qakmani 'there outside' etc., so ya'ni seems to reflect a comparatively recent trisyllabic form *iya~ni and an earlier *iyani or *iya~ni < *iyani. The same sort of lengthening of the short vowel in trisyllabic forms is found also on the Asiatic side, but not generally, cf. St. Lawrence (tape recording) amvraq, dimin. of amiq 'skin' vs. arnaraq, dimin. of arnaq 'woman'; iqa'Luk = Kusk. Hinz eqatluk, but Nauk. Jenness eqaL'uk, EE. iqaluk 'fish'; narwya = St. Michael Rasmussen naro'jaq, but Chug, narua, EE. nauya(q) 'sea-gull'; anvyu 'snow', with the same prosodic

212

KNUT BERGSLAND

pattern as the contract form umvlik 'captain, owner of an open skin boat', EE. umialik, but on the Alaskan side loss of the velar spirant and lengthening of the consonant before the secondary diphthong: Nun. arriu, St. Michael Swadesh anniu, and Naukan Jenness anyi, perhaps = anii, with the late contraction characteristic of Asiatic Eskimo, EE. aniu and, with preservation of the spirant before a syllable heavier than the last short one, Nunam. aniyuPyaq 'snow house'. The prosodic structure of the various western dialects remains to be clarified. But the lengthening of the medial vowel in trisyllabic forms could hardly be dated to 'Proto-Yupik', being apparently also later than the abbreviation of the geminates. On the Alaskan side, it could rather be connected with the high pitch on the same syllable in Aleut, as on the second a in haqakur 'he comes'. The lengthening of a in ya-ni apparently is later than the loss of the initial syllable in Kusk. Barnum hwa 'ne, Hinz vane = (x)wani, EE. uwani, Wales — with a regular loss of the intervocalic spirant — uani 'here (where I show)' (cf. Wales oani < awatii; awun < apun 'snow on the ground' etc.), and Chapl. xwani, Sir. mini id. — different from Sir. manr, EE. maa-ni 'here, in this country etc.'. The Sirenik m-, which presupposes the secondary xw- (or w-) found in the neighbouring dialect, has an exact parallel in the westernmost dialect of Aleut: Attu ma-y(u) 'here, hither', marja- < *ma-ma- '(I) do now', Atka (1826, 1860) wa-ytts, wa-ma-, E. Al. wa-yun, yu-ma- (so also modern Atkan). Cf. further Sir. mit- 'flower, bloom; rise (of sun), begin (of the new year)', EE. and Kusk. Hinz uit-, Kod. wit- 'open one's eyes; break (of day); crack (of ice) etc.', ui-luq 'Blue mussel (EE.); small spoon (Kusk.)', Aleut wayyi-r, Attu mayyi-r 'Blue mussel; small knife' (suffix -yyi-, -yy-). The initial (x)w- of the respective neighboring dialects is very rare, and manifestly secondary, so the change to m- in the westernmost idioms of Eskimo and Aleut could hardly be regarded as accidental. But in between there are some thousand miles of ocean, so there could hardly be any historical connection either. For a different reason, the same must be true of the respective neighboring dialects. In the 1st p. pronouns, which have the same stem and in the singular the suffix -ya 'I', one finds the same set of correspondences: Sir. mfrja, Chapl. xwaya, Nauk. MenovsSikov (without indication of quantity) wi, wiya, Jenness uwatja, Wales uaya, farther east uwatja (EGr. era), Kusk. Barnum hwe, rel. hwegna, Hinz ve, vinga, Nunivak Jenness kwiya, Chug., Kod. xwi. The relative forms vinga etc. seem to have got a second -ya 'I' added, for the final -/[/] of the simpler forms apparently reflects -aya as in Kusk. Hinz ile = ilii 'his comrade; part of it', EE. ilaya, ilaa, Chapl. ilaya etc.; Kusk. itle = (t)Li 'he', EE. ilaa; cf. also the loss of -y- in ayag-tu-a 'I went' etc., EE. -ya. After an initial syllable the -y- is preserved, as in Kusk. Hinz kannga 'o', so the contraction presupposes a trisyllabic form like the EE. form uwaya rather than a disyllabic one like Chapl. xwaya, which, thus, is separated from the Alaskan forms also by the Naukan form wi. The contract -ii < -aya is opposed to -aa < *-ara as in qayaa 'his kayak' (qayaq), Kusk. Hinz angia 'his open skin boat' (angiaq), Chapl. ayyaa, Sir. atjyira < *ayyara ( -ii vs. -aa or the like. If, however, Nauk. wi reflects the same contracted form as the Alaskan (x)wi[j], the Alaskan form (i')ii[(] 'he' would seem to have been contracted later, for the dialect of Naukan has iLija 'he', with preservation of the nasal after the syncopated vowel as in Chapl. (i)Ltja (Sir. la-tja), cf. pi. rel. iLyita and EE. ili-yit 'they themselves'. The Alaskan form goes rather with EE. ilaa < *ila-ya < *iTi-ya, so that the Naukan form wi is more reasonably explained as an independent reflex of uwar/a. In Chaplino, the simple sound shift represented by the forms xwani, xwarja, has also taken place after the loss of an intervocalic r : wamt-, Sir. urmit- < *ura-mit-, Greenl. ua-t-, uapa 'deceives him', ussapai < *ura-ut(i)- 'is deceptive with them: takes too many of them (things) for himself. The overriding innovation, which sufficiently explains most of the particular points of agreement between the one or the other of the Asiatic dialects and the dialects south of Norton Sound, apparently is the change of the prosodic pattern, i.e., primarily the abbreviation of the geminates which seems to have taken place independently on each side of the ocean. On the Alaskan side, a new initial sequence yu- has come into being also through the metathesis of iw- before -a-, e.g., Kusk. yualu, Nun. yualuR, Chug, yua luq but Sir. ivila, Chapl., Nauk. ivalu, EE. iwalu and, with a different metathesis, Greenl. ujalo = uyalu 'sinew'; Kusk. yuar-, Chapl. ivar-, EE. iwaq-, Greenl. uyar- 'search' vs. Kusk. iva-uq, Greenl. iva-voq 'hatches eggs', in Aleut, perhaps, respectively, rmi[-]t- 'chose, pick out', and/or imyar, Atka imyar 'fishline for deap sea; to fish in deap sea' vs. miya- 'hatch eggs'. In Aleut the -w- has merged with -m-, so the Alaskan WE. yu- < iw- is a local innovation. The sequence yu- of Alaskan Western Eskimo also reflects an earlier iyu-, with palatalization or simple loss of the velar spirant (cf. above): *iyu- or *iu- > yu-. In the following word the velar spirant has become intervocalic through the loss of a post-consonantic r, an innovation common to the western dialects other than the idiom of Sirenik (where the r is preserved in all positions, as such or as c): Kusk. Hinz yoringauq 'is stiff, frozen', beside iyuR-tu-a 'I stiffen' (transcribed from Barnum), Chapl. iyur-niq 'lump, clot' etc., Greenl. igssor-poq 'stiffens, coagulates' < *iyfuq-. In the dialects in question the post-consonantic r is usually preserved as y or s (z) before i, but dialectally also sometimes before u: Chapl. du. iyuyik, Sir. iycix, EE. iyruk, Nauk. ixuk and St. Michael Swadesh iyyuk presumably < *iyyuk, Nunivak Hammerich irj3cu{x) 'testicle'; Chapl. quuk 'swan' (also 'log', EE. qiruk), EE. quyruk, Kusk. Hinz kuguyuk, presumably = quy(u)yuk, Nunivak Jenness quyusuk. The details indicate a parallel development rather than a 'Proto-Yupik' innovation. The same would seem to be true of the following remarkable identity: Chug., Nun. and Chapl. Ruq 'arrow, cartridge', Sir. qarci'R, qarruR, EE. qarruq, Canada qaryuq (y < r as independently in WE., cf. also Attuan y < d), Greenl. qarssoq = qarsuq. The

214

KNUT BERGSLAND

loss of the a between the two uvular consonants apparently was conditioned by the soundshift qarruq > *qaruq and hardly is more surprising than the loss of i (rather than i) in the same position: Kusk. Hinz qrutoq < *qiru-t-tuq 'is cold (a person)', Chapl. Rut-, Sir. qvr'iRt- id., EE. qiu- 'freeze to death', cf. WE. aru-, EE. au- 'rot' etc. vs. WE., EE. niri- 'eat'. In Nunivak yw-yoR, Kusk. yu-goq 'gets out of a boat or sledge; disembarks', yu-ga 'takes it off (clothing)' the earlier innovation is the loss of the initial n- before i, common to Western Eskimo and Aleut: EE. niu- intr. 'get out of a boat etc.', tr. 'take, put out; ladle out; unload etc.', Greenl. niggupa < *niyu-ut(i)- 'gets out with it or him', Chapl., Sir. uxt-, wxt- 'disembark', with vowel assimilation as in tuyu-, EE. tiyu- 'take' and syncope < *iyut- < *iyyut- 'get out with himself', Aleut iyu'take or pull out (e.g., load out of a boat, eggs out of a bird's hole in a cliff)'. The initial n- has also been lost before i in the following words: EE. nii-uq-tuq 'drills', Greenl. niggit < *niyi-un 'whirl in a drill', Sir. iy[-]n[-]/?i-, Aleut iy-Say- 'to drill'; EE Canada niur 'swallow (liquid), sip', Chapl. iuRtaquq 'inhales, takes a sip', iur-nfq 'inhalation', Sir. iruR-(t-) 'smell, snuff', iciraLiR 'inhalation', Aleut iSu-t- 'snuff', idrut- 'swallow'; EE. niu 'leg', pi. Alaska, Canada niut (Greenl. secondary nisut — nissut, cf. the normal nuna-t 'lands' etc.), with loss of -r- as in qiu- above, WE. iru, Sir. ira 'id.', Aleut iru- 'go away, abandon' and/or vRi- 'to limp', possibly with the suffix -ri- as in iqya-ri- 'go in a kayak; swim (of a duck)'. Before i", on the contrary, the n- is preserved in Western Eskimo and has regularly merged with t- in Aleut (as also before a and u), e.g., Greenl. nivi-va 'seeks it, keeps to it', Chapl. nlvi'-, niv-aquq 'sticks, adheres to something', Aleut tmu- 'weigh on, lie heavy on'; Esk. nimiq{-) 'band, wrapping; to bind, tie up etc.', Atka tmay 'Nereocystis Luetkana, a very long seaweed formerly used for fishlines (imyar)', tma-dyi-r 'Atka mackerel (a fish with conspicuous dark 'bands' around its body)'. On the Asiatic side, however, the initial n- is preserved before an i in Chapl. tiivaqa[a] 'pours it from one vessel into another', Greenl. nivag-pa 'flings it away; shovels it off', Alaska Jenness niwak-tuq 'digs', perhaps Aleut im- 'to turn over'. Also the curious Sirenik form yayiy-, yayuy- 'dig' may have had the initial n- : < *yuyay- < *nuyay- < *niway-, or *yuyay- < *yiway- < *niway-, cf., for example, yayviR, Chapl. Nauk. nayvaq, EE. narwaq, Kusk. nanvaq 'lake'; yayira, Chapl. narwya etc. 'seagull' andyav'irtt-, Chapl. navrut- 'exchange'; yuvix, Chapl. nuvak, EE. nuwak 'saliva'; yivit- 'stick to', Chapl. riiv'i- etc. above. Rather than postulating a phonemically palatal *«-, one may perhaps look for an explanation in Chukchee. In modern Aleut, where there are only three vowel phonemes, i, a and u, the dental consonants are quite strongly palatalized before an i of any origin, and there are also some words which have an initial n- only in the eastern dialect, e.g., ir, (n)iri-, (ri)iri-la'baidar, open skin boat', (n)iri-lu- 'to paddle in a baidar'. Before the ancient *i, opposed to *i (*a), the dental consonants may have been palatalized in the whole Eskimo-Aleut area, for there seems to be no certain example of an initial *t- before this i in either language, in Aleut only c- and s-, in Eskimo only c- or s-, the c- being

THE ESKIMO SHIBBOLETH "INUK/YUK"

215

confined to Alaskan Western Eskimo, where s-/s- reflects the secondary y- as in suk etc. or the usual c- in cases of recent syncope such as sta'man, Kod. 1848 ctaman, EE. sitamat, sisamat, Nauk. sitamat, Chapl. stamat, Sir. sitima- 'four' (Aleut sic in, siciy). The loss of the initial n- apparently was conditioned by the following i, i.e., by a rather trivial palatalization, and could have taken place at any time, just as the cluster *-ny- has been abolished in Eskimo at least twice. In the modern dialects the cluster -ny- seems to occur only in Naukan, viz. as the reflex of -rjy- (cf. Nauk. inlu, Chapl. n'irjlu etc. 'house' above): Jenness anyaq, Chapl. atjyaq, Sir. ayylR, Kusk. Hinz angiaq, Nunivak Jenness ayizaq 'open skin boat' (EE. umiaq), probably < *atju-yaq, cf. EE., WE. aijuar-, Sir. ayicir- < *arju-rar- 'to row or paddle in an open skin boat', ar/uar-un etc. 'oar, single paddle oar'; *-rar- is a well-known frequentative suffix and -yaq could be the causative suffix -yar-/-sar-, Aleut -ya- 'try constantly to make —', the stem being perhaps atju- 'catch, hunt' (or else the cognate of Aleut aya- 'side, space alongside, longitudinal half, match etc.', cf. atja'ru- 'single paddle oar; pole on the left side of the kayak to prevent the fishline from cutting the kayak skin'). An earlier cluster *-ny-, due to the reduction of -ayu- to -yu-, has been reduced to -y- on both sides of the Bering Sea in the name of the interesting fish which in a Copper Eskimo tale recorded by Jenness is addressed by the ptarmigan as kanayuq kanna... 'the sculpin down there ... you there, you with the big fore-head', cf. Kusk. cam-yuq 'you (far) down there!' (demonstrative stems kana- and cama-, EE. sama•): Rasmussen Kusk. kayo-qupak, Nunivak pi. ka^yu-t, Jenness sg. kaiyu, Sir. Gondatti kaynirax = kay-yiraR, Chapl. kayu, dimin. kayurjiRaq (Bogoraz kayutjixak), Wales Rasmussen kaniyoraq, Jenness kaneyuaq, Nunam. kanayuq, pi. kanayut, Greenl. kanajoq, kanasut, NGr. kanioq, kanisut = -ss-. As pointed out a long time ago by Bogoraz, the Chukchee have borrowed the word in its full form: kanayo-lyan, pi. -t (Chukchee suffixes), i.e., either from Asiatic Eskimo at a time earlier than the reduction of -nay- to -y-, or from the Eastern Eskimos of the Diomede Islands or the American side of the Bering Strait, bypassing the people of Naukan — if these have or had the form kayu (in Naukan, Gondatti got an entirely different name for this fish); if not, the form kayu is still more likely to have come about independently of the Alaskan kaiyu etc. The irregular plural forms (and other geminated forms) in Eastern Eskimo of the well-known word qayaq 'kayak' indicate that the word reflects an earlier form *qanyaq: Canada qainnat, Greenl. 1750 kainet, modern qainat = qaannat with the regular merger of ai with aa before a consonant, and Alaska qayyat, in WE. regularly abbreviated to qayat. Perhaps it is a derivative, with the same suffix -yaq as in at]{i)yaq, of the verb qani-, Greenl. qani-va 'goes part of the way with him', qani-t- 'be near' etc., either in the sense of accompanying the open skin boat paddled by the women, as the Eskimo men do in their seasonal travels, or in the sense of coming close to the game (Aleut iqya- 'kayak', rather than being the cognate of the Eskimo synonym, could be derived with the same suffix from the above mentioned word ir 'baidar,

216

KNUT BERGSLAND

umiak'). Anyway, the denasalization in the Alaskan form qayyat must be later than the general merger in Eastern Eskimo of the geminated counterpart of -y- with that of -s- as in nuyaq, pi. nutsat, WE. nuyaq, nuyat 'hair of head', nasaq, pi. natsat, As. WE. nasaq, nasat, Al. WE. nacaq, *nacat 'hood'. If, as it seems, the earlier plural form was *qannyat or *qanyyat (in clusters, the gemination does not show up in the modern dialects), the passage to qayyat in Alaska could simply be part of the general Eastern Eskimo denasalization of the nasals (m, n, rj) before most oral consonants, which east of Alaska has come on top of a local innovation, the general merger of a dental consonant with a velar one before a dental or labial consonant, e.g., Chapl. qan-li-, Wales-Barrow qalli- 'come closer' vs. ay-li-, ayli- 'grow, become bigger', Canada and old Greenl. qayli-, ayli-, respectively derived from the just mentioned word qani-(t-) 'be near' and from arj'i- 'be big' with a regular syncope of the stem vowel i (vs. ami-li- 'become narrower' etc.); cf. Chapl. aLpa, Wales atpak, Canada akpak, old Greenl. akpa 'guillemot'. Before the uvular stop -q-, both the -n- and the -/- (-L-) have merged with -t- in most of the dialects, but in the dialect at the Bering Strait (Diomede Islands, Wales, King Island) the nasal has been kept apart, having been assimilated directly to the stop, e.g., WE. tanqiq, Wales taqqiq 'moon' vs. WE. miLquq, Wales mitquq 'feather, hair (of body)', WE. atRar-, Wales atqaa-tuq 'descends', but farther east tatqiq, mitquq, atqaq-tuq, in Labrador and Greenland (by 1724) with full assimilation taqqiq, miqquq, aqqar-puq (variously written). Thus, the denasalization before oral consonants, which is one of the most important Eastern Eskimo innovations, could not possibly date from a "ProtoEastern Eskimo". The eastern form qainnat could hardly have yielded qayyat, for also in Alaska there is a sequence -ainn-. But it could possibly reflect an earlier *qaynnat, with a consonantic -y-, whether or not there was a metathesis *-nny- or *-nyy- > *-ynnbefore the general denasalization (cf., for example, anruaq 'amulet', east of Hudson Bay aryuaq, Greenl. arnuaq). The former existence in Eastern Eskimo of a preconsonantic y would explain also the otherwise quite obscure single -q- in the word niaquq, Alaska [nea-], pi. niaqqut 'head', Nauk. nayqitq with a voiceless y, Sir. iciqi'R, iyiqu(pax) < *nay(l)quq, Chapl. and Alaskan WE. nasquq; cf. Kusk. Hinz najangara 'bows to him', Greenl. najangar-poq 'sits nodding in his sleep', Wales neayar-toq 'falls asleep while seated', perhaps also Chapl. nay-raq 'mountain' and/or the contract derivative Kusk. nayeq, pi. nayit, EE. natsiq, natsiit 'fiord seal, Phoca phoetida' (the hunter approaches it by imitating the nodding movements it makes when basking on the ice etc.). In Eastern Eskimo, a single intervocalic -q- has regularly passed to -r-, and -k- to -y-, after a vowel other than that of a short initial syllable, as in the above mentioned word isaruq, WE. yaquq etc. < *iya-quq 'wing' (with the same suffix -quq as in niaquq 'head'), and in putuyuq, pi. putukkut, Chapl. putukuq 'big toe'; in the dialect at the Bering Strait also after a short initial syllable, e.g., turu- < tuqu- 'die', ayia < akia 'its price' etc., but neaquq 'head'. In the latter dialect a veritable 'Lautverschiebung' has been taking place, the geminates being very much

THE ESKIMO SHIBBOLETH "INUK/YUK"

217

shorter than farther east, the stops other than -t- having become spirants (and s voiced) and the corresponding spirants mostly zero, e.g., Wales tiu- < tiyu- 'take'. 19 But the secondary -y-, as in putuyuq, has not been lost, which indicates that the 'Lautverschiebung' may have been part of the general Eastern Eskimo 'weakening' process, rather than an innovation superimposed upon it. Anyway, at the time when the intervocalic -q- was passing to -r-, all the dialects must have had the same form nayquq as the dialect on the other side of the Bering Strait, on the American side presumably pronounced *\neyquq] or the like, cf. \_-eya-'\ = -aya- etc., and later switched to [neaquq] = niaquq, with abolishment of the preconsonantic -y-, in the whole area where -q- had already passed to -r-. Both innovations are coextensive with the denasalization and define modern Eastern Eskimo but the geography obviously does not yield any valid measure of the 'time-depth'. Both in Chaplino and south of Norton Sound (Unalakleet) the form nayquq, preserved in Naukan (and indirectly in Sirenik), has become nasquq, just as a secondary -y-, reduced from -ay-, has become -s- after a stop in the same dialects, e.g., Sir. aqv < *aqiya 'belly', aqiRqi'R < *aqiyaquq 'stomach', Chapl. aqsa, aqsaquq, Nauk. aqyaq, Wales ariaruq, Nunam. aqiaruq [-ea-\ Greenl. aqayak, aqayaruq, pi. aqayaqqut id., Kusk. aqsik 'stomach', aqsaquq 'belly', Nun. aqsaq (Hammerich aksix), aqsaRquq, Chug, aqse, aqxaruq id., E. Aleut aqiya- 'pelvis, posterior', Atka du. a'qaya-y, Attu a'qiya-ya-y 'hips'. This word constitutes an almost exact parallel to ka{i)yu < kanayuq, the difference being that the intermediate step -i(y)a- has a slightly wider geographical extension. To explain the passage of a voiceless y to a sibilant there seems to be no need to assume a direct contact between Chaplino and the Yukon area, nor an immigration of Eastern Eskimos to the Seward Peninsula. Just as the dialect of Naukan forms the transition between Asiatic Eskimo and Eastern Eskimo, the yi/fc-dialect of Alaska would rather seem to form the transition between the latter and the dialects of the Pacific coast. For 'man' the j>uA>dialect of Alaska has two words: (1) Hinz angun = EE. atjun 'man', with poss. suff. Greenl. anguta 'his father', N. Alaska aijutaat 'her husband, the head of the family' ('their man', also of a couple without children), Chug, atjun 'old man', derived from aij'i- 'be big': atj-un, rather than from atju- 'catch, hunt', cf. EE. arjrjun < *arju-un 'game'; (2) Barnum, Hinz taru = Asiatic Esk. Bogoraz taru, Canada tau, and, with the suffix -uraq 'resembling', Greenl. 1750 taursak = tausaq, E. Greenl. taautar, the equivalent of yuk/inuk in the shaman's language. In East Greenland 'a man' is tikkak, the same as 'old male seal (which has an evil smell)', whereas in 1884 e~ (inuk) was replaced by tdq, apparently the shamanistic word taaq < *tauq, but merged with taaq, rel. tarrup, N. Al. tarrum 'darkness' < *tariq, *tar(i)r-um, taaq-tuq 'it is dark', Aleut tar-ti-kur 'it gets dark'. In Aleut, 'a man' is tayaru-r, in the eastern dialect the near synonym of Esk. inuk, yuk, western anr-arina-r 'having breath, living person', E. Al. 'woman', in the west rather 'male' (vs. " Cf. D. Jenness, "Notes on the Phonology of the Eskimo Dialect of Cape Prince of Wales, Alaska", Intern. Journ. of American Linguistics, IV (1927), 168ff.

218

KNUT BERGSLAND

ayaya-r 'female'). The word could easily be a derivative of taya- 'trade, buy' with the frequentative suffix -ru- 'have (do) much, many', cf. EE. tauqsiq- 'trade, buy', a regular derivative with the suffix -liq- 'provide with' of the word found in Greenl. as taorte 'that which is given in exchange'. The latter, in turn, looks like an agent noun *tauq-t'i 'making trade', possibly < *taruq-t'i (regular loss of -r- after a vowel other than i), so that if this word has anything to do with Aleut taya-, the shamanistic taru, tau apparently equals Aleut tayaru-: *ta(y-)ru(q), cf. the Esk. frequentative suffix -ruq- in EE. nii-uq-tuq 'drills' etc. Trading was an important business of both the Eskimo and the Aleut men and brought them great distances, for example, from North Alaska to the region of Chaplino and from the western end of the Aleutian chain to Prince William Sound and the Tlinkit etc. And word taboo was practiced both by the Eskimos and by the Aleuts (Veniaminov) and may have had far reaching effects, depending upon particular situations impossible to reconstruct. The lexical differences between Aleut and Eskimo and between the various Eskimo idioms are not surprising, nor analyzable in quantitative terms. One may also ask how one could possibly distinguish between borrowing and 'genuine development' in this part of the world. From taaq < *tariq 'darkness' are derived EE. tarraq, Chapl. tarntraq 'shadow, reflection' < *tar(ir)-n(i)r-aq and EE. Nunam. tarniq 'man's life, spirit', Greenl. tarneq, tame 'soul' < *tar{i)r-niq, cf. Greenl. kingu-neq 'that which comes after in time or space', Nunam. pi. kirju-nr-it 'tribe', from kiyu- 'back, hinder part (as of kayak)', Aleut kirj- 'subsequent time; younger sibling' etc. To the former derivative corresponds Kusk. yunraq etc. 'young person' < *iyu(y)-n(i)r-aq, with syncope in the suffix -n(i)r- and thereby also loss of the stem consonant due to the additional suffix -aq. To the latter derivative would correspond Aleut suyanr- 'young person, young man' < *iyuy-(i)niq, with a later syncope of the suffix, as in suyanr-a < *iyuyi'nr-a, and an 'auxiliary' -a- < -i- (*a) possibly in other similar cases as well. There seems to be no particular reason for assuming borrowing in just this case, which is just one of several instances of an intriguing agreement between Aleut and Pacific Eskimo. In addition to his 'soul' and his name, Esk. ati'q, atq-a, WE. atR-a, Aleut asa-, asaa human being has also a body: Nunam. inu-u-niaq-tuq 'tries to be an inuk, tries to make a living', in simple English: 'lives'. In Atkan Aleut a living human body, especially the trunk, in the religious translations also the flesh, is called iyyu-, Salomatov 1860 irjiyu-, whence tyin iyywsa- 'to wriggle', cf. the suffix -yu- in aYu-y du. < *at-yu- 'the buttocks', at-, Esk. atl- 'the lower part'; kayayu- 'the upper part of the body', kay-, Esk. kayiq(-) 'the top (as of a plant), the upper part'; Greenl. tingajuk 'a caribou's beard', tirj'i 'pubic hair', Aleut cya- 'hair of body, fur' etc. The Aleut word could possibly reflect the same form as the Eskimo shibboleth, to be reconstructed as *itjyuk. If, as it seems, the treatment of the intervocalic -y- after an initial i- depended also upon rhythmical factors, the simple word *iyyuk might in Aleut have been treated differently from the stem in the derivative suyanr-, pos-

THE ESKIMO SHIBBOLETH "INUK/YUK"

219

sibly < *iyyuy-(i)niq, and the Eskimo forms suk etc. could possibly be due to the very numerous derivatives, or else *iyyuk could have become Esk. *iijuk generally etc.: the shibboleth being necessarily unique, i.e., different from other words, there could be no neogrammarian 'law'. Therefore, the Aleut word could also be a later derivative of the same word, *iyu-yu- or the like (cf. Esk. *irju-yu-lik 'swell, wave' above?). Or it could be another derivative of the same stem, cf. WE. iylu-, EE. iylti'the one, longitudinal, half of a symmetrical body (Chapl.), the one of a pair' with a 'singulative' suffix as in niu-lu- 'single leg: the root of a plant'. Two such halves make a pair: Greenl. igdlu-gik 'having each other as igdlu-, cf. Aleut iyacir, iyacari-y du. 'married couple' (-ari- 'have'), whereas a human body is a symmetrical whole, cf. E. Aleut pi. anyi-n 'intestines' vs. du. anyi-y 'mind'. With a possessive, i.e., relational, suffix EE. inu-a, Kusk. yu-a, Chapl. iw-a, Sir. yuyw-a mean 'owner', e.g., 'owner of the house', 'the host', to which corresponds in old Atkan Aleut (Salomatov 1860) iyisy-a 'its owner', ulam i. 'the master of the house, the host' (St. Luke 12 : 39 etc., Labrador iglub inua, Kusk. nim yua). The word seems to contain a suffix which in place names means 'provided with' or the like, e.g., cuyusyis pi., name of islets, cuyu- 'sand'. A host has a house, ula-, but necessarily also a guest, to be received in friendship face to face and referred to as itja-n 'that one sitting over there (in front, not far)', Chapl. iy-na, rel. irj-um etc., so it is reasonable to think that the stem is the demonstrative iy-, the 'moving' counterpart of the static ik- in Aleut ika-n 'that one over there, e.g., a house on the other side of the bay' vs. aka-n 'that one far away, at an unlimited distance in the horizontal direction from the speaker (in front)'. The near synonym of Esk. irj-lu- is in Aleut the simple stem arja- 'side, space alongside, longitudinal half, the one of a pair', whence i.a. arjad- 'the opposite side' (Esk. aki-), ayadwti,- 'adversary, enemy' (Greenl. akeraq), ay a-si- 'enemy; (drum) song' and ayayri-y du. 'twins', apparently the 'longitudinal, distant' equivalent of the 'transversal' iy-. Thus, it seems, the shibboleth is in some way embedded in the realistically abstract system of demonstrative orientation which pervades a large part of the Eskimo and Aleut vocabulary and forces upon the observer a strictly structuralistic attitude. If an inupiaq, yupik 'real man' presupposes some less real people such as the itqilit 'louse egg people', born of an outcast girl and a dog, an inuk, yuk 'living human being' obviously presupposes another, as a friend to be received face to face (iya-n) or as a rival to be pushed away (aya-). To the linguist, the shibboleth reveals a world of binary oppositions. The comparative problem includes no 'probability', for the various innovations turn out, on closer inspection, to be concatenated rather than random. The traditional 'Stammbaum' model could obviously not account for the intriguing similarities between the Asiatic dialects, especially that of Chaplino, and the Alaskan dialects south of Norton Sound. These are perhaps most reasonably explained as concatenated series of parallel development but the possibility of direct contacts, i.e., a constant traffic, across the ocean could not be entirely dismissed. Between St. Law-

220

KNUT BERGSLAND

rence Island and Cape Chaplin, a distance of about 50 statute miles, there was notoriously much traffic — trading and fight. Between St. Lawrence Island and the Yukon area there are about 130 miles of ocean but according to The United States Coast Pilot the Northeast Cape (1462 feet) of this very big island "can be seen on a clear day a distance of 35 miles [40 st.m.] or more" and Cape Romanzof on the Alaskan side is 900 feet higher. In the Aleutians there was also a notorious traffic between Kiska (Rat Islands) and the easternmost of the Near Islands (nearest to the Russian promyslenniki of the 18th century), Agattu and Shemya, a distance of some 160 statute miles with only the small, uninhabited, island of Buldir in between, and these islands are much smaller than St. Lawrence Island. Shemya is very flat, but Agattu, Buldir and Kiska have high mountains, respectively 2080, 2152 and 4004 feet, so in clear weather one did not loose sight of land. In the Aleutians, however,

221

THE ESKIMO SHIBBOLETH "INUK/YUK"

n» j

,«•1

< AGATTU BULOIR Om

&

EL

Si

KISKA ^J » «

1

1

TO

•

l

the fog is neither rare nor predictible, so the route was hardly safer than the hypothetical one between St. Lawrence Island and the Yukon area: from a geographical point of view, the problem of the interrelationship of the Eskimo dialects is a question of hours, fog and native daring. The archeaologist could not be of any great help, for sculls and harpoon heads do not speak dialects and one cannot dig in the sea. The linguist has to rely upon his own methods, i.e., upon a meticulous scrutiny of the linguistic facts. UNIVERSITY OF OSLO

V. BESEVLIEV

PROTOBULGARICA

1. E I N E N E U E

INVENTARINSCHRIFT

I

M Jahre 1964 wurde bei Feldarbeiten in der Ortschaft Souk Punar beim sogenannten Devedzi Jolu im Gebiet des Dorfes Dläzko, Kreis Sumen, ein Fragment von einer neuen protobulgarischen Inschrift gefunden. Es wurde von V. Antonova 1 in einer schwer zugänglichen lokalen Zeitschrift veröffentlicht. Leider ist die Publikation so unzulänglich, daß sie eine neue Veröffentlichung unentbehrlich macht. Die Inschrift ist auf einer viereckigen, 0,625 m hohen, 0,41 m breiten und 0,17 m dicken Kalksteinplatte eingemeißelt. Die Buchstabenhöhe schwankt zwischen 0,04 und 0,07 m. An der Rückseite befinden sich zwei Vierecke verschiedener Größe, die ineinander übergehen und mit einer gemeinsamen Rinne enden. Sie dienten wohl als Weinpresse oder zum Opfern. An den beiden Seitenflächen sind einige von den sogenannten protobulgarischen Zeichen und zwei Rosse ganz schematisch eingeritzt. Der Stein ist z.Z. im Archäologischen Museum von Sumen untergebracht. Das Fragment, das einen Teil von der rechten Seite einer größeren Inschrift darstellt, lautet:

OBAPCA HCOBPE I ZTHCT KEIX OP ATSEXI A O t % A IA; P N Z i k• E V Fl• A E• •N 1 "HOBOOTKPHT paHHo6i.jrrapcKH Haflimc OT C. ¿IJIMKXO", H3eecmm Ko/mpoezpad, III (1965), 25-29.

m

Hapoduun

My3eü-

223

PROTOBULGARICA

Bemerkenswerte

Buchstabenformen:

, O

(Zeile 4: O),

£;

Ligaturen:

H + C , $ • In Zeile 1 und Zeile 3 ist das Zeichen / K zwischen den Buchstaben P C

Z T

und £3 T

eingeritzt. Das Zeichen ist wohl später hinzugefügt wor-

den und hat mit der Inschrift nichts zu tun. Die Inschrift läßt sich folgendermaßen lesen und ergänzen: OBAPCA HCOBPEI ZTHCT KG l XOp-

224

v . BESEVLIEV

5

a xou

'¿%\ Ko-

[piKia p . . K a ] g [ i ] ö i a pv£'

KE VITAE>i Aus den Zeilen 5 und 6 ist zu entnehmen, daß das Fragment zu einer großen protobulgarischen Militär-Inventarinschrift gehört, deren Zahl bereits auf 7 steigt. Es ist schwer zu sagen, was am Anfang der Inschrift gestanden haben mag. Die Inventarinschriften beginnen gewöhnlich mit dem Titel einer protobulgarischen Militärperson, z.B. Pbl Nr. 48: ßayaxoup ßayaivou XopctKia fjv ojioß vy', KacriSia ne'. Die erhaltenen Buchstaben in den ersten drei Zeilen lassen sich zu keinem der bis jetzt bekanntgewordenen protobulgarischen Titeln ergänzen. Die von Antonova vorgeschlagene Ergänzung KoX]oßap[o]pBHTi GtjirapcKH xaHOBe", FoduuiHUKb na Hapodmn

My3eü 3a 1922-1925

1926), 2 7 9 f . , 301 f . ; "OcTartUH o n > e3HKa Ha ÄyHaBCKHTe npaötJirapH",

na öbmapcKUH apxeo/iozmecKu mcmumym,

(Sofia, Mseecmwi

V, 1928/29 (Sofia, 1929), 132-136;

J . N e m e t h , "IlpoH3xofli>TI> Ha HMeHaTa ' K o 6 p a T ' H 'EcnepHxi>'", M3eecmuH

Ha

ucmopmecKoeo dpyotcecmeo, XI-XII, 1931/32 (Sofia, 1932), 175-177. 2. Wahrscheinliche

Personennamen

BoptjQ, Bopiar| bleibt unklar. Darin steckt vielleicht die iranische Endung -an, oset. -on, altir. -äna. 36 Über die Etymologie von racrcsiç s. Zgusta, Personennamen, 89 f., 432; Abaev, 166. Eine andere Etymologie bei Duicev, 330 f. Kàpôafioç,37 bulgarischer Fürst (777-803), läßt sich mit dem iranischen Personennamen Kartham, Kapx|ir|ç38 vergleichen. Ob der byzantinische Personenname Kapôànriç, auf den Duicev39 aufmerksam macht, hierher gehört, ist unsicher. Dieser könnte eventuell auch iranischen Ursprungs sein. Der Name hat kein Gegenstück in den türkischen Sprachen. Über die türkische Etymologie siehe W. Tomaschek, Zeitschrift für die österr. Gymnasien, 28 (1877), 685. KoßpätoQ, Ko0ßpaTOQ,i0 bulgarischer Fürst in Südrußland ( | 642), der keine Parallele in den türkischen Sprachen hat, wurde von H. Grégoire41 mit dem iranischen Personennamen XopoaGoç oder XopouaOoç42 verbunden. Über die Etymologie des letzteren s. Zgusta, 169f.; Abaev, 168; M. Vasmer, Russisches Etymologisches Wörterbuch, 3. Bd. (Heidelberg, 1958), 261. Uber die Ersetzung des iranischen xdurch q- s. Menges.43 Die altbulgarische Form KoyptTb oder KoypTt in der bulgarischen Fürstenliste44 läßt sich aus Koußpcrroq über *KoyßpaTb > *KoyBpvrb unter Ausfall von -B- nach -oy- ableiten. Über die türkischen Etymologien s. Moravcsik, 161 f. MocmmbiS Icirgu Boilas bei den Zaren Symeon (893-927) und Petär (927-969) erinnert stark an die iranischen Personennamen Mocraoç 46 oder MCKTTOÇ, Macrxouç.47 Bei den letzteren Namen ist der Ubergang von a in o im Altbulgarischen anzunehmen. Die Endung -nib bleibt jedoch unklar. Über die Etymologie der iranischen Namen s. Zgusta, 116f.; Abaev, 173. Opoopmy, Q|iopxay, Mopiaycov, Moupxayov, 48 bulgarischer Khan (815-831), der weder ein Gegenstück in den türkischen Sprachen noch eine sichere türkische Ety35 30 37 38 39 40 41 42 43 44 46 48 47 48

CIRB, 859f. s.v. ; Zgusta, 89f. Nr. 94. Abaev, 225 ff. Moravcsik, 153. Justi, 158. Duicev, 332f. ; vgl. auch Moravcsik, 154. Moravcsik, 161 f. Byzantion, 17 (1944/45), 116, und La Nouvelle Clio, 4 (1952), 322f. CIRB, 907 s.v. ; Zgusta, 169 § 256. Menges, 108 und Anm. 1. Besevliev, Pbl, 306 Nr. 79, 4 und 318. Ebenda, 299 Nr. 69, 1-2. CIRB, 885 s.v. ; Zgusta, 312 § 675. CIRB, 882 s.v.; Zgusta, 115ff. § 146. Besevliev, Pbl, 251 f. ; Moravcsik, 217f.

PROTOBULGARICA

229

mologie hat, wurde von H. Grégoire 49 dem iranischen Personennamen Moup5ayoç 5 0 gleichgestellt. Der Anlaut O- macht dabei jedoch gewisse Schwierigkeiten. Nach Grégoire soll der Anlaut O- der griechische bestimmte Artikel sein, was aber zu verwerfen ist, da der Name auch bei den westlichen lateinisch schreibenden Chronisten Omurtag lautet. Auch die Vermutung von H. Menges, 51 daß es sich hier wohl um die 3. Person des türkischen hinweisenden Pronomen handelt, hat wenig f ü r sich. Es liegt hier wohl eher ein prothetischer Vokal vor, der in manchen iranischen Sprachen gelegentlich auftritt. 6 2 Es verdient dabei, hervorgehoben zu werden, daß eine Reihe protobulgarischer Namen wie (X)Kopar|ç, Ü)veyaßov[aT|]ix CJIOB B pyccKOM H3bnce K cpaBHHxejibHOMy 3THM0Ji0rmecK0My cnoBapio pyccKoro H3MKa", Hoeue donoAuemn u nonpaem (1905), 21. • Vasmer berichtete irrtümlich, daß Gorjaev der Entlehnung zugestimmt habe. ' A. FIpeo6pa»eHCKHö, 3muMOAOz. cAoeapb pyccK. H3. (1910ff.), s.v. 8 ü38. ornò. pyccK. H3„ XIX, 296ff. • M. Vasmer, Russ. etym. Wb., I (1953), 704. 10 W. A. Christiani, Über das Eindringen von Fremdwörtern in die russische Schriftsprache des 17. u. 18. Jh., Bln. Diss. (1906), 22. 11 A. Brückner, Slownik etym. jqz. polsk. (1927), 282.

H. H. BIELFELDT

234

aus deutsch Kunst- entlehnt, haben aber das 'n' bewahrt! Die Herleitung von Kycmäpb < Künstler ist m.E. falsch. — Andererseits können die bisherigen semantischen Begründungen für die Zugehörigkeit von Kycmapb zu Kycm so wenig überzeugen, daß wohl gerade das Vasmer bewog, zu der alten Herleitung aus dem Deutschen zurückzukehren. Schon in Efron's Enzyklopädie wurde abgelehnt: Kycm 'Strauch' wurde als Symbol der Familie aufgefaßt, weil der Strauch aus einer Wurzel in viele Zweige auseinanderwächst, wie die Familie mit ihrer Verwandtschaft; weil die KycmapnaH npoMbimjieHHocTb in der Familie betrieben wurde, sei die Wortbildung von Kycm her erfolgt. Kogen argumentierte: HCKOHHO POACTBCHHO CI» Kycm H nepBOHanaJibHO Morao o6o3HaiaTi> 'peMecjieHHHKa, 3aHHMaiomaroca o6pa6oTKofi KycroB'; aber diese Erklärung wird auch durch die Behauptung nicht überzeugender, die erste Art der Kycmapnan npoMbmijieHHOCTb sei das Korbflechten gewesen, und das Material waren Strauchgewächse, z.B. Weiden; denn noch viele andere Handwerke mit ganz anderem Material gehören von Anfang an zur KycmapnaH npoMMinjieHHocTb. — Das von Dal' neben dem für das Moskauer Gebiet verzeichneten KycmapmK

'Kleinhandwerker'

für

Sibirien

verzeichnete

xycmapnuK

'Gpo^ara,

KOToptifi npaneTCH B Jiecy no nycma\C ist eine andere Ableitung, von Kycmapb 'Strauch', und kann

selbstverständlich

die Verbindung

von

Kycmapb

'Klein-

handwerker' mit Kycm 'Strauch' nicht stützen. — Das als 'Übertragung' aus der botanischen

Bedeutung

'Strauch'

bezeichnete

Kycm

'rpynnoßoe

oöteflHHeHHe

(icaKHX-JiHÖo npeflnpHaTHH, apTejiefi, opramoanHH h T.n.)' findet seine Entsprechung in der gleichen Übertragung in französ. brauche, deutsch Branche, Zweig. Für diese Bedeutung von Kycm fand ich kein Zeugnis vor 1947 ;12 jedenfalls muß man von ihr bei der Erklärung von Kycmapb 'Handwerker' m.E. absehen. Kycmapb 'Handwerker' ist nicht aus dem Deutschen entlehnt, sondern von Kycm 'Strauch' abgeleitet. Aber hierfür fehlt bisher die Begründung aus der Wortbildung und Semantik. — Ableitungen auf -apb in der Bedeutung einer Person sind im Russischen von Substantiven, Adjektiven (dmdpb—dmuü)

und Verben (nücapb—

nucämb) gebildet worden. Auch innerhalb der Ableitungen von Substantiven ist das semantische Verhältnis zwischen dem zugrunde liegenden Substantiv und der A b leitung mit -apb verschieden, vgl. z.B. cnoeäpb (in dem Wörter enthalten sind) — cAÖeo; epamäpb (der mit dem T o r zu tun hat) — epamä; z/iaeäpb (der das Haupt ist) — znaeä. Die Ableitung auf -apb ist heute im Russischen wenig produktiv. Von den im 11.-14. Jh. im Russischen bezeugten über 40 Bezeichnungen handelnder Personen mit -apb13 aus Substantiven sind nur fünf bis in die gegenwärtige Schriftsprache erhalten, und auch ihr Gebrauch ist sehr begrenzt: Bpamapb und ncapb muß man fachsprachlicher Terminologie zurechnen; auch 3onomapb, soweit es nicht veraltet ist; KopMuäpb ist veraltet; Mbimäpb gilt nur für den berühmten Zöllner des 12

Cdoeapb coep. pyccK. jium. H3., 5 (1956), s.v.

H. B. HypMaeBa, "CymecTBHTejibHue c cy(J)(J)HKCOM -apb co SHaHeHHeM aeäcTByiomero jnma BflpeBHepyccKOMH3buce XI-XIV BB.", Hcc/iedoeauuH no ucmopmecKoü AeKcmoAozuu dpeeuepyccKoao H3bwa (M., 1964), 260-271. 13

235

RUSSISCH "KYCTÄPb, KyCTAPfl : KyCT"

Neuen Testaments. Die anderen bis ca. 1400 so zahlreich bezeugten Wörter auf -apb verschwanden oder wurden durch Bildungen mit anderen Suffixen ersetzt. Doch treten nach dem 14. Jh. noch wieder neue Ableitungen mit -apb auf, darunter Entlehnungen aus dem Polnischen und Deutschen; für die erst nach dem 14. Jh. erstmalig bezeugten Bildungen auf -apb fehlt eine Zusammenstellung. Kycmapb ist vom 15.-18. Jh. reich bezeugt in der Bedeutung 'mit Büschen bewachsenes Gelände, Anzahl von Büschen'; diese Ableitung mit -an hat also eine von Kycm 'Strauch' nur wenig abweichende Bedeutung. Dagegen liegt in Kycmapb 'Kleinhandwerker' ein ganz anderes semantisches Verhältnis zu Kycm vor, also eine andere unabhängige Wortbildung. Kycmapb 'Kleinhandwerker' ist erst im 19. Jh. bezeugt; in der genannten Abhandlung 13 wird keines der beiden Kycmapb erwähnt. Die in Kycmapb 'Kleinhandwerker' vorliegende Wortbildung und Bezeichnung beruht auf der Vorstellung der Gleichheit bzw. Ähnlichkeit des Kleinhandwerkers mit dem 'Strauch'. Die Ähnlichkeit zwischen beiden besteht in der UnVollständigkeit: in der Geringheit des niederen Strauches im Vergleich zum höheren Baum und der nur nebensächlichen Charakter tragenden Tätigkeit, geringen Leistungsfähigkeit, usw., des Heimhandwerkers im Vergleich zur Haupttätigkeit in der großen Werkstatt. Diese Erklärung wird übrigens in Efron's Enzyklopädie erwähnt, aber nur, um sie abzulehnen: "Kycmapnax npoMbiuuieHHOCTb, KaK 6oJiee Mejiicas, ynojioßjiHeTca KycmapmiM pacTeHUHM no cpaBHeHHio c 6oJibIUHMH

flepeBbaMH,

KOTOPHM cooTBeTCTByioT KpynHbie —

Hanp. m / — Fem + Obi + Gov

(R 5) Gen. Masc/Neut. Sing. '-Adj " + Det 0 s / - Plur - Fem + Obi - Gov

(R 6) Nom/Acc. Fem. Sing./Plur. + Det ~ - Obi + Plur 0 e/ Plur + Fem

Det " Plur Masc Obi Gov _

252

MANFRED BIERWISCH

(R 7) Gen/Dat. Fem. Sing./Gen. Plur. + Det + Obl ' + Plur r / 0 - Gov Plur + Fem

(R 8) Dat. Plur.

0

n /

+ + + +

Det Plur Obl Gov

The rules (R 1) to (R 8) clearly correspond to the eight syncretism fields of pattern (5) above. The indication of [— Adj] in (R 5) is needed because only proper determiners have /er/, while adjectives take the affix /en/ in this case: dünnen Weins, not *dünnes Weins "diluted wine". We will account for these adjective forms in a later reformulation of the rules. Each of these rules must be applied only once to a given complex symbol, because otherwise we would derive such impossible forms as *diesererer etc. This convention is for the time being only an ad hoc stated condition. We will see later that it is deeply connected with basic principles of morphology and grammar in general. The above rules can be somewhat reduced by certain notational transformations. Thus, if we allow two rules such as (13) (a) to be collapsed to (13) (b): (13) (a)

A A

B / CX_ E / CZ_ B / E /

(b)

X_

z_

_YD JJD _Y _U / C_

_D

then we can also restate with a slightly different convention, for instance, (R 1) and (R 2) as (R 1'): (R 1') Nom/Acc. Masc. Sing.

0

Jr / [ - Gov]} \n / [ + Gov]}

1

+ + -

Det Plur Masc Obl

Notational conventions of this kind are by no means arbitrary. They must be chosen in such a way that they yield a simpler formulation if and only if this simplification expresses a significant generalization.28 The principle of double environments, illustrated by (13) (b) and (R 1') is of this type, as is known from other evidence. Of course, (R 1') allows no judgment as to the relevance of the sparing of four features. We could further reduce (R 1) to (R 8) on the same principle by collapsing 28 For detailed discussion of this problem see N. Chomsky and M. Halle, "Some Controversial Questions in Phonological Theory", in: Journal of Linguistics I, Nr. 2, 1965, and N. Chomsky and M. Halle, The Sound Pattern of English.

253

SYNTACTIC FEATURES IN MORPHOLOGY

them all to a single rule, extracting the recurrent feature [ + Det] from all the environments and stating it as the general environment for the whole collapsed rule. We would save then seven further features. Reducing the collapsed rule as much as possible by means of split environments we come to the following complex rule, which also accounts by subrule e) for the affix /en/ of genitive singular masculine and neuter: (IR 1') Pronominal Inflection r / [-Gov]} T _ " n / [ + Gov]}' |_+ Mase s / — Mase] — Fem

J

m / [ + Gov] in / [ + Adj ]} ]. T ] (s / [ - Adj ]] J [ - Gov J •

e I

r /

+ — —

+ n /

+

Plur Plur] Fem Obi Plur] Gov Plur = Fem Plur= Gov

- Obi Plur Fem Obi + Det

fe]

This formulation clearly shows the greater differentiation of inflection within masculine and neuter singular as opposed to feminine and the entire plural: 13 syntactic features are needed for subrules a) through f) and only 11 for the rest. This corresponds somehow to the principle of greater syncretization within marked categories. Comparing now (R 1) through (R 8) with (IR 1') we see that instead of 49 syntactic features in the former formulation only 26 are needed in (IR 1') which in addition accounts for the inflection of the genitive of adjectives. (We are counting thereby 0 as a syntactic feature and we will replace it below by an overt feature. If we were to ignore it, we would have 41 vs. 25 syntactic features. The number of phonological features is 11 and 12 respectively.) This is obviously an important simplification. But even this leaves us with a rather complex set of subrules. One should therefore think of a simplification which is explicitly based on the well-known interplay of markedness and neutralization.

254

MANFRED BIER WISCH

4.3. Neutralization Rules. — If a certain distinction is suspended in the environment of a given category C, then, according to a widely assumed principle, it is either neutralized only for the marked value of C or for both the marked and the unmarked value, but never for the unmarked value alone.29 Thus gender is neutralized in the pronominal inflection of German in the environment [ + Plural], the distinction between governed and non-governed cases is neutralized for the direct, but not the oblique cases in the neuter and in the plural, and in the presence of [ + Feminine] also for the oblique cases. It is never neutralized for the oblique cases only. By the way, one may ask whether this principle is valid since, as was already pointed out, the syncretism of nominative and accusative in the neuter and their distinction for the masculine would be a counterinstance given that the neuter is the unmarked gender. However, the principle in question states only a condition on neutralization in case there is one. It says nothing about whether and where neutralization really takes place in a given language. This means that morphological neutralization in particular languages cannot be accounted for by universal conventions but must be expressed, if necessary, by particular rules. These rules would be for the pronominal inflection as follows: (NR 1)

[ + Plur ]

(NR 2)

Masc"| Obi

(NR 3)

J

[ + Fem ] -

[=

— Mase Fem

[ - Gov ]

[ - Gov ]

These neutralization rules, of course, must precede the inflectional rules and (NR 1) must precede (NR 2) and (NR 3), the ordering of the latter two being immaterial. (NR 1) through (NR 3) is the most economical set of neutralization rules I have been able to find. Let us see now what the incorporation of them into the morphological component of a German grammar would save with respect to the proposed inflectional rules. Since after (NR 1) [ + Masc] and [ + Fem] never occur in the environment of [ + Plur], the specification of [— Plur] in (R 1), (R 2), (R 6), and (R 7) becomes superfluous. In the condensed form (IR 1') [— Plur] is abundant only in the subrules g) and h). However, the saving of four or two features respectively is at the expense of three feature specifications in (NR 1). Therefore it yields no simplification. (NR 2) and (NR 3) do not save any features and are therefore completely worthless. More important than feature counting, however, is the fact that many of the neutralization phenomena are already handled in a very general and more natural way by the principle of reference to natural classes only and partly by the use of double 20

See, for instance, J. H. Greenberg, Language Universals. A similar principle follows from Hjelmslev's laws of suspension and syncretization, as far as I understand them. See La Catégorie des Cas. Jakobson, in his "Beitrag zur allgemeinen Kasuslehre", is more careful in hinting only at certain asymmetries in inflectional patterns.

255

SYNTACTIC FEATURES IN MORPHOLOGY

environments. Thus the irrelevance of gender specification for the plural forms is accounted for by referring to the plural only by [ + Plural] without mentioning the gender features. By this principle, moreover, we have dispensed with postulating ad hoc that the gender appearing in the plural is the neuter as is stated in (NR 1). This observation, by the way, holds also for other inflectional patterns of German. Instead of using neutralization rules, we have to look for a more adequate way to improve (IR 1'). 5. A G R E E M E N T R U L E S

5.1. Feature Duplication. — Before further discussing the inflectional rules I will outline the rules which have been presupposed until now, namely those accounting for concord in German noun phrases. As mentioned above case-, number-, and gender-features are partly lexicon-inherent, partly base rule inherent, and partly transformationally introduced (only for nouns). From nouns they must be transferred to determiners and adjectives. This will be accomplished by a rule of the following form: (T 1) Noun Phrase Agreement

SD:

1 SC:

f [ + Det]l _

W — [np X

1 2

l t + Adj]f

2 3

3 4

5

6

4

"a Plur " ß Masc y Fem - Z np] S Obi e Gov " 5

U

7

By (T 1) the specified features of the fifth element in the structural description, which is the only noun directly dominated by the noun phrase in question, are added to the determiner and to all adjectives belonging to the same NP. The placement of (T 1) among the transformational rules is not a trivial question. (T 1) has to follow on the one hand the rules which introduce case features, including the nominalization rules, and also the relative clause rule and those for the reduction of relative clauses to attributive adjectives. On the other hand in order to get the correctly inflected forms in noun phrases without a noun, it has to precede pronominalization and deletion of lexically empty nouns. That is, (T 1) must be part of the transformational cycle. We cannot go into the details here any further. 5.2. Inflected and Uninflected Forms. — As already pointed out adjectives must be inflected only if they are dominated by a node NP in the surface structure. It must be noted that determiners are sometimes not inflected either, or have, as it were, zero inflection. The indefinite article ein, the negative kein and all possessive pronouns

256

MANFRED BIERWISCH

mein, unser, dein, etc. remain without an affix in the masculine and neuter nominative and neuter accusative. These facts require an indication whether adjectives and determiners have to undergo inflectional rules or not. This can be done by introducing a feature [ + Inflection] under appropriate conditions making it part of the context restrictions in the inflectional rules. The rules providing this feature are somewhat clumsy because of the special environment. (T 2) and (T 3) are the simplest formulation I was able to find. We make use here of Boolean functions of features. The notation is to be understood as: F and/or G must be present, not necessarily both, but at least one. (T 2) Uninflected Determiners

[ + Det]

[ - Rule (T 3)] / [ N P

+ Indefinite + Negative + Possessive — Plural — Feminine — Oblique •

UNPRODUCTIVE

z o p prekratil and vice versa in case of a reverse conversion. It is believed that semantic competition between infinitive (representing the verb as a word class) and noun can be ascertained in most of the other infinitive occurrences listed by van Hoik under the label of zero competition. Thus, by applying different transformational operations, such as replacement {a => b), expansion (a => b + c), reduction (a + b c), and addition {a => a + b), to van Hoik's illustrative examples one obtains the following conversions: (1) On perestal kurit' => On brosil kurit', a => b; On brosil kurit' => On brosil kurenie, a => b ; On brosil kurenie => On prekratil kurenie (a => b). The direct transformation On perestal kurit' => On prekratil kurenie would have to be interpreted as representable by a formula a + b => c + d, implying some sort of 'complex replacement' not particularly specified in standard transformational grammar.27 However, as shown by the above series of transformations, such 'complex replacement' can be derived from a sequence of simple replacements; cf. also our preceding remarks on the semantics of the context). (2) On vysel guljat' => On vysel dlja togo ctoby guljat' (a => a + b); On vysel dlja togo ctoby guljat' => On vysel dlja guljan'ja (or, rather, progulki; a + b => c). (3) Ser'ëzno ja govorit' ne umeju => Ser'ëznogo ja razgovora ne znaju (or, with additional permutation: a + b => b + a, Ser'ëzno ja govorit' ne umeju => Ser'ëzno govorit' ja ne umeju; Ser'ëzno govorit' ja ne umeju => Ser'ëznogo razgovora ja ne znaju). Here, again, the nominalization (govorit' => razgovora) is concomitant with a shift of the governing verb {ne umeju => ne znaju) so that we have to do with a 'complex replacement' (o + b => c + d) rather than with a 'simple replacement' (a => b); cf. also the same phenomenon in instance (1) above. (4)

Cto ze vy obedat' ne idëte? => Cto ze vy na obed ne idëte? (a => b + c).

(5) Ona ob odnom tol'ko mectaet: poexat' za granicu => Ona ob odnom tol'ko mectaet: (o) poezdke za granicu {a => b, or if the preposition is repeated, a => b c).

minimal grammatical information. It says nothing either about the participant of the narrated event or about the relation of this event to other narrated events and to the speech event. Thus the infinitive excludes person, gender, number, taxis, and tense" (cf. note 15 above). See also id., "Zur Struktur...", p. 77; for similar interpretations of the functional (informational) load of the Russian infinitive, see further A. V. Isacenko, Die russische Sprache..., p. 346, and O. S. Axmanova, Ocerki po obScej i russkoj leksikologii (Moscow, 1957), p. 108. 27 Cf., e.g., E. Bach, An Introduction to Transformational Grammars (New York-Chicago-San Francisco, 1964), pp. 70 and 74.

PREDICATION AND THE RUSSIAN INFINITIVE

289

(6) Vse vyxodili v sad, grefsja na mesjace => Vse vyxodili v sad, dlja grenija na mesjace {a => b + c)-28 (7) Byf starym — eto tjazelo => Starosf — eto tjazelo (a + b^=> c). Or, perhaps somewhat more naturally, Stanovifsja starym — eto tjazelo => Starenie — eto tjazelo, a + b=> c. With starosf 'old age' a more natural phrase would be Starosf — tjaiela, where, however, by omitting the pronoun eto, the anticipative character of the infinitive phrase and its transform would be lost. As for the rest of the infinitive expressions, i.e., the conditional infinitives (8), the repetitive infinitives (9), and the twin infinitives (subject constituent, 10a), listed by van Hoik as having zero competition but being stylistically restricted, only the subject constituent infinitive of the twin infinitive construction can clearly be considered as being in semantic competition with a noun while of the two other infinitive uses only the 'repetitive infinitive' can be said conditionally to compete with the use of a noun (cf. below). None of these infinitive uses belong under the heading of zero competition, where they now appear in van Hoik's diagram. As concerns the conditional infinitives (without a subordinator; type Posmotref, tak eto cudo), they should be considered but a variety of the clause-equivalent infinitive phrases (van Hoik's 'subordinate infinitive clauses') introduced by esli or some other conditional conjunction. 29 Thus, e.g., the relationship between four semantically equivalent expressions can be represented as follows: Poslusaf, kaietsja odna u nix dusa = Esli poslusaf, kazetsja odna u nix dusa :: Poslusaef (i) kazetsja odna u nix dusa = Esli poslusaes\ kazetsja odna u nix dusa.30 28 On the transformational rules involved in the nominalization of Russian reflexive and nonreflexive verbs (into deverbative nouns), including some pertinent comparison with related phenomena in Polish, see D. S. Worth, in American Contributions to the Fifth International Congress of Slavists, Volume I: Linguistic Contributions (The Hague, 1963), pp. 374-375. 29 For further examples, see A. G. F. van Hoik, op. cit., p. 59. 30 The transformational relationship (in the sense of Z. S. Harris) obtaining between these four sentences can be represented by the diagram appearing on the next page. Symbolization of types of transformation:

replacement: a=> b addition : a => a + b (or b => a + b) deletion : a + b => b (or a + b a)

expansion : a=> b + c reduction : a + b => c

The formulae in parentheses need be considered only if a + b implies a fixed order of succession. Cf. also E. Bach, op. cit., pp. 70-82 ("Derived P Markers"). — T h e above scheme represents but a fragment of a 'transformational derivation' (as described in N. Chomsky's earlier work on syntactic structure), that is, it does not provide a framework for the complete mapping of the sentence-structures involved. All four sentences, being complex ('subordinate clause + main clause' and 'clauseequivalent infinitive phrase + main clause', respectively), must be considered non-kernels (cf. note 23 above), arrived at by a generalized transformation from two (or even three, cf. below) underlying kernels: {ty)posluSaes' + kaietsja (+) odna u nix duSa, concatenated by the semantic factor of conditionality ('if ... then ...'), which lexically is normally realized by a subordinate (conditional) conjunction (esli, i f ) , but, in theory at least, also could be verbalized by some adverbial phrase such as 'on the condition of': *'On the condition of your listening they seem to have (but) one soul'. Strictly speaking, one may even have to assume three underlying kernels, since kaietsja odna u nix duSa is

290

HENRIK BIRNBAUM

As opposed to the esli + infinitive phrases, the conditional infinitives without a subordinator are in fact stylistically restricted (though only in the 'weak' sense as defined above, i.e., as belonging to the stylistic domain of lively speech in general rather than to the domain of folkloristic — proverbial, etc. — style). Regarding the so-called repetitive infinitives (type ¿if ne zivet), it may seem hard to classify this special, stylistically conditioned use of the Russian infinitive. It serves primarily to express a certain emphasis, anticipating and summarizing an action or a paratactic structure combining two predications: one expressed by kazetsja and one implied in the zero form of the copula: odna # u nix dusa or odna u nix # dusa (also, with permutation: a + b=> b + a, duSa u nix # odna). The paratactic structure is interchangeable in CSR with the hypotactic one: kazetsja, cto odna u nix dusa.

PREDICATION A N D THE RUSSIAN INFINITIVE

291

state-of-being to be further qualified (as to person, time, and modality; i.e., as to the basic predicative categories) by the finite form of the same verb only in the continuation of the sentence. Thus, the use of the infinitive is here a particularly clearcut example of its serving as an unqualified and unspecified token representing the verb system as a whole. Semantically speaking, therefore, the 'repetitive infinitive' competes with a noun (usually a deverbative noun, thus an instance of nominalization) occurring in a prepositional phrase or, to be exact, its semantic (though not syntactic) equivalent: ¿if ne zivet => Cto kasaetsja zifja (zizni), to (on) ne zivet. Similarly, Ponimat' ne ponimaju => Cto kasaetsja ponimanija, to ja ne ponimaju. All such transformations are expansions (a => b + £')• Cf. further also a series of transformations such as ¿if ne zivet => *Prazdno iif (on) ne zivet => Prazdnoj zizni (on) ne vedet (but hardly without any qualification: ¿if ne zivet => *2izni (on) ne vedet). Notice, on the other hand, that if one adds a cto as a subordinator in the (to be sure, rare) sense 'as for', the infinitive expression assumes predicative function: Da cto smotref — to ty sam smotri (quoted, after van Hoik, from Tolstoj). The exact function of the 'repetitive infinitive' needs therefore to be further investigated. It also deserves mention that the distinction between a preposed completive (or completivesupinal) infinitive and the 'repetitive infinitive' is not always very sharp and easy to establish, in particular in cases of ellipsis. Cf., e.g., Proigryvaf ja ne umeju (preposed completive infinitive) vs. Proigryvaf, ja ne znaju kak (to wit, proigryvaf; repetitive infinitive). Finally, as concerns the twin infinitive construction, it was already mentioned that the first of these infinitive occurrences can be considered an equivalent of a noun (Volka bojafsja ... => Bojazri volka ...; ¿izri proiif ... => Prozitie zizni ..., etc.) while the second infinitive occurrence can hardly be regarded simply as a substitute for a finite verb; rather, it belongs to the predicate constituent in the same way as a predicative noun (in English school grammar often misleadingly called 'predicate nominative'); i.e., it completes the meaning of a linking verb in the broad sense, including synonyms and substitutes of the copula verb 'to be'. Here, one has to allow for a considerable variety in the scope of the semantics of the copula verb, centered around such meanings as 'implies, means, equals, amounts to, is tantamount to', etc. (Russian znacit, ravnjaetsja, — to ze cto, etc.): Volka bojafsja (znacit) v les ne xodif; ¿izn' proiif ne (to ze cto) pole perejti, and so forth. The predicate constituent infinitive of the twin infinitive construction must, therefore, be considered as being in syntactic (and semantic) competition with a — predicative — noun. 14. To sum up, then, the preceding considerations, it can be proposed that the structure of the total semantic spectrum of the Russian infinitive (within which predicative function — in its three major syntactic contexts and further subclassifiable in terms of more subtle syntactic-semantic categories — forms an integral part) be represented somewhat differently from the way elaborated and diagramed by van Hoik. Omitting here the subcategories just indicated as well as some other details

SPECIAL ORIENTATION » COMPETITION SEMANTIC Zero

o

H £ w t— «« O j

< w

• H U £ Q O

Vb. (Inf.) vî. Noun Inf. vs. Fin. vb. Vb. (Inf.) vs. Noun Inf. vi. Fin. vb. (1) Completive (11) Subject (2) Completive- ^ ^ int. m clause inf. supinal HHl equivalent^ (3) Supinal phrases with? (4) Explicative (5) Anticipative

ÎH

Z h-t H
al-sadr, lit. 'noise from the chest') is specifically mentioned in connection with some of the majhura, and as being absent from the mahmusa, and voicelessness (§awt'al-fam, lit. 'noise from the mouth') in connection with the mahmusa and, less clearly (fuwayt min 5al-fam, lit. 'a small noise from the mouth') in connection with the pausal allophones of q, j, t, d and 6.17 Thus while the phenomenon of voicing was perfectly well-known to Slbawayhi and his teachers (though the action of the vocal cords may not have been recognized as its source), one gets the impression that it was considered irrelevant or unnecessary to mention it in the catalogue of distinctive features, a point to which I return below. However 15 Thus first in 'al-Sirafl's commentary in the 10th century (Fleisch, 1958, 206, fn. 1), and later in 'al-'Astaraba6f, 13th century (ibid.); in 'al-ZamaxSari, 12th century (Cantineau, 1946, loc. cit.), etc. 16 A detailed analysis of Slbawayhi's description of pausal allophones appears in my "Fronting", sec. 4.46. The text on which it is based is Kitab II, 310,7 to 311,8; cf. also Schaade (1911), 10; Bravmann (1934), 25. 17 Cf. my "Fronting", sec. 4.3, 4.46. The identification of these terms for voice and voicelessness, though mooted previously (cf. Schaade, 1911, 6-7) is due largely to Fleisch (1949-50), 233f.; (1958), 200f.

THE 'SONOROUS' VS. 'MUFFLED' DISTINCTION IN OA PHONOLOGY

301

that may be, at least two authoritative later sources, one of which quotes Sibawayhi himself, explicitly give presence of voice as a characteristic trait of the majhura and absence of voice as a characteristic of the mahmusa. These have been brought to light and minutely examined by Fleisch (1956, 1958) and his conclusion has already been mentioned in sec. I above. A few comments on these texts follow: (1) The later of the two statements is due to 'al-'AstarabaQl (13th century) and was presented and analyzed by Fleisch (1958); the bulk of it was also translated and analyzed by Alarcon (1925), though its exact source was not known until Fleisch's work. 18 It describes in some detail the various characteristics of the dichotomy (including presence or absence of nafas, full or weak contact, presence of tawt'al-sadr vs. sawt'al-fam) and clearly reflects earlier doctrine, including the statements found in the passage attributed to Sibawayhi by 'al-Sirafi. That passage was found by Troupeau (1958) and Fleisch (1958) in 'al-Slrafl's unpublished commentary, 19 and published by Fleisch in footnote 3 of his 1956 book. It is worth translating in full; the following attempt adopts Fleisch's readings where the mss. vary or are unclear. My own interpolations are in brackets. Said Sibawayhi: What distinguishes the majhura from the mahmusa is that you cannot pronounce the majhura clearly unless it includes noise-from-the-chest [viz. glottal tone or 'voice']. All the majhura are of this sort; the sound [that accompanies them] coming out of the chest and flowing in the throat. However, the m and the n have their sound coming out of the chest and flowing in the throat and the nasal cavity, so that what flows in the nasal cavity becomes a nasal twang that mingles with what flows in the throat. This is shown by the fact that if you hold your nose and pronounce either of these sounds, you will find them both distorted. As for the mahmusa, their sound comes from their points of articulation [only], and that is what brings out the sound. Their contact ['i'timad, lit. 'pressing, leaning against' see fn. 12] is not like that of the majhura, and so the sound is brought out of the mouth weakly. This is shown by the fact that if you speak softly, you whisper these [same] consonants, something you cannot do with the majhura. If you say Saxsun, what produces these consonants [viz. s, x and i] is noise-from-the-mouth, though they are followed by noise-from-the-chest [viz. the voice of the ensuing vowels] after having been produced by noise-from-the-mouth, so that they can be perceived and understood. In such cases the noise-from-the-chest is the same as the voice which you raise after [consonants produced with] noise-from-the-chest: note that you say qadama either softly or aloud, but if you say it aloud you produce a different sound [than if you whisper, unlike the mahmusa, which sound alike whether you whisper or speak aloud]. (2) The reference to voice vs. voicelessness as a feature of the dichotomy is, as 18 Alarc6n (1925), 191-192, translates a passage which he found in Sprenger (1862), I, 321-322. The work is an 18th century dictionary of Arabic and Persian technical terms in which the definitions are taken from authoritative (but usually unidentified) sources. The passage is identical with that quoted by Fleisch, though the first few lines, which contain nothing essential, are missing in Sprenger. 19 Troupeau (1958), gives a thorough analysis and many quotes, in translation, of a portion of the commentary dealing with phonetics. On pp. 176-178, the matter that concerns us here is dealt with, and three statements, of which two hark back to Sibawayhi himself, are mentioned; one of these is the passage published by Fleisch and translated below. From a remark on p. 181, it appears that 'al-Sirafi held this statement from 'Abu 'al-Hasan 'al-'AxfaS, Sibawayhi's most distinguished pupil and first promulgator of the Kitab.

302

H. BLANC

Troupeau and Fleisch have pointed out, quite unmistakable. However, the passage contains another indication that is just as clear and compelling, and that explains the final words of the quotation from the Kitdb adduced in par. (A) above, where the terse statement that consonants can be pronounced either aloud or in a low voice is at first sight puzzling. Voicelessness, we are now told, can be heard in the mahmusa at all times, no matter whether spoken aloud or whispered; on the other hand, voice can be heard in the majhura only when spoken aloud or clearly, but if whispered they sound different, i.e., have little or no voice.20 Since the analysis of pausal allophones (cf. above) also indicates that at least some of the majhura were partly or wholly devoiced in pause, it is no wonder that Slbawayhi did not use presence vs. absence of voice as the single overall distinctive feature of the dichotomy in the succinct statement of the Kitdb. That is why he states, in discussing the presence or absence of nafas, (cf. the quotation in par. (A) above) that consonants can be pronounced either softly or aloud, the implication being that in either case the majhura remain distinct from the mahmusa by virtue of the feature of nafas, which is not quite as true of the feature of voice. In short, the literal intention of the term mahmusa may well have been 'whisperable', viz. "sounds that can be whispered [without changing them]", a sense quite in keeping with the uses of the passive participle; conversely, the literal intent of majhura may have been "sounds that CAN be spoken with jahr, i.e. loudness, sonorousness, resonance". (3) When we are told that "all the majhura" need to be voiced when pronounced distinctly or aloud, must we conclude that this holds literally true for hamza, q and t as well? Fleisch and others answer in the affirmative in the case of /, 21 almost as positively in the case of q,22 and a voiced hamza is thought possible by some, 23 impossible and hence erroneously classified by others. 24 That hamza was classed with the majhura because it was often reduced to a mere vocalic onset or interlude is possible; that it fits the characterization of the Kitdb in that it lacks nafas is certain. As for q, that it had voiced allophones in Slbawayhi's time is clear, that it had voiceless allophones as well is practically as certain ;2B it fits the characterization of the Kitdb by virtue of its unaspirated character. 26 As for (, it is examined separately below. 20

Fleisch (1958), 207, understands the matter of speaking aloud vs. speaking softly somewhat differently. The two additional passages adduced by 'al-SIrafl (cf. preceding fn.) refer less clearly to the impossibility of pronouncing the majhura in a low voice (scil. without changing or devoicing them) but the text of the Kitdb, plus the elucidation furnished by the passage just translated, are obviously more authoritative. 21 Fleisch (1949-50), 236, fn. 1; Schaade (1911), 13; Cantineau (1946), 100. 22 Fleisch, ibid.; cf. also Cantineau (1946), 103; Schaade (1911), 14. 23 Cantineau (1946), 117, 118; Garbell (1954), 233. 24 E.g., Fleisch, ibid.; Schaade, 14-15. 25 The pronunciation of q in Old Arabic is one of the central questions to which my "Fronting" addresses itself; cf. esp. sec. 4.4, 4.7. Ancient and modern evidence is adduced to show that, before the 10th century, q was predominantly pronounced as an unaspirated postvelar stop with voiced [G] and voiceless [q] as main allophones. 26 Evidence for unaspirated q (vs. aspirated k) may be found in my "Fronting", sec. 4.7, 5.1.

THE 'SONOROUS' VS. 'MUFFLED' DISTINCTION IN OA PHONOLOGY

303

(4) Slbawayhi says (II, 455,9) that t without its 'itbaq (velarization) would be d, just as s would be s and d would be d. This seems to point to a voiced t, as does the occurrence of a voiced t in some Yemen dialects (cf., e.g., Schaade, 1911, 13). In point of fact, the Yemenite t is not voiced, but voice-indifferent, i.e., has both voiced and voiceless allophones, and is, like d, merely unaspirated and lenis (cf. e.g., Garbell, 1958, 307). This might well have been what Slbawayhi had in mind, as was shrewdly pointed out by Meinhof (1920/21, 94, fn. 1). There are other dialects27 in which t, though voiceless, is unaspirated, like d, but unlike t, so that it is not exactly the 'emphatic' counterpart of either. Another statement of Slbawayhi's (II, 452) lists, in a brief survey of non-canonical variants, "a t that is like /". This could mean either a t that was partly or fully de-emphaticized (in which case the equation '/ minus emphasis = d' is contradicted) or it could point to the existence of a fully voiceless t, as in many dialects today and as attested to by other medieval sources (cf. sec. Ill below). In either event, it is clear that t was not purely and simply a voiced stop, though it is still possible that Slbawayhi had this in mind for his standard pronunciation. The evidence from ancient transcriptions and loans from and into Arabic weakens the case for voiced t still further. Arabic t never renders foreign [d], but always foreign [t], often demonstrably unaspirated, 28 and vice-versa, Arabic t is rendered by foreign t, not d, except in the Spanish area, where t is rendered now by t, now by d, sometimes in the same word (Steiger, 1932, 149-160). Persian, Turkic and other Islamic alphabets regularly use the Arabic letter t for voiceless sounds, never, to my knowledge, for voiced ones. No present-day dialect so far described has a fully voiced t. (5) In view of all that has been said above, it is hard to avoid the feeling that the omission of voice from the characterization of the majhüra-mahmüsa distinction in the text of the Kitab is not accidental. The terms themselves, which can be rendered as 'sonorous' and 'muffled', are no doubt strikingly similar to terms for voiced and voiceless in several other traditions 29 and may, indeed, have been coined primarily to express such distinctions as that between [z] and [s], [d] and [t]. Certainly, Slbawayhi was familiar with the phenomenon of voicing, and aware of the connection between it and the dichotomy we have been studying. We can only guess at the reasons why he chose (assuming he did choose) not to mention voice as a criterion for differentiating between the two series, but the passage reported by 'al-Síráfí indicates rather plainly, I think, one reason for the choice; so does his treatment of pausal allophones. He was aware of the fact that t is always distinct from d but 27 E.g., Mesopotamian and some Beduin dialects, cf. refs. in my "Fronting", sec. 5.1. In such dialects, an emphaticized [t] is thus not exactly [{], being aspirated, nor a [d]. A Baghdadi [tl has (on me) the acoustic effect of being midway between [(¿] and [t]. 28 Cf. my "Fronting", sec. 4.7, and references cited there; different explanations of why t renders unaspirated Latin and Greek [t], are given by Cantineau (1946), 118; Fleisch (1949-50), 235-236. 29 Alarcón (1925), 290, fn. 1, duly notes this for Spanish 'sonoras' and 'sordas'; Cantineau (1946), 118, does the same for French 'sonores' vs. 'sourdes'. Others may be added, e.g., Russian zvonkie vs. gluxie, outmoded English 'sonant' vs. 'surd', though of course all may hark back to a single origin.

304

H.BLANC

whereas t is always voiceless, d is not always voiced, but might, in pause, in whispered speech and perhaps elsewhere, be partly or wholly devoiced [d].30 The 'invariant' feature, then, that distinguishes t from d is something else than voice, though voice is an important additional feature. Sibawayhi hit upon presence vs. absence of (pure) breath, and it has been seen how this can be interpreted. This would imply, inter alia, the existence of phonological thinking more than a millenium before the formulation of the distinctive feature principle, but I think the implication would nowadays seem quite acceptable to most linguists. (6) There remains the troublesome matter of full vs. weakened articulation. If t and d remain distinct even when d is devoiced, the contrast fortis vs. lenis would almost inevitably come to mind as the distinctive factor. According to Cantineau's reading of the text, this is precisely what Sibawayhi was saying, except that it was his d that was tense and his t that was lax. The reasons for doubting this reading, though it is not ruled out a priori, have already been stated. A more clearcut indication, describing a fortis-lenis relationship more compatible with general experience in Arabic and contrary to Cantineau's reading, can be found in Ibn Sina.

III. IBN S l N A

In the early 11th century, the great scientist and philosopher Ibn Sina (Avicenna), composed an independent description of the production of speech sounds, concentrating on Arabic, but adducing examples from Persian and other languages.31 His training and point of view are physiological, anatomical and empirical; his terminology has little if anything in common with Slbawayhi's (one gets the impression that he never read the Kitab), he does not try to erect any sort of system, and his account is the only one known to me that is completely independent of Arab grammatical tradition. The Arabic sounds he describes are, in some cases, different from those described by that tradition and closer to present-day pronunciations.32 This probably reflects his empirical frame of mind and his lack of concern for the archaizing norms 30

And, of course, similarly for the other mahmusa and majhura respectively, to one extent or another. It cannot be accidental that hamza is not listed with the other majhura stops as having special (probably devoiced) allophones in pause: voicelessness in the glottal catch is hardly anything special; Sibawayhi even states explicitly that hamza is no different in pause and in context (Kitab, II, 310, 22). Devoicing without loss of contrast is rather common in many Peninsular and Eastern Beduin dialects, wherein the /g/ ( < q) of /sug/ 'market' is, in pause, voiceless, lenis and unaspirated, vs. the Ik/ of, e.g., /abuk/ 'your father', which is voiceless, fortis and aspirated; similarly /sud/ 'black (pi.)' vs. /byut/ 'houses', etc. 31 The two versions in which this work has come down to us were published, with a Persian translation, by Xanlari (1954-55); a Spanish translation of one of the versions appears in Alarcon (1925), 284-288, 298-306, and a German translation of the same version in Bravmann (1934), 112-131; an English translation of the other version may be found in Semaan (1963). For further bibliographical details, cf. Blanc (1966). 32 Thus his t is clearly voiceless, his d is a stop, his q seems to be voiceless, though he indicates a [g] pronunciation for q that was common among 'the Arabs' of his day; his j is clearly an affricated [g]; for details and exact references, see Blanc (1966).

THE 'SONOROUS' VS. 'MUFFLED' DISTINCTION IN OA PHONOLOGY

305

preferred by the orthoepists. If I have expressed reservations as to the term 'Old Arabic' when applied to what Sibawayhi deals with (see fn. 7 above), it clearly must be dropped altogether in the case of Ibn Sina. Ibn Sina knows of no overall dichotomy such as mahmüsa vs. majhüra. He refers, rather vaguely, to certain vibrations that occur with the spirants z, ó, g and are absent in s, s, 6, and this is about all he has to say about voicing, assuming this is what is meant. 33 He is, however, somewhat clearer in the matter of strong vs. weak articulation: [g], [g], [d] are said to have an occlusion that is weaker ('ad'af) than that of [k], [5], [t], [t], whose occlusion is stronger ('asadd); [b] and [p] stand in the same relation to each other, if only because [p] is called bff musaddada, 'strengthened b'; and [g] is weaker than [x].34 Here, then, are lenis and fortis pairs rather clearly described as such, though not systematically and not in terms of an overall distinctive feature; moreover, the lenes are evidently voiced and the fortes voiceless, though this is not spelled out; the [p] is the Persian sound, as are [g] and [5], and whatever doubts there may have been over voicing in Arabic, there is none with regard to Persian. The [g] is also described as the sound "used by the Arabs nowadays instead of the q", a reference to the voiced reflex of q that is still typical of Beduin speech today. The [t] is voiceless, an emphatic [t], not an emphatic [d]; this is shown not only by its description with 'strong occlusion' together with [t], and opposed to the 'weaker occlusion' of [d], but also by the statement that velarized [ /] is to plain [/] what [t] is to [t].35 Ibn Slna's description is not only independent, but also isolated. For centuries after him the grammarians continued the Sibawayhi tradition and spoke of majhüra with 'full' closure and no breath vs. mahmüsa with 'weakened' closure and breath. He thus provides the only explicit indication that what could otherwise be inferred was actually the case at least at one time and place, namely that at least some voiced sounds were lenis and some voiceless sounds fortis.36 It must be kept in mind, however, that apart from the distance in time, the relationship between the Arabic described by Ibn SIna and that described by Sibawayhi is hard to estimate. While the latter clearly describes a pronunciation that was considered acceptable for good speech and recitation (together with some less acceptable variants) in 8th century Iraq, Ibn Sina is far less explicit. Both he and Sibawayhi were, moreover, native speakers of Persian, but unlike Sibawayhi, Ibn Sina spent all his life in Iranian regions where Arabic, though the language of culture, must have been spoken natively by a minority and in a number of varieties of which nothing, it should be pointed out, " Xánlari, 16-18, 39-40. " On [g] and [x], cf. Xánlari, 39-40; on [g] vs. [k] and [g] vs. [6], cf. ibid., 21, 43; on [d] vs. [t] and [{], ibid., 17, 40; on [b] vs. [p], ibid., 24, 44; cf. also Blanc (1966). " Xánlari, 23, 44. " Alarcón (1925), 305, fn. 1, was the first to note the connection, in Ibn Sina's description, between voicing and weak occlusion: "También atribuye en ciertos casos, el matiz especial, característico de las letras sonoras a una oclusión más débil". I have tried to work this out in some detail in Blanc (1966).

306

H. BLANC

is known. Still, there is nothing in the description itself that casts doubt upon its authenticity; on the contrary, it is on the whole, as has been mentioned, quite similar to present-day pronunciations, notably that of modern Classical Arabic. Nothing that Ibn Slna says can, of course, prove anything with regard to Slbawayhi's phonetics ; but, taking all the data into consideration, the proponents of the view that the majhura were lenis and the mahmiisa fortis seem to be vindicated.

IV. CONCLUSION

It is useful to distinguish between (A) the facts of Arabic pronunciation in Slbawayhi's time; (B) his analysis of the facts and his nomenclature; (C) an adequate modern analysis and nomenclature. (A) The facts, so far as they can be reconstructed, seem to be that Arabic in 8th century Iraq had two voiceless, aspirated and probably fortis stops (k, t), six unaspirated and probably lenis stops (', q, j, t, d, b) of which at least some had both voiced and voiceless allophones; voiceless spirants that were probably fortis; voiced spirants, sonorants, and semi-vowels that were probably lenis, and of which some may have had voiceless allophones in pause. There were differences between formal pronunciations and less acceptable ones, as well as dialectal variations, but there are no indications of essential differences in the system as a whole. (B) Beyond the description by point and manner of articulation, Slbawayhi arranges all the above phonemes into two series, such that one includes all sounds with nafas, '(voiceless) breath', viz. the voiceless spirants and the aspirated stops, and the other includes all sounds without nafas, viz. all phonemes with any degree of voicing and all unaspirated stops. These he calls respectively mahmusa 'muffled', and majhura 'sonorous', and the terms may possibly have been coined to designate 'voiceless' vs. 'voiced'. Their impressionistic nature made it possible, it may be mooted, to retain them even after definitive analysis showed voicing to be less relevant as a distinctive feature than presence or absence of nafas. The statement, attributed to Slbawayhi but absent from the Kitab, that all the majhura were voiced and all the mahmusa voiceless, contains an explicit reservation about whispered speech that foreshadows the more definitive statement of the Kitab. The statement assigning 'full' contact to the majhura and 'weakened' contact to the mahmusa need not refer to tenseness vs. laxness. An indication that matters were, in fact, reversed, i.e., that voiced (unaspirated) obstruants were lenis while voiceless obstruants were fortis may be found, albeit two centuries later, in lbn Slna. (C) Would a modern phonologist attempt to fit all the consonants of a system such as outlined in (A) above into a single dichotomy? My own feeling is: not necessarily. If he did, however, he might, in deference to the genius of Slbawayhi and his teachers, retain for Old Arabic a terminology that would fit both the putative facts and their traditional description; Garbell found the formula: all the mahmusa are

THE 'SONOROUS' VS. 'MUFFLED' DISTINCTION IN OA PHONOLOGY

307

'breathed' and all the majhûra are 'non-breathed'. If we step slightly beyond the traditional analysis and wish to couch our description in terms of a more generally recognized distinctive feature system, we can retain Meinhof's and Jakobson's fortis vs. lenis at least for the obstruants. JERUSALEM

REFERENCES Alarcón, M. A., "Precedentes islámicos de la fonética moderna", Homenaje ofrecido a MenendezPidat, III (Madrid, 1925), 281-308. Blanc, H., "Les deux prononciations du qâf d'après Avicenne", Arabica, XIII (1966), 129-136. , "The fronting of Semitic g and the qäl-gäl dialect split in Arabic", Proceedings of the Conference on Semitic Studies, Jerusalem (to appear). Bravmann, M., Materialien und Untersuchungen zu den Phonetischen Lehren der Araber (Göttingen, 1934). Cantineau, J., "Cours de phonétique arabe" (1941), in Etudes de Linguistique Arabe (Paris, 1960), 1-125. , "Esquisse d'une phonologie de l'arabe classique", BSL, XLIII, 1 (1946), 93-140. , "Le consonantisme du sémitique", Semítica, IV (1951), 79-94. Cohen, D., Le dialecte arabe (lassäniya de Mauritanie (Paris, 1963). Fleisch, H., "Etudes de phonétique arabe", MUSJ, XXVIII (1949-50), 225-285. , L'arabe classique — Essai d'une structure linguistique (= MUSJ, XXXIII, 1956,1-156) (Beirut, 1956). , "Maghüra, mahmüsa, examen critique", MUSJ, XXXV (1958), 193-210. , Traité de philologie arabe (Beirut, 1961). Gairdner, W. H. T., "The Arab Phoneticians on the Consonants and Vowels", The Moslem World, XXV (1935), 242-257. Galmés de Fuentes, A., Las sibilantes en la Romania (Madrid, 1962). Garbell, I., "Quelques observations sur les phonèmes de l'hébreu biblique et traditionnel", BSL, L, 1 (1954), 231-241. , "Remarks on the historical phonology of an East Mediterranean Arabic Dialect", Word, XIV (1958), 303-337. Ibn Sinä, see Xänlari. Jakobson, R., "Mufaxxama — The 'emphatic' phonemes in Arabic" (1957), in R. Jakobson, Selected Writings, I (The Hague, 1962), 510-522. , and Halle, M., "Phonology and Phonetics" (1956), in R. Jakobson, Selected Writings, I (The Hague, 1962), 464-504. and , "Tenseness and Laxness", in R. Jakobson, Selected Writings, I (The Hague, 1962) 550-555. Meinhof, C., "Was sind emphatische Laute und wie sind sie entstanden?", ZES, XI (1920-21), 81-106. Schaade, A., Sibawaihi's Lautlehre (Leiden, 1911). Semaan, K. I., Arabic Phonetics — Ibn Sinä's Risälah on the Points of Articulation of the Speech Sounds (Lahore, 1963). Sîbawaihi, Le Livre de Sibawaihi, ed. H. Derenbourg (Paris, I, 1881; II, 1889). Sprenger, A., et al. (eds.), Kassäf 'iffilähät 'al-funün, A Dictionary of the Technical Terms used in the Sciences of Musalmans [sic] (Calcutta, 1862). Steiger, A., Contribution a la fonética del hispano-árabe y de los arabismos en el ibero-románico y el siciliano (Madrid, 1932). Troupeau, G., "Le commentaire d'Al-Sïrâfi sur le chapitre 565 du Kitâb de Sïbawayhi", Arabica, V (1958), 168-182. Trubetzkoy, N. S., Grundzüge der Phonologie (Prague, 1939). Völlers, K., "The System of Arabie Sounds as based upon Sibaweih and Ibn Yaïsh", Transactions of the 9th International Congress of Orientalists (London, 1893), II, 130-154.

308

H. BLANC

Wallin, G., "Über die Laute des Arabischen und ihre Bezeichnung", ZDMG, IX (1855), 1-69; XII (1858), 599-655. Xänlari, P. N., ed. and transl., Maxärij 'al-hurüf yä 'asbäb hudüO 'al-hurüf (Teheran, 1333 A.H.S./ 1954-55).

M O R T O N W. B L O O M F I E L D

THE SYNCATEGOREMATIC IN POETRY: FROM SEMANTICS TO SYNTACTICS "Poets must seek, not as Wordsworth thought, words in common usage, but a powerful and passionate syntax." W. B. Yeats

P

R O F E S S O R Jakobson, who has made us aware of so much in language and its functioning, was the first linguist who distinguished between the metaphoric and metonymic or the symbolic and contiguous dichotomy in the linguistic process. 1 I propose to carry a little further this distinction in language, employing it in a slightly different way, in order to show its usefulness in studying literary work. My emphasis will be upon metonymy — the role of contiguity and linguistic context — which as Professor Jakobson has said has been neglected in comparison with metaphor. It has long been recognized 2 that words have two major roles in language — to refer and to indicate internal relations. In the one, words point outward to concepts, macro- or micro-cosmic; in the other they point inward to their sentence or constituent context. The first convey meaning: they name; the second convey grammatical function: they connect or relate. One function is vertical; the other horizontal. Words or morphemes in their referential role need the world of meaning to disambiguate them; words or morphemes in their syncategorematic role need only the linguistic or grammatical world. The difference between these two functions is often seen as a difference between two types of words, although I think there is really not such a sharp distinction between these functions as such a verbal division implies. Charles Morris in discussing semiotic, the theory of signs, designates these functions by the terms, seman-

1 Fundamentals of Language, Part II ( = Janua Linguarum, 1) (The Hague, 1956), 76-82. • See, e.g., Aristotle, Poetics, XIX 7ff. ( = 1456bff.) and Rhetoric III, 5, 2 ( = 1407a). See also R. H. Robins, Ancient and Medieval Grammatical Theory in Europe with Particular Reference to Modern Linguistic Doctrine (London, 1951), 43 et passim. "Besides the words which are used to name ideas, we need those which signify the connection of ideas or propositions" (Leibnitz, New Essays Concerning Human Understanding, III, 791, trans. Alfred G. Langley, 3rd edition, La Salle, Illinois, 1949, p. 364).

310

MORTON W. BLOOMFIELD 3

tic and syntactic. Other terms have been employed, usually indicating a distinction in words rather than in function: autosemantic and synsemantic, lexical and grammatical, full and empty. Bertrand Russell calls syntactic words incomplete symbols. They have also been called form or functional words. The medieval logicians used the 'syncategorematic' (as opposed to 'categorematic') for these grammatical words, a term which has the advantage to us today of being relatively uncluttered by various irrelevant associations. Syncategorematic words provide the structure of the sentence or proposition, whereas the semantic or referential words are the variables. To put it in more modern terms, syncategorematic words are the words used as such in the pre-morphophonemic rules (although there are other syncategorematic words which come in on the morphophonemic or lower levels). Logic is a study of the possible manipulations of the syncategorematic element in propositions, even if the Aristotelian symbolization is replaced, as it is in modern symbolic logic, by less palpable and more mathematical symbols. Syncategorematic words are more predictable than categorematic words. The possibility of filling the gaps in an open sentence like the following are practically endless: The in the . On the other hand any speaker of English would probably be able to fill the following gaps without much trouble: dog lives kennel. He may be wrong, but his possible fillers are very few. Syncategorematic words function, it will be noted, metonymically, that is, in relation to their verbal or morphemic environment. Categorematic words function metaphorically inasmuch as they symbolize concepts. The metonymic-metaphoric distinction can then be applied to words in their language use. 'Metaphor' must be employed somewhat loosely in this context and not strictly as Professor Jakobson employs the term in his writings. 'Metaphor' is here not being used as Aristotle defined it — the finding of similarity in dissimilarity — but rather in a looser but more etymological sense of 'carrying over' or 'transferring'. 'Metonym' is also being used more broadly, somewhat in the tense of 'contextual'. In ordinary language usage, it is extremely doubtful whether any hard or fast division between these categories of words or morphemes can be made. This is why referring to the function of words rather than to the words themselves has advantages. At the most we can say that some words are more syntactic (or semantic) than others. Syntactic words have some meaning even if on a different level, often only in metalanguage, from that of semantic words. The fact that one cannot ordinarily be substituted for another even though a perfectly well-shaped sentence will result, shows that they have a different meaning. 'The man' does not mean the same as 'a man' in English in spite of the fact that 'the' and 'a' share a formal class — that of article — together. I think everyone more or less recognizes the semantic aspect of syncategorematic words. It is, however, less commonly recognized that lexical words have some syntactic or relational role, even when undistinguished by a syn3

Signs, Language, and Behaviour (New York, 1946), 219.

THE SYNCATEGOREMATIC IN POETRY

311

tactic bound morpheme as in inflectional languages. The parts of speech into which these words may be put, for instance, provides an elementary syntactic category. The part of speech to which a word belongs determines its possible sentential roles, which are made more precise by word order or syntactic markers. Take the word home in English. This word which would ordinarily be considered a semantic word can function in certain roles and only in certain roles in a sentence. Home is relatively open syntactically and can act as a verb, noun, adjective, or adverb, depending on context. The context immediately establishes its syntactic function. He came [ ]. adverb [ ] is where the heart is. noun He hit a [ ] run. adjective The pigeon will [ ] on its release, verb Anyone having a knowledge of English rules can immediately classify the syntactic role of the word by virtue of the context. Owing to the prevalence of functional shift, the relative ease with which words can be shifted in English from one part of speech to another, most English semantic words are relatively open syntactically. However, some words are more restricted than others, with a syntactic role limited to two or even one possibility. In the latter case, the part of speech to which it belongs may ordinarily, if the word is known to the speaker, be determined at once. Pronouns and adverbs ending in -ly* are usually syntactically closed in English, for they are generally incapable of functional shift (except in metalanguage). In the case of adverbs with -ly the presence of a bound morphemic ending, which is the syntactic part of the word, forces its syntactic function to a single role. In inflected languages where functional shift is relatively limited, the syntactic morphemes, if unambiguous, perform a similar role. When they are ambiguous, experience solves the possible ambiguity. Mensa in Latin could be the ablative of a noun mensa or the second person singular imperative of of a verb mensare. However, the speaker knows of no verb mensare, whereas he knows of a noun, mensa. -s in oral English is more ambiguous, but the context determines whether it is a genitive or a plural or belongs to a noun or a verb category. Most endings in inflected languages, however, make the problem a rational not an empiric one. In any case, lexical words carry, so to speak, some kind of grammatical load which can easily or with a little difficulty be interpreted by means of sentence context. Semantic categories in one system of grammar may become syntactic in another. Until recently the animate-inanimate distinction in English was regarded as a purely semantic distinction, but it is now realized that such a distinction has grammatical implications and indeed determines the compatibility of certain nouns and verbs. It becomes a grammatical rule rather than a semantic feature in generative-trans-

4

I am aware of the fact that there are -ly adjectives in English too.

312

MORTON W. BLOOMFIELD

formational grammar. In other words, words and morphemes squint in two directions at once — towards the world of concepts and towards the world of words, although in almost every case the squint in one direction is stronger than in the other. Syncategorematic words then are words which function MAINLY in terms of context — intrasententially: "words with inherent reference to the context."5 Categorematic words are words which function MAINLY in terms of the world of meaning, extrasententially. In terms of the listener or reader, words and immediate constituents function transitively and intransitively. They are both means to other words or concepts as well as ends in themselves. In general the artistic use of words works to increase the intransitivity of words. This feature of literary language is similar to the notion of 'foregrounding', to use the happy translation of P. L. Garvin, which Jan Mukafovsky of the Prague School suggested was its main characteristic. Poetic language calls attention to itself. The 'object' of words which are mainly lexical is the world of meaning; the 'object' of words which are mainly functional is its own verbal or syntactic context. Because of the heavy metonymic rather than metaphoric function of syncategorematic words, paradoxically we become more aware of such words as words. When words lead to other words we are more conscious of their own wordhood than when they lead to concepts. When words lead to concepts, we tend to forget the word in the concept. In other words the transitivity of syncategorematic words increases their own intransitivity. The transitivity of categorematic words on the other hand decreases their own intransitivity. Inasmuch as one of the main purposes of literary art is to increase meaningfully intransitivity, the problem of the artist is much greater when working with semantic words than with syntactic words. Syntactic words carry much intransitivity by their very presence; whereas to increase the intransitivity of semantic words poses a great problem. Poetry not only says something about the world, it also says something about the language, about its own form of expression.® Poetic language not only communicates concepts; it also communicates itself to the reader and listener. Poetic devices enable the artist to do so. Poetic devices are the artist's weapons in his fight against the transitivity of lexical words. Again Professor Jakobson makes a similar point although not quite in our terms when he writes, "The function of poetry is to point out that the sign is not identical with its referent. Why do we need this reminder? Because along with the awareness of the identity of the sign and the referent (A is A:), we need the consciousness of the inadequacy of this identity (A is not A:); this antinomy is essential, since without 6

Roman Jakobson, "The Cardinal Dichotomy in Language", in Language: An Enquiry into its Meaning and Function, ed. Ruth Nanda Anshen ( = Science of Culture Series, VIII) (New York, 1957), 161. • "If there is one meaning which the metrical pattern enforces on all language submitted to its influence, it is this: Whatever else I may be talking about, I am talking also about language itself" (John Thompson, The Founding of English Metre, New York-London, 1961, 13).

THE SYNCATEGOREMÀTIC IN POETRY

313

it the connection between the sign and the object becomes automatized and the perception of reality withers away." 7 Almost all theorists on the nature of poetry are aware of the fact that poetic devices function at least partially for the purpose of getting attention and of emphasizing certain words. To list the many critics who have discussed these aspects in various ways and in various terms would be otiose and unnecessary. Sigurd Burckhardt puts the matter in his article "The Poet and Fool and Priest" 8 in a most arresting way and in a manner close to my approach. I propose that the nature and primary function of the most important poetic devices — especially rhyme, meter, and metaphor — is to release words in some measure from their bondage to meaning, their purely referential role, and to give or restore to them the corporeality which a true medium needs. ... He [the poet] can ... drive a wedge between words and their meanings, lessen as much as possible their designatory force and thereby inhibit our all too ready flight from them to the things they point to. Burckhardt is interested in the phenomenon from a psychological point of view. He emphasizes the fight of the poet with his material, and his necessity of stripping it of its ordinary force as used in everyday life. The media of the painter and sculptor and composer do not need this stripping : words do. My interest is more linguistic than psychological. The stripping does not interest me as a manifestation of the poet's will but as a grammaticization of language. 9 What I am arguing here is that language affords in its normal functioning a means of self-attention. The syntactic function of morphemes, words and phrases leads to an intransitivity which makes us aware of their own forms and leads us back to the sentence. Poetic effort, among other things, is directed to increasing the syncategorematic and decreasing the semantic element of language. The poet heightens the normal syntactic function of his words and phrases as much as he can by means of poetic devices. Now, literature has other goals than foregrounding. It presents a point of view, it organizes experience, it attempts to catch attention, not to speak of social and other indirect aims. But great literature, especially poetry, has a linguistic message as well, and in the last analysis this is what gives it much of its greatness. This is ' Victor Erlich's translation of a sentence from "Co je poesie", Volné smery, XXX (1933-34) in his Russian Formalism: History — Doctrine (= Slavistic Printings and Reprintings, IV) (The Hague, 1955), 154. 8 ELH, XXIII (1956), 279ff. The quotation is from pp. 280-281. • Murray Krieger makes a similar point, but I think explains it in the opposite (and I think, wrong) way when he writes: "It [Poetic language] is language in rebellion against the ways in which we normally use it as a counter for things; it is language that subverts its normal auxiliary function of denying its own terminal existence in order, instrumentally, to lead us to the world ; language that proclaims itself as substance and its own world of multiplying meanings as sovereign" ("The Poet and His Work — and the Role of Criticism", CE, XXV, 1964, 408). The important, though not the only feature of poetic language, is its terminal quality, its intransitivity, not its instrumentality. Cf. Roger Caillois, Babel, orgueil, confusion et ruine de la littérature (Paris, 1948), 238 ff.

314

MORTON W. BLOOMFIELD

why paraphrase even when necessary, destroys a poem or a finished work of art. It has always been recognized that a poem gives fresh meanings to words. "La poésie est ainsi re-sémantisation du discours."10 That it also re-grammatizes discourse has not I think been widely recognized. Poetry is a kind of syntactization of referential words. They become more métonymie and less metaphorical. Poetry has a syntactic as well as a semantic density. It should also be noted that certain genres like the novel are less concerned with foregrounding than others. What I have to say applies more forcefully to poetry than to prose. Yet poetic devices are by no means rare in novels or short stories, but they are not so persistent or frequent. Let us look at some of the better known poetic devices with this function in mind. It is of course clear, I hope, that the confines of an article make any kind of complete analysis impossible. Nothing less than a handbook would fulfil the task of showing how poetic devices work in enforcing foregrounding and intransitivity. However, a few words may not be inappropriate and will perhaps make my point clearer. (1) Under the broad head of repetition, we may group a number of well-known poetic devices: rime, sound-repetition, alliteration, assonance, meter and rhythm, parallelism, verbal or grammatical repetition and so forth. All of these gain their primary force from their linguistic context and first of all operate intrasententially or intersententially. The semantic power must of course combine successfully with the linguistic repetition to be accounted aesthetically satisfying. None of these poetic devices can be mechanically applied. When yoked with semantic significance, they produce great literature. Rime, sound-repetition, alliteration and assonance are different types of segmental phonetic repetition that effects an immediate sound linkage and captures a kind of primitive attention. The linked sounds provide a 'macro-context'11 in which intraand inter-sentential patterns become perceptible. They force us to dissociate partially words from their normal meanings and perceive them as words. Meter and rhythm are also phonetic typics of repetition relying on segmented time and supramental phonetic representation. They are more strongly intersentential, since in order to gain their effect they must persist over a minimum period of time.12 They link together sentences and clauses and set up a foregrounding pattern for the whole poem — a kind of contextual adhesive. Rhythm and meter also operate powerfully on the physiological apparatus and have a deep biological effect. This effect is not my concern here, but it certainly links together the reader or listener 10

Paul Zumthor, "Stylistique et poétique", in Style et Littérature (The Hague, 1962), 34. See Michael Riffaterre, "Stylistic Context", Word, XVI (1960), 213. 12 There are some very short poems — of two or three lines — in which the rhythmical or metric effect, when it is relied on to any extent, is based on close juxtaposition. Occasionally we get a strong effect from the rhythm of a phrase or sentence; these cases are rather rare and the foregrounding is often subordinated, though not of course eliminated, to the referential power and wit of the words — or sometimes to repetitive devices other than rhythm and meter. Poems generally, however, are at least four and usually more lines long. 11

315

THE SYNCATEGOREMATIC IN POETRY

and the poem. The rules of meter and rhythm must also be distinguished from the actual performance of the verse. They pertain to the deep rather than the surface grammar. Parallelism affects more directly the verbal and syntactic level of communication. It operates both intra- and inter-sententially and is a more obvious type of repetition than the phonetic types. Robert Frost's "Acquainted with the Night" could be analyzed at great length, but I shall examine it only for the purpose of looking at some different types of repetition. I have been one acquainted with the night. I have walked out in rain — and back in rain. I have outwalked the furthest city light. I have looked down the saddest city lane. I have passed by the watchman on his beat. And dropped my eyes, unwilling to explain. I have stood still and stopped the sound of feet. When far away an interrupted cry Came over houses from another street, But not to call me back or say good-bye; And further still at an unearthly height One luminary clock against the sky Proclaimed the time was neither wrong nor right. I have been one acquainted with the night.

1 2 3 4 5 6 7 8 9 10 11 12 13 14

'Light' rimes with 'night' and 'lane' with 'explain'. The second word in each of these pairs is linked to the first. The odd last word in the second line of each triplet becomes the major rime of the next triplet, for 'rain' (line 2) rimes with 'lane' and 'explain'. Besides the referential value of 'night', we have a syncategorematic value which forces us to link it with 'light' making us aware of both words as words and setting up an opposition (which is picked up again in reconciliation in the second to last line "Proclaimed the time was neither wrong nor right", reinforced by the rime 'right' which drives us back to 'light' and 'night'). Rimes as always exhibit assonance, and the [ai] diphthong sounds through the poem. There is another riming word 'height' in line 11 and [ai] occurs in 'eyes' (line 6), 'cry' (line 8), 'good-bye' (line 10), 'time' (line 13). Other sounds like [ei] also reverberate through the poem. Throughout the poem the five beat rhythm both in deep structure and performance runs steadily except for line 11 where only 4 beats can appear in performance — "And further still at an unearthly height." The stress there on 'still' emphasizes its ambiguity — a semantic feature which causes us to waver syntactically. 13 Except for the final couplet, the poem is written in terza rima with the unrimed 18

A type of syntactic ambiguity. See below, p. 317.

316

MORTON W. BLOOMFIELD

final word of the middle line of each triplet picked up as the rime of the next. The odd rime of the fourth triplet becomes the rime of the final couplet which is also the rime of the first couplet. Alliteration also effects linkage within lines as in "I have stood still and ¿topped the sound of feet" (line 7). Less noticeable we have two '1's' in line 4, which also echoes the / of 'light' in the preceding line. Alliteration links 'cry', 'came', 'call' running over lines 8, 9, and 10. Line 10 offers us ftut, ¿>ack, and bye. These are the main alliterative linkages. This sonnet also offers a number of significant parallelisms on the verbal and grammatical level. Words like 'I', 'night', 'rain', 'still', 'acquainted' occur more than once, and we also have a fascinating inversion 'walked out' (line 2) and 'outwalked' (line 3), all emphasizing the profound despair of the repetition of nature and life — the lack of interest which a perpetual cycle creates. This is all heightened by the repetition of the first line in the last line, giving the effect of a circle which never gets anywhere. The first seven lines are each complete sentences. Note the repetition of the perfect tense "I have been, walked, outwalked, looked down, passed by, dropped (with a variation by which 'have' is only understood), stood, stopped ('have' understood), and finally 'been' again." The perfect with its sense of completed action comes again and again. There are only two other finite verbs, both in the preterite, emphasizing the pastness of the past as seen in the coming of the 'cry' (line 9) and the proclaiming of the 'clock' (line 12). (2) Under the broad head of deviation, we find ungrammaticality of various types — violations of the syntactic or lexical rules of English.14 Lexical deviation belongs properly to semantic behavior and is a normal process of the use of words. Grammatical deviation of various sorts shocks us into intransitivity and is hence of our immediate concern. That poets deliberately violate the rules of grammar is a commonplace. These violations force us into an awareness of syntactic context and of the word or words themselves and are part of the general metonymic process. This aspect has not I think been noted before. When Dylan Thomas writes 'a grief ago', 'all the sun long' or 'robed in the long friends', 'His golden yesterday asleep upon the iris' or E. E. Cummings, "Blow soon to never and to twice / (blow life to isn't: blow death to was)", the syntactic deviations are arresting. More common in the poetic tradition is variation from normal word order and original conversions of words into new parts of speech. All syntactic deviations are attention-getting and serve the function of making us more aware of the metonymy of words and phrases. They all help to dissociate words from their meanings and make us aware of words as words. Deviations of style, tone, metrical pattern, even genre are also common. Sudden shifts in person, point of view, direct and indirect narration: all these are types of 14 See Samuel R. Levin, "Deviation — Statistical and Determinate — in Poetic Language", Lingua, XII (1963), 276-290.

THE SYNCATEGOREMATIC IN POETRY

317

15

deviation which effect foregrounding. One can at present do little more than point them out. (3) The third heading is ambiguity, especially syntactic ambiguity. It is not a common occurrence, but it occurs frequently enough to be thought of as a poetic device. Here a close reading of the context is necessary to resolve the ambiguity if indeed it is resolvable. Even if it is resolvable, it also always suggests its other possibility. We have already referred to one of these ambiguities above in the word 'still' in Frost's poem where the ambiguity of adjective and conjunction/adverb remains unresolved, in suspension, until 'proclaimed' in line 13 favors the second alternative. With a similar suspended ambiguity Archibald Macleish writes "To feel the always coming on / The always rising of the night" where 'always' suggests a noun until 'of converts it into an adjective. We find another type of syntactic ambiguity in English when an inverted or possible inverted word order creates suspension as to what immediate constituent a word belongs. In "And brass eternal slave to mortal rage" (Shakespeare, Sonnet 64), 'eternal' hovers between 'brass' and 'slave'. Or in "and beauty making beautiful old rhyme" (Sonnet 106) where 'beautiful' may be an attributive adjective (with 'old rhyme') or a predicative adjective (after 'making'). All these syntactic ambiguities serve to push us to scrutinize the context, hesitate over it and become aware of the syntactic possibilities. They partially separate their lines or sentences from their meaning and create an intransitivity which bars the easy leap to referential meaning. The richness of poetry is syntactic and metonymic as well as semantic and metaphoric; and one ignores the former at the cost of the understanding which can open new vistas upon language and life. Poetry and verbal art must rest upon the manipulation of the powers of language. We not only perceive new meanings in the extralinguistic world in great poetry, but we see new possibilities in the world of language itself. HARVARD UNIVERSITY

18

Even the shift from count to mass noun effects such a shock. See for example Shakespeare's "the twilight of such day / as after sunset fadeth in the west" (Sonnet 73) where 'day' becomes a mass noun for the nonce.

IL T. EOrATMPEB

HMnPOBH3AI];HJI XY^O^CECTBEHHtlX MATEPHAJIE HA

IIOBECTEÎÎ

JIYBOHHBIX H

IIECEH

H

HOPMbI

IÏPHEMOB XVIII

B.,

KAPTHHKAX, O

EPEME

H

HA

HA^IIHCEfï CKA30K,

e B. IL AflpnaHOBOH-IIepeTii h fl. C. JlHxaneBa (cTp. 12) HHTepecHbi 3aMenaHHH o crane h pH({)Me hobccth o OMe h EpeMe. Pha HHTepecHbix h ueHHbix 3aMenaHHH o xyfloacecTBeHHoii 4>opMe npoH3BeaeHHH

pyKonHCHoii h ycTHoii TpaflHUHH o EpeMe h oMe Haxo/iHM b CTaTbe P. ^Ko6coHa 'TpaMMaTHKa II033HH H n033HH rpaMMaTHKH".3 1 a 8

ffpeennn u Hoean POCCUH, I, JVs IV (1876), 359-368. A. TaJiaxoB, Hcmopun pyccKoù cjioeecHocmu, I, H3fl. 3 (MocKBa, 1880), 503. Poetics — Poetyka — IIosmuKa (The Hague, Mouton, 1961), 397-417.

HMIIPOBH3AUHfl H HOPMbI XyflOHCECTBEHHblX IIPHEMOB B pyccKOM

esthétique

nepeBoae

KHOTH A b r a h a m M o l e s ,

319

Théorie de l'information et perception

( p e f l a K T o p B . B . H ß A H O B ) B I r j i a B e § 5 "HHOMy-TO n o Ü M a j i H , E p e M y - T o c r p e G j i n — B n e c H e 5 — 2 p a 3 a . E p e M a - T O y r n e j i , OMa y6e»caji, E p e M a BbipBajica — B necHe 5 — 2 p a 3 a .

Heocojibicy

322

3th

n . r . BorATMPEB pflflOM

CToamne

napajuiejibHbie

cHHOHHMHiecKHe

pa3M

hocat

(JjyHKUHio

JieHTMOTHBOB. 6 K p o M e T o r o aHajiH3 n o B e c T e f l , jiyÖKOB h a6yjibi h hbjmiot-

ch loci communes.

Otmcthm, hto HeK0T0pbie napHbie npeflJioaceHHa BCTpenaioTCH b pa3JiHHHbix npoH3BefleHHax 6e3 H3MeHeHH«, KaK to : " E p e M a c e j i b JioflKy, O o M a b 6othhk" — b n o ß e c T a x I , I I , I I I , I V . BcTpenaioTCH napHbie npeAJioaceHH» — loci communes, b kotopwx cjiobo öomttuK 3aMeHHeTcaflpyrHM6jih3khm eMy no 3HaHeHHio cjiobom: 6othhhok, 6othh, HejiHoneK, B e c J i o n e K :

b

j i o j i K y , a C>oMa-To

O x , E p e M a c e j i fla

JioflKy, a O o M a

"EpeMa-To ceji

b

b 6othhhok". ( 1 ) b 6othk>. (ck. 1)

E p e M a K y n n j i JioflKy, O o M a HejiHoneK. ( 6 ) E p e M a K y n n j i jioflKy, OMa B e c j i o n e K . ( c k . 3 ) pa3bi " B o t E p e M a 3 a m e j i b uepKOBb, a OOMA b a j i T a p b " ( c k . 1) KaK loci BCTpeiaioTCH

b

HecKOJibKHX n o ß e c T H x , CKa3Kax

h necHax.

communes

stom H a ö j n o a a e T c a b

Ilpn

oflHHX c j i y n a a x He3HaHHTejibHaa BapHaTHBHocTb, b a p y r a x —

6ojn>iiiaa:

E p e M a B i u e j i b uepKOBb, O o M a b o j i T a p b . ( I )

b

EpeMa

iiepKOBb,

oMa Ha n a n e p T b " ( I I I ) ;

b

O o M a b npHTBop" (II). K p o M e T o r o mm B c r p e n a e M TaKHe BapnaHTbi, r ^ e cjiobo ifepKoeb 3 a M e H a e T c a cjiobom KAupoc: " E p e M a c T a j i Ha KpbiJioc, a O o M a - T o Ha a j i T a p b " ( 1 ) ; " C r a j i O o M a Ha K J i n p o c e , a E p e M a b a j i T a p e " ( 4 ) ; " E p e M a - T o CTaji Ha kphjioc, O o M a - T o b a j i T a p b " (ck. 3).

Be3 H3MeHeHH« ß c T p e n a i o T c a napHbie n p e a n o a c e H H a —

loci communes:

"Bot

E p e M a K y n n j i j i o i n a f l b , a < 5 o M a - T o x c e p e 6 u a " b HecKOJibKHX 4>ojibKjiopHbix n p o H 3 B e a e H H a x : 2 , 6 , 1 8 , CK. 3 .

B 3thx xce ÖJIH3KHMH n o

napHbix

npefljioaceHHax

oTuejibHbie c j i o B a

3aMeHaioTca

apyrHMH,

3HaieHHIO:

• Herbert Peukert, "Die Funktion der Formel im Volkslied", Poetics — Poetyka — IIo3muKa, 525-536.

Mho» cAaHa b nenaTb cTan.« o ^ymcunax

jiefiTMOTHBOB

b

pyccKoä

6ujihhc.

HMHP0BH3AUHH H HOPMbI Xy,HO>KECTBEHHbIX ÜPHEMOB

323

EpeMa KynHji MepnHa, a OoMa-To »cepe6na (3, CK. 1) EpeMa Kynaji Jioinaflb, oMa-To cojioBKa. (7) H o , 3aMeHHH oflHo CJIOBO apyrHM, nepenncHHKH pyKonHceñ h HcnojiHHTejiH neceH h cKa30K He HapyinaioT npH STOM, KaK MM BH^HM, HH npneMa cHHTaKCHHecKoro napajuiejiH3Ma, HH CHHOHHMHH STHX pa3. napHbie coneTaHHH CHHOHHMHHCCKHX yHKUHio: noKa3aTi> CXOACTBO 06pa30B flByx HeyaaHHHKOB H CXOJICTBO HX HeyziaHJiHBbix noxoxcfleHHH. HeKOTopbie napHbie cHHOHHMHnecKHe npefljioaceHHH, nocTpoeHHbie no npHHUHny CHHTaKCHHecKoro napajuiejiH3Ma, eme Kpenne cnaHBaioTca Mexcay COÓOH PH(J)MOH H accoHaHcaMH: EpeMy TonKaMH, a í>oMy nHHKaMH. (jiy6. 1; 5) EpeMa B Becjia rpeGeT, a í>OMa OÜHH paicn 6epeT. (Jiy6. 1) EpeMa B nyxcoM, a í>oMa He B CBOCM. (III, IV) EpeMa IHH6KOM, a OoMa 6POCKOM. (II) EpeMa KPHHHT, a OoMa 6ojibuie 3biiHT. (I) OTflejibHbie napHbie pHcfiMOBaHHbie cHHOHHMHiecKHe (J)pa3bi, noBTopaiomHecH B oflHOM H TOM »ce npoH3Bex(eHHH no HecKOJibKy pa3, Hrpaior pojib JICHTMOTHBOB. TaK, HanpHMep, (J>pa3bi: EpeMa TO GOKHT He orjiaHeTCH, a OoMa-To 3a HHM ayeT He ocTaHeTca. BOT

noBTopaioTca B oflHoíí necHe ( 2 ) flBa pa3a. B STOH »ce necHe c 3aMeHoñ cjioBa dyem CJIOBOM mHHem Te »ce napHbie npefljioaceHHa noBTopaiOTca eme ppa. pa3a, T.e., B 0flH0H H TOH »ce necHe OHH BCTpenaioTca 4 pa3a. C 3aMeHOH cjioa dyem CJIOBOM nynum a r a napHbie npe&Jio»ceHHH n0BT0pai0TCJi B necHe 3 flBa pa3a: BOT EpeMa TO SexcHT, He orjiHHeTca, A OoMa-To 3a HHM jiynHT He ocTaHeTca (3). B noBecTH I 6e3 HSMeHeHHa noBTopjuoTca 2 pa3a napHbie npe/iJioaceHHa "EpeMa 6HJI njieuiHB, a OoMa mejiyziHB". flBa pa3a BCTpeiaioTca Taicace B O^HOM h TOM »ce npoH3BeaeHHH npejuioaceHHH: BOT HpeMa CTaJi Hbipjm», a XaMa 3a HHM roHHTb (15). OTflejibHbie pH(J)MOBaHHHe napHbie npeflJio»ceHHJi noBTopaioTCH 6e3 HSMeHeHHa KaK loci communes B pa3JiHiHbix npoH3BefleHHHx: ' CM. rjiaBy "CHHOHHMHHeciaie coieTamw H HX pojn» B ycTHOfl ncminj" B KHHre A. n . EBreHbeBofl, OtepKU no H3biKy pyccKoü ycmnoü no33uu a lanucux XVII-XX ee. (MocKBa-JIeHHHrpafl, 1963), 254-296.

324

n . r . BorATbiPEB

MBHO (2, 3, 9, 11, 18). ô p o c K O M , a O o M a UIH6KOM (14, III, II). A e 6HJIH, a 3>OMy He n y c r a j i n (IV, V). B n y x c o M , a O o M a He B CBOÊM (III, IV).

E p e M a - T o Ha EpeMa EpeMy EpeMa

AHO,

a O 0 M a - T 0 TaM

E p e M a HrpaeT, a O o M a HaneßaeT (I, I I ) . H a E p e M e KOJinax, a Ha 3>OMe acejiBaic ( I V , I I ) . Cjieayiomne

n a p H b i e pH(j>MOBaHHbie n p e f l j i o x c e H H » B C T p e n a i o T c a B

pa3HopoflHtix

o ô p a ô o T K a x TOH »ce (J)a6yjiti c H e x o T o p t i M H H3MeHeHH«MH: O o M a He y c T o a j i , a E p e M a y ö e a c a n ( 4 ) . E p e M a y G e a c a j i , a < t o M a He oTCTaji ( V ) . E p e M y fle n o r j i a 3 a M , a OMy n o o i a M ( I V ) .

(III).

E p e M y ÔHTb CTajiH n o y u i a M , a oMy n o r j i a 3 a M

C r a n E p e M y 6 m b n o r j i a 3 a M , a oMy n o y u i a M ( V ) .

HaraM,

EpeMa no

a 3>0Ma n o njienaM

(V).

B p a 3 H o p o f l H b i x o 6 p a 6 o T K a x c i o a c e T a BCTpenaioTCH pncJjMOBaHHMe cHHOHHMHHecKHe cHHTarMM,

nocTpoeHHbie

c

HeKOTopbiM oTKJioHHeM o T cHHTaKCHHecKoro

napaji-

jiejiH3Ma:

BOT

E p e M a crraji

TOHYTB, O o M y

3a Horn

THHYTB (2, 17, 6, CK. 3).

B O T E p e M a yTOHyji, a < 5 o M y TO noTHHyji ( 7 ) . E p e M a TO yTOHyji, a O o M y - T O H o p T yTHHyji (CK. 4 ) . O o M a n o T O H y j i , E p e M y n e p T noTHHyji ( 5 ) .

H O o M y NOTAHYJI (11). NOTOHYJI, a X a M a HBO n o T H H y j i (14).

BOT

E p e M a yTOHyji,

BOT

HpeMa

KaK E p e M a noTOHyji, a X o M y 3 a K o c y noTHHyji ( 1 0 ) .

BOT

OoMyixiKa

NOTOHYJI h

E p e M y noTHHyji

(4).

I l p H B e f l e M HeKOTopbie n a p H b i e pHoMá B o j r r á p b ( I V ) . B O T E p é M a 3 a r a e n B u é p K O B b , a O o M á B a j i T á p b (CK. 1). E p é M a n p n n i e j i B uépKOBb, a O o M á B o j r r á p b (13).

EpéMa "

B u é p K O B b , 4>OMÁ B n p H T B Ó p

(II).

H . B . T o r o j i b B " 3 a M e i a H H H x nnn r o c n o a a K T e p o ß " , o T M e i a a 6 o J i b m o e CXOACTBO E o ô n i H C K o r o

H ¿(OÖIHHCKOTO, rOBOpHT H 0 6 HX OTJIHHHH : "^OÔHHHCKHS HCMH05KK0 B b i m e H C y p b e 3 H e e BOÔHHHC K o r o , HO EOÔIHHCKHË pa3BH3Hee H acHBee

floôniHCKoro"

("PeBH3op").

n . r . BorATBiPEB EpéMa t o CTaJi Ha Kpbinoc, OMá- Ha apyrôË (I, IV). Bot EpéMa Kyn&ri nómaflb, a OMá to ¡fcepeóuá (2, 6, ck. 3). EpéMa KynHJi Jiómaoi», OMá-TO coJiOBKá (7). EpéMa K y n n j i MépHHa (nponapoKCHTOH), OoMá xcepeóiiá (ck. 1), (3), (10). EpéMHHa (nponapoKCHTOH) He éaeT, 4>oMHHá He seséT (12, 2). EpéMHH t o He Traer, a omhh t o He Be3éT (3). y EpéMM to He é«eT, y omb'i to He Be3éT (6, 9, 18, ck. 3, 1). y EpéMbi He THHeT, y ®omm He Be3éT (7, 6, 14). Bot HpéMHHa hh TWBTB, a Xomhhí He BH3én. (15). EpéMy b méw, OMy b tojtjkh (I, II, III). Bot EpéMy b3hjih b méio, a OMy b tojiíkh (ck. 1). Bot EpéMy t o no mée, a OMy b tojihkh (2). TaM EpéMy t o no mée, a OMá b cochák (IV, 2, ck. 1). EpéMa b 6epé3HHK, a «PoMá b cochák (V). EpéMa c móth, a 4>OMá e TeHHTbi (IV). y EpeMbi 6bum rycjm, y í>OMbi opráH (I). y EpeMbi t o rycjiH, y 4>oMbi flOMpá (II, III). EpéMy AyÖHHOK) (ayÓHHoñ?), oMy naTorÓM (II). EpéMy no MÓ3ry, OMy no pyabi (II). EpéMa b3äji flyóímy, a ®OMá BH30By (I). EpéMa B3HJI nájiKy, OoMá naToxcÓK (II). EpéMy AyÓHHoñ, ®om^ b5I3obóíí (I). EpéMy fly6ÄHoö, «Dom^ pbiiarÓM (I). B3HJIH c coôoô EpéMa námcy, a í>OMá mn6ÓK (TV). Ox, EpéMa cxBaran nájiKy, a ®OMá Kocápb (ck. 1). En $0Má lecHÓK, a EpéMa péabKy (4). EpéMa en péflbKy, oMá lecHÓK (6). EpéMa KynHJi iryKy, OoMá necHOK^ (4, 10). EpéMa ne 3a pény, OoMá 3a HecHÓK (IV). PéflbKa 3anáxjia, necHÓK 3aBOHSJi (V). EpéMa 3aKynáeT, a ®OMá npoaaéT (IV). EpéMa 3a CBá®>6y, a OoMá croBopàri (IV). EpéMa 3a fléBKy, a í>OMá 3a noná (IV). EpéMa b poro3ÉHe, a ®OMá b KynbKé (IV). y EpéMbi nycTo, y Oomú Hiraeró (TV). EpéMa na bflBépn,a í>0Má b okhó (IV). EpéMa t o 3a KÓpo6, a OMa t o 3 a c y a n y « (3). EpeMa KynHJi c y M K y ; OMa nonropá (12). EpéMa cxBaTHn my6y, a OMá Ka (10). Bot EpéMa cxBaran pÉ3bi, a 4>OMá craxápb (12). Bot EpéMa t o 3a pÓ3y, a OMá 3a coMá (IV). EpéMa B 6anaxÓHe, OMÁ B rnaôypé (14). y EpéMbi HE pyÔHT, y OMM HE 6epéT (14). EpéMa TO nocéan, a OMá Ha 6yróp (IV). BOT EpéMa B3HJI 3a BÓXOKH, a í>OMá-TO 3a ymy (9). EpéMa TO B aepéBHio, a XoMá TO BO cenó (10). Y EpéMbi aepéBHH, y 4>OM¿I cejibuó (I, II). EpéMa cen Ha JiáBKy, a í>OMÁ B cxaMbiò (III). H a EpéMe mànica, He OM6 K O J r a á K (I). KaK EpéMa-TO B mámee, a XOM4-TO B nuibpncy (10). EpéMa B ABÉPH, OMÁ B OKHÓ (I, II, III). EpéMy CHCKájiH, OM^ Hanum (I, III, IV, V). EpéMa HacHJiy Bb'ipBancH (nponapoKCHTOH), a OoMá y6e*cán (Jiy6. 1). HeKOTopbie h 3 Bbiiiie npHBeaeHHbix HäMH n a p H b i x n p e a j i o H c e H i m BCTpenaioTCH n o HECKOJIBKY p a 3 B O^HOM H TOM ace npoH3BeaeHHH H H e c y r (JJYHKUHIO JieñTMOTHBa: OH E p e M y TO ayÔHHOH, a OOMY TO BH3OBOÍÍ (1 — ,nBa p a 3 a ) . E p e M y B m e i o , i MOH . . . M b i paccMOTpejiH 30 npoH3BeaeHHH o EpeMe H 3>0ME: noBecra, Ha^nncH Ha Jiy6Kax, a TaKace (JioJibKJiopHbie npoH3BeaeHHa: CKa3KH H necHH. H 3 HHX B flByx npoH3BeaeHHHX M H He HaxoflHM HH napHbix npeaJioaceHHH " A " , HH napHbix npefljioaceHHH " B " . Pa36epeM 0TFLEJN>H0 Kaxcaoe H3 28 NPOH3BENEHHH o $ o M e H EpeMe, B KOTOPHX BCTpenaiOTCH napHbie rrpejuioKeHHH " A " H " B " . B

13 npoH3BeaeHH«x HMCIOTCJI TOJibKo napHbie npewioaceHHH " A " :

Jiy6. 1, B

necHHx: 1, 2, 6, 7, 9, 13, 14, 15, 18; B cKa3Kax: 1, 3, 4. K p o M e Toro B 9 npoH3BefleHH«x Hapnay c npefljioaceHHHMH " A "

cnopaflHnecKH

BCTpenaiOTCH npeflJioxceHHH " B " : B noBecTH I V npeajioaceHHH THna A — 26 pa3,

rana

B n o B e c r a I I I npefljioHceHHH THna A — 14 pa3,

THna B — 1 pa3

B noBecTH

I I npeAJioaceHHfl THna A — 13 pa3,

THna B — 1 pa3

B noBecTH

V npeftnoxceHHH rana A — 10 pa3,

rana

B noBecra

B — 1 pa3

B — 1 pa3

I npefljioHeHHH THna A —

8 pa3,

THna B — 1 pa3

B necHe

3 npe^JioaceHHH rana A —

9 pa3,

rana

B necHe

4 npefljioaceHHH rana A —

4 pa3a, THna B — 2 pa3a

B — 1 pa3

B necHe

10 npefljioxceHHH THna A — 13 pa3,

THna B — 1 pa3

B necHe

12 npeaJioaceHHH rana A — 12 pa3,

rana

B — 2 pa3a

CPE^H paccMOTpeHHbix HÍMH npoH3BeaeHHH B Tpex necHHX (8, 11, 16) H B OÍIHOH jiy60HH0H KapTHHKe (Jiy6. 3) KOHua X I X B. BCTpeiaeTCJi HecKOJibKo napHbix npeajioaceHHH " B " H HeT napHbix npeflJioxceHHH " A " . B OFLHOÑ necHe (5) — 6 napHbix npeflJioxceHHH " B " H OAHH pa3 — napHbie n p e a jioaceHHH " A " . B necHe 17 —

3 napHbix npeflJioaceHHa " B " H OAHH pa3 napHbie

npeajioaceHHH " A " . Pa3o6paHHbie HaMH 28 npoH3BejieHHH noKa3biBaioT, HTO HHCJIO napHbix n p e a jioxceHHH " A " 3HAHHTEJIBHO npeBajinpyeT Haa HHCJIOM napHbix npefljioaceHHH " B " , T.e.,

npeanoxceHHH,

B KOTOPHX

CHCTeMaraHecKH

c

HMeHeM

oMa B KyjibKe EpeMa B jianTax, a O o M a B HHpKax, EpeMa B lyacoM, a O o M a He B CBOCM (IV).

T o ace BCTpenaeM B necHax : O o M a aa EpeMa 6MJIH 6paTeHHHKH, npOKypaTHHHHHKH. OHH nHJiH-ejiH cjiaflKO, aa HOCHJIH xoporno : EpeMa HOCHJI poroacy, a OoMa-To Topnbe (1). EpeMa co OMOH 6buiH 6paTejibHHHKH ; OHH JiaflHo acHBajra, xoporno xaxcHBajra; EpeMa-To B poroxace, O o M a B Topnnme (3). BOT EpeMa H oMa poflHbie öpaTb«. OHH BMecTe HCHJIH, OHH cjia^Ko ejra: eji O o M a necHOK, a EpeMa peflbKy (4). JKHJIH-ÖMJIHflBa6paTeHHHKa, Asa CKJiaflOHHHKa, EpeMaflaoMa. OHH HCHJIH GojiHe CJiaBHo HHJIH, ejIH CJiaflKO, XOflHJIH xoporno. EpeMa eji peflbKy, oMa necHOK. EpeMa B 6ajiaxoHe, OMa B ma6ype (14). 3TOT ace npneM M H BCTpeqaeM B cGopHHKe CKa30K A. M . CMHpHOBa: "ÎKHJIH 6WJIH

HMITP0BH3AIIHSI H HOPMbI XyflOJKECTBEHHblX IIPHEMOB flBa

6 p a T a , OMa a a

EpeMa.

B3¿iyMajiH p e ô î r r a ,

331

npncoBCTOBajiH n a i i i e H K y n a x a n . ...

y p O f l H J i a C b p 0 3 K b X O p O I I i a , H f l p e H a , KOJIOCHCTa, BOJIOKHHCTa. KOJIOC OT K o j i o c y

—

H e c j i b i x a T b r o j i o c y ; C H o n OT C H o n a —

—

n a j i K a M H K H f l a j i H ; c y c J i O H OT c y c j i o H a

n e p e r o H a M H TOHÎUIH" (CK. 4 ) . 3Ta

UHTaTa nepeKjimcaeTca c BbiuieuHTHpoBaHHbiM

(J>parMeHTOM H3

HapoAHoñ

FLPAMBI " M H H M M H 6 a p H H " . PaccMaTpHBaa

npHBeaeHHbie

(J)pa3bi — O K C i o M o p o H b i , Mbi BHAHM, HTO n e p e BapbHpOBEJIH HX. n p H M e p a x oKCioMopoHHbix $ p a 3 n o c j i e f l y r o m n e

IIHCHHKH H H C n O J I H H T e j l H , H M I i p O B H 3 H p y H , Bo

npHBeaeHHbix

Bcex

HaMH

4 > p a 3 b i o T p H u a i O T n p e ^ b i f l y m a e : e c j i H B n e p B O H OME NPHBOFLHT Hac K CJIE^YROMEMY 3AKJNOHEHHIO.

06pa3bi H

E p e M b i H oMbi p a c i c p b i B a i o T C H H e TOJibKO o n a c a H H e M HX BHCUIHOCTH

HX NPHKJNOHEHHH,

HO

napajuiejiH3MOM pa3 c CHHOHHMaMH

TAK»CE XYFLOXCECTBEHHBIMH

OAHOH C T o p o H b i — o

O K a » AO M

H3

3THX

EpeMe,

"repoeß";

c

NPHEMAMH : flpyrofl

pH(J)MOH,

CHHTAICCHHECKHM

— o O o M e ; (J)pa3aMH-

KOTOpaa

CJiyaCHT

n a p a j u i e j i b H b i x CHHOHHMHHCCKHX OMM H, HAKOHEU, c B o e o ô p a 3 H b i M p a c n p e a e j i e H H e M y a a p e H H Ö .

oTMenaeT, HTO pyKonHCHoii TpaflHimH, H

B . I I . AflpHaHOBa-IlepeTU n p a B H J i b H o

OoMe

CBH3bK> 06pa30B

oneHb

HeycToiÎHHB

H B

noBecTH o EpeMe 13 y c T H b i x BapHaHTax.

TCKCT B

H

RIPH a H a i i H 3 e np0H3Be,ueHiiH o E p e M e H d>oMe Mbi cTpeMHJiHCb H a MHoroHHCJieHH b i x n p H M e p a x n o i c a 3 a T b , HTO H n e p e n n c H H K H n o B e c T e ñ , c o c T a B H T e j r a H a a n H c e f t H a JiyÖOHHblX

KapTHHKaX H HCnOJIHHTCJIH (J)OJIbKJIOpHHX npOH3BÊÎ(eHHH,

BapbHpyioT

nncbMeHHbie

H y c T H b i e TCKCTH, H e H a p y m a a

HopM

HMIipOBH3HpyH,

xyaoacecTBeHHbix

npHeMOB.14 " C e p 6 c K H e rycjmpbi", KaK r m i i i e T P . .HKOÔCOH, " H HX c p e a a 3 a M e n a i o T H Hepe^KO ocyac^aiOT BCHKHH yxjioH OT CHJiJia6HHecKOH cxeMbi ANMECKHX neceH H OT NOCTOHHHoro MecTa, T.H. ue3ypw, omnwdb ne 6yòymi e cocmomuu onpeòeAumt, e teM ace cocmoum OWUÓKCT (noflH. n.E.). 1 5 ECJIH c e p ß c K H e r y c j i a p b i o c y x c a a j r a BC»KHH y x j i o H OT C H J i J i a Ô H H e c K o i i c x e M M , aHajiH3 MHoroHHCJieHHbix

npHMepoB

H3 O M e n O K a 3 b I B a e T , HTO H C n O J I H H T e j I H (JlOJIbKJIOpHblX n p O H 3 B e a e H H H ,

o

EpeMe

TO H

HMnpOBH3HpyH

" B. n . AjpaaHOBa-nepeni, Pyccmn deMOKpamuuecKOJt camupa XVII e. (MocKBa-JIeHHHrpan, AH CCCP, 1954), 241. 11 06 HMHpOBHSaOHH B HapOOHOM TBOpieCTBe CM. n . r . EoraTbipCB, "TpaflHIUW H HMnpOBH3aiIM B HapoOTOM TBopHecTBe", VII MeoKÓympoÓHbiü Kompecc aumponoAozmecKux u arrmoepaffiuiecKux myx (Mocicea, aseycm 1964 a.) (MocKBa, Hayxa, 1964). A. M. AcraxoBa, "HMnpoBH3auM B pyccKOM (J)0JU>KJi0pe (ee (JwpMti H rparaubi B pa3Hbix xcaHpax). CneuHHKa 3).4 It is as important for my own position as for K-P's to take the evidence as still valid and to regard do as extraneous to the verb phrase, so I cannot allege it against their use of will. But it was necessary to look at the consequences if the other alternative were chosen. (4) The futurity of the imperative raises two questions: whether the imperative is necessarily future, and, if it is, whether will has any bearing on the matter. The first question is a pseudo-question within the framework of the K-P imperatives as I am assuming it. If a command is an order that is to be carried out, it is necessarily understood as referring to the future — futurity is part of the definition of 'command'. From this standpoint, K-P's emphasis on will becomes not a desire to prove that commands refer to the future but simply an interest in detecting the underlying linguistic correlate of futurity. For the traditional grammarians the futurity of imperatives had a larger relevance, and I shall return to it later when we are able to step out of the straitjacket of strict commands. But the second question is pertinent. The motive for raising it, from the generative standpoint, seems to be that since the expression of tense is obligatory in the kernel, something needs to be done to involve the imperative in a tense, and will is the answer. But K-P are careful to show that a sentence like You will go home is ambiguous, and to account for the imperative interpretation they posit an Imperative morpheme. Thus presumably will in the imperative carries the basic notion of futurity and Imperative is added to it. For more than one reason, this is not very satisfying. In the first place, if the you will in the imperative interpretation of You will go home already is accompanied by the Imperative morpheme, then so is both the you will and the he will in the imperative interpretation of You will sit here and he will sit there, and we have the makings of a third-person imperative. In the second place, will often is not future. Accept 4 One argument against do as an auxiliary is that it does not quite parallel the auxiliary do as a. carrier of affirmation. For the question Did he come or not? a normal emphatic affirmative answer is He did come. For the question Shall I come or not? the answer Do come might be heard but would not be felt as a carrier of emphatic affirmation (for this, Yes, come is better) the way Don't come would be felt as a carrier of emphatic negation.

THE IMPERATIVE IN ENGLISH

339

the job, if you will, and I asked him to accept the job, if he would, clearly relate to Accept the job, will you?-, and yet in I asked him to accept the job, if he would, but he wouldn't, then or later, the wouldn't is not necessarily future to the past time axis; it means 'He refused'. The confusion on this point is occasioned not only by the ubiquitous tense associations of will but also by reading too much importance into action verbs. In I keep telling him she's stepping out on him but he won't do anything about it one has the illusion of futurity because in the nature of action verbs there has to be a time span between the telling and the acting. But in I keep telling him she's stepping out on him but he absolutely won't believe it the meaning is that he absolutely refuses, now, at the time of the telling. The reference is to willingness, and whatever futurity one infers is just as readily to be inferred if will you is replaced by do you agree, do you so incline, etc. Will you is an appeal to willingness. Its function is to cajole, like that of please, as we have already noted. Help me, will you? means Help me, are you willing to? If this is true, then the imperative will is not a future will plus an Imperative, but is a homonym of the future will. It follows that we do not interpret the imperative to refer to the future because of the implied presence of will, since even if that implied presence is conceded, it is the wrong will. WHEN we interpret the imperative to refer to the future we do so because that is in the nature of the stimulus-response relationship. 5 The same relationship and the same interpretation, with no assist from any will, is present in the following examples: "Where are you going?" "I'm going on a picnic, and you're going along." "What are you having there?" "We're having some ice cream, and you're having some with us."

II. O T H E R I M P E R A T I V E S , R E A L O R A P P A R E N T

Going along with what I suppose K - P had in mind, up to this point I have narrowed the field to imperatives that were obviously commands. It is time now to consider whether there was a justification for this. Overlaying an Imperative on a basic future would be another instance of defeating the hopes of those generativists who look for confirmations in historical development. The imperative connotations of will preceded the future connotations, as can be seen in certain residues today. There are constructions referring to the future that fail to get by if 'willingness' is excluded. Compare 6

If If If If

he dies tomorrow, and I hope he does... (*he will) he gets caught, and I hope he does... (*he will) the door blows shut, and I hope it does... (*it will) the road they're going to build goes uphill, and I hope it does... (*it will)

with If he comes, and I hope he will (or does)... If she seems acceptable, and I hope she will (or does)... If he climbs uphill, and I hope he will (or does)...

340

DWIGHT BOLINGER 1. Conditions

The commonest non-command use of the imperative is in if-then sequences like the following: Spare the rod and spoil the child', Cut ourselves off from that source oj income and we're ruined; Break that vase and I'll break your neck. It is obvious that these are no more commands than is Can I have a nickel? an inquiry about one's ability to own five cents. What is less obvious is that they are not imperatives. To exclude them from imperatives, or try to distinguish a pure set of imperatives that are commands and nothing else, may simplify the analysis of commands but it complicates that of conditions. Consider the following conjunctions: A fellow gets a few gray hairs and they think he's ready for the ashheap. A telegram comes and you're sure its bad news. Macy's advertises a sale and the whole town goes crazy. or adjective clauses: Any man tries to talk back to his wife usually catches it. The employee who presumes to advise his boss risks losing his job. or especially adverb clauses of time: When you accept I'll accept. Once you see the effect of those hard words, you'll want to speak more softly. To the last example one might add the explicitly conditional if to counterbalance the once clause: but if you don't, I guess I'll have to resign myself to your continuing to be rude. It seems that cause-effect and condition-consequence are not limited to specific structures. It would seem easier to call the structures what we have traditionally called them, and assign their conditional uses to figures of speech within discourse. The speaker who says A telegram comes and you're sure it's bad news is saying in effect A telegram comes — just imagine this; then imagine what comes next — you're sure it's bad news. This is exactly the type of discourse that includes all phases of story-telling: "just imagine." And we don't on that account categorize story-telling statements separately from fact-reporting ones. But there is a bit more to the question of imperative conditions. We can trace it by starting with sentences containing impersonal you, which are especially common: You try to please somebody and all you get is a kick in the pants. You have half a degree of temperature and they put you to bed; you look a little pale around the gills and they act like it's time to call a priest. You tell him anything and he just looks at you blankly.

THE IMPERATIVE IN ENGLISH

341

These are all normal without the and, e.g., the last one: You tell him anything, he just looks at you blankly. This suggests that the and may well be parasitic. If we so regard it, then all barriers between if conditions, conjunctions, and imperatives are virtually flattened: ([If] you) tell him anything, (and) he just looks at you blankly. The fact that anything rather than something can be used in any one of these three telescoped conditions is significant. Now compare these with the well-known aphetic questions: ([Do] you) like it? ([Would] you) care to come along? ([Can] you) tell any difference? I suggest that the same thing has happened with conditions, and that in at least some cases there has been an aphesis of the initial if or if you, which produces something with all the appearance of an imperative and accounts for those supposed conditional imperatives that least resemble commands. The process of aphesis explains some conditional "imperatives" without the subject you: ([If] I) give you a nice present, (and) you don't appreciate it. ([If] they) expect any kind of sensible answer from him, they don't get it. ([If] I) buy myself a few pretty clothes, (and) you act like you'd been robbed. ([If] they) give themselves the least advantage, (and) the rest of the gang hollers unfair. It also explains the ungrammaticality of certain conjunctions of two apparent imperatives : ([If] we) take our medicine (and) we get well. T a k e our medicine and get well.6 Take your medicine and get well. Since no aphesis is possible in the second verb, only genuine imperatives will get by. Aphesis appears to occur pretty freely with all subjects except third singular: If she buys herself a few pretty clothes, her husband acts like he'd been robbed. She buys hereself a few pretty clothes, (and) her husband acts like he'd been robbed. *Buys herself a few pretty clothes, (and) her husband acts like he'd been robbed. •

Grammatical, of course, as a fragmentary answer to the question What should we do?

342

DWIGHT BOLINGER

Nor do matters seem to be helped if, on the assumption that this may not be aphesis after all, we use the plain verb stem instead of the inflected form: ?Buy herself a few pretty clothes, (and) her husband acts like he'd been robbed.7 Buy myself a new suit, (and) my wife raises the roof. There seems to be no similar aphesis of preterit forms: ([If] you) sell at too low a price, (and) you make no profit. *Sold at too low a price, (and) you made no profit. The only thing that seems to work consistently well is aphesis of a subject whose verb is identical in form to the plain infinitive. The other way of looking at it, which is to assume that this is the plain infinitive attached to various subjects including third person, does not seem inviting because of the exclusion of third-singular subjects; I give some more examples of the latter: ?Shake down too many people, (and) he gets caught. Shake down too many people, (and) they get caught. ?Tell herself that it's true, (and) she ends up believing it. Tell myself that it's true, (and) I end up believing it. ?Get careless, (and) he gets hurt. Get careless, and you get hurt. The best, of course, is you, but the acceptability of the others is improved by having a verb form that resembles the one with you. Of course, you and its verb form are the same as for the true imperative, and we must now see what the latter's impact may be. It is most dramatic with the verb to be: If I'm critical, my friends hate me. I'm critical, (and) my friends hate me. *Am critical, (and) my friends hate me. *Be critical, (and) my friends hate me. 7

As a reprise, third singular can get by, but then has the support of a parallel sentence fragment. Thus I find "Sooner or later he's got to stop running." "Give himself up and they'll hang him." a bit strained, but not "The only thing for him to do is give himself up." "Give himself up and they'll hang him." (I have provided two contexts for an example from Randolph Quirk). On the other hand, oneself is normal: Make oneself a slave to another person and one loses all self respect. This — given its formality by contrast with impersonal you — suggests that perhaps the conditional ties ought to be with the older //-clauses with subjunctive: If one make oneself a slave etc. So for Jespersen's example (V, 24.3 4 ) Give you women but rope enough, you'll do your own business, we may add if one. The factor working to preserve one but not he nor she in these conditions is probably its indefiniteness. The more hypothetical the condition the better it suits conditional 'imperatives'.

THE IMPERATIVE IN ENGLISH

343

If they're scoundrels, people admire them. They're scoundrels, (and) people admire them. *Are scoundrels, (and) people admire them. *Be scoundrels, (and) people admire them. If you're a scoundrel, people admire you. You're a scoundrel, (and) people admire you. *Are a scoundrel, (and) people admire you. Be a scoundrel, (and) people admire you. If they're left behind, nobody will pity them. •Be left behind, (and) nobody will pity them. Be left behind, (and) nobody will pity you. If we're as good as our word, others will trust us. Be as good as our word, and others will trust us. All constructions with the verb to be, including the passive voice, appear to exclude all subjects other than you and inclusive we. The plain infinitive stem of to be is be. No subject, including you, is compatible with this form. But you is compatible with it in the imperative, and the effect extends to inclusive we. So it develops that with this one verb the imperative becomes decisive. It seems that the so-called conditional imperatives are a confluence of two streams: one from aphesis of subject (and possibly i f ) from non-third-singular verbs, the other from the imperative. The fact that the non-third-singular is the same in form as the imperative makes it difficult to tell where the water is from once the two streams have met. There are two restrictions that partially clarify this, one on the use of any and the other on the use of statives.8 The restrictions on any shape up as follows: If you find any tickets, we'll go to the movies. *Find any tickets and we'll go to the movies. If you write any letters I'll mail them for you. *Write any letters and I'll mail them for you. If you make any soft cereal for tomorrow's breakfast I'll feed some to the baby. *Make any soft cereal for tomorrow's breakfast and I'll feed some to the baby. 9 If you meet John, he wants you to go home with him. *Meet John, and he wants you to go home with him. 8

These restrictions were brought to my attention by George Lakoff. It is necessary to peel off irrelevant examples of any. An any implying 'that there may be' does not count here. Usually such instances carry modifiers, e.g., Just go out and bring in anybody who hasn't already signed up, Make a list of any missing items and the police will help. Sometimes not: Let us in on any news (that there may be). The contrast here is not between an indefinite any vs. an 'any whatsoever' but between an indefinite any and an any = if any: 'Let us in on the news, if any.' 9

344

DWIGHT BOLINGER

The incongruousness of the last example without any exposes the nature of the restriction, which is less obvious in the others: 'imperative' conditions must be true conditions, that is, conditions of intrinsic consequence. The use of an explicit if makes it possible to form outlandish pseudo-conditions like There's a man in your office, if you want to see him ('If you want to see a man you can, because...'). 'Imperative' conditions are limited to those whose consequences are the automatic result of the condition. In * Write any letters and I'll mail them for you the willingness to do a favor is not consequent upon the writing of the letters. The any forces the clause to be a condition, and it is the wrong kind of condition because the consequence is not automatic (a some would make the sentence a command, and the problem does not arise: Write some letters and I'll mail them for you). On the other hand, if John is such a friendly fellow that he no sooner sees you than he invites you to his house, then Meet John, and he wants you to go home with him becomes grammatical. Other examples: Pay any attention to people like that and they never let you rest. Invent anything new and the public goes wild about you. This is the reason why conditions with IMPERSONAL you, like these last two examples, are so common. They are the sort of generalizations about the consequences of virtue and folly, mostly the latter, that make up a large part of folk philosophy. But particular conditions are perfectly grammatical if the consequence is intrinsic: Try any harder and you'll bust a gut. Find any proofs and I'll believe you. We are now in a position to ask what the relevance of any is to conditions vs. commands. We have seen that there are conditions that exclude any, and exclude it qua conditions, not qua commands. On the other hand, we would expect imperatives that are commands also to exclude any, in the light of obviously ungrammatical examples like *Let me have any of your time tonight, will you? *Please go with anybody to church tomorrow. *Write anybody a letter of condolence.10 Nevertheless there are some apparent commands that admit any: All right then. Put up your fists. Let me hear you call me any of those names. Go ahead. Just try it. If you're real brave, try him with any of those wild ideas. 10 Not even EMPHATIC any will work in an imperative of the first instance: *Hey, George! Call anybody on the telephone!; * Well, Mary, tomorrow your vacation begins — go Anywhere!', *I want to relax — tell me dny story. But as second instance, they get by: I want to hear a story — tell me any story; I just can't make up my mind where to go — Go anywhere! You're a free man! In Joe's OK — ask dnybody! we have a concealed condition: Ask anybody and he will tell you that Joe's OK.

THE IMPERATIVE IN ENGLISH

345

These are dares with unnamed consequences. The consequence, when named, is a threat: the speaker wants to present it as inevitable. The result is a condition with an intrinsic consequence. Conditions that would not get by otherwise do if they contain threats. The sentence Make any soft cereal for tomorrow's breakfast and I'll feed some to the baby could be said by a person who dislikes soft cereal and threatens to upset the household by feeding it to the baby, who will kick up a fuss. The sentence Break one of your teeth and your face will be ugly is a normal condition; the sentence Break one of your teeth and I'll take you to the dentist is either a silly invitation to break a tooth (and hence probably a command, as in Break these pieces of candy and I'll take them in to the guests) or it is a condition containing a threat, the dentist being the punishment. The above examples can be completed: Let me hear you call me any of those names and I'll show you what's what. Try him with any of those wild ideas and see what happens. Such vague consequences are frequently omitted.

Threats are interesting in that

they are often rigged to sound like commands — this is the nature of a dare: Just you make any of those smart-aleck remarks of yours and you're going to get slapped down. Aside from the cases cited, I believe that any is a serviceable test for commands vs. conditions. With statives, we find that intrinsic consequences are again significant: If you like her I'll introduce you to her. *Like her and I'll introduce you to her. If you like her her friends will love you. Like her and her friends will love you. If you own this property I'll buy it from you. *Own this property and I'll buy it from you. If you own a piece of property you get taxed unmercifully. Own a piece of property and you get taxed unmercifully. If you understand Chinese I need you for a teacher. "Understand Chinese and I need you for a teacher. If you understand Chinese you can get any of these jobs. Understand Chinese and you can get any of these jobs. If you are John Smith this message is for you. *Be John Smith and this message is for you. If you are sick they put you to bed. Be sick and they put you to bed. If it's six o'clock I'm going home. *Let it be six o'clock and I'm going home.

346

DWIGHT BOLINGER

If it's six o'clock all the workmen rush for home. Let it be six o'clock and all the workmen rush for home. Statives thus appear to be a test. If they can be readily used the construction is a condition with intrinsic consequence. Taking any and the statives together, it seems a safe generalization that straight conditions (conditions with no admixture of commands) using the simple verb stem must be conditions of this type. There are of course many straight conditions to which neither of the two tests applies. If commanding the action would be absurd, it is usually safe to assume a straight condition : Eat too much candy and you get a stomach ache. Be overly polite and people will be suspicious. Disobey your parents and you'll get punished. Dynamite the White House and you'll get arrested. But in other examples without context it is impossible to tell whether condition or command is intended : Eat your spinach and you'll be strong. Be our candidate for president and we'll win. Join the Navy and see the world. In still others the command is obviously uppermost : Hand me that hammer, will you, and I'll nail this down. Step this way, please, and the doctor will see you. We conclude that if K-P intended to exclude conditions they were right, though it is now clear that we cannot exclude all apparent imperatives that CONTAIN or IMPLY conditions, for some of them are truly commands. 11 2. Wishes, hopes, advice We noted earlier that 'futurity' was, for K-P, part of the definition of 'command'. It seems that 'possibility of compliance' is also part of the definition, for they star examples like Believe the claim, Understand the answer, Want more money, Hope it rains, and observe (77) that it is "anomalous to request someone to do something which he cannot willfully choose to do". To adhere to this condition results in some curious exclusions. First are sentences to which no traditional grammarian would deny the status of commands, but would recognize as figures of speech: Go to hell. Be an angel and hand me that comb. Tell it to the Marines. 11

Some further examples of conditions, with don't, appear on p. 361.

THE IMPERATIVE IN ENGLISH

347

Second are certain sentences with statives : You be young and I'll be old. The make-up man can manage it. Remember that name. Some day it will be famous. Just hope you have the courage when the time comes. Here. Take the money. Get in through the back door. Leave the bag on the table. And hope that nobody sees you. Just be very clear on this point. I want no misunderstandings. Are you ready now? May I proceed? Now hear this. Hear me out. I have only a few more remarks. Hear the difference between these two passages. Play them over again, Joe. Third are wishes of various sorts and facetious advice: Be happy! Get well! Sleep well! (Psychologist to young man on how to get ahead:) Choose the right parents. (Barry Goldwater, same purpose:) Inherit a department store. That these are felt to be commands is suggested by certain pretty clearly injunctive contexts, e.g., Be happy! Look on the bright side! Get well — that's an order! Sleep well — mind you, now, I mean it. Get over that headache, now. Get busy and inherit a department store. Another requirement would be a distinction between affirmative and negative commands. For example, ¿»^-passives are awkward (unlike get passives, e.g., Get paid off before noon today; I want to leave early), but not in the negative: Don't be examined by a chiropractor; go to a real doctor. Certain statives are better in the negative, others are worse: I've been thinking of owning a place as soon as my wife and I move. — No, no. Don't own if you have a chance to rent. Owning presents too many responsibilities. Confess, but don't hope for any forgiveness. He's too hard-hearted. Don't believe that pack of lies. Can't you see how dishonest he is? ?Don't hear the difference between these two passages.

348

DWIGHT BOUNGER

The difficulty with any such criterion as what one can or cannot wilfully do is that it is too hard to tell.12 When Canute commands the waves or Jesus says to the man with leprosy, Be clean, it is better to think of figures of speech, or of the speakers as deluded or possessed of a higher power, than to try to decide whether what they say is a command or not. 13 I believe that no useful generalizations relevant to the task here can be made about advice or hopes or wishes as distinct from other imperatives, and that it is beside the point to ask whether in an exchange like They complain that they can't read my letters. — Write more clearly. That way they won't complain, we have to do with a command, a piece of advice, or a condition (Write more clearly and they won't complain). It is true that there are interesting combinatory effects, and there must be countless restrictions with particular verbs or classes of verbs that need to be spelled out, but our business now is to get a perspective on imperatives as a whole.

III. IMPERATIVES IN PRESENT AND PAST

Possibility of compliance and futurity are fundamentally a single criterion. The present and the past cannot be acted on. The future can. With this requirement laid aside, we can now ask whether imperatives may be past or present. The question is pertinent, for it is raised by traditional grammarians for whom lack of any tense but future was not a criterion but an empirical fact — they just had trouble finding examples. Jespersen is dogmatic: "The imperative always refers to the future, often the immediate future" (V, 24.17). Poutsma gives a way out: "Uncertainty of fulfillment naturally clings to a command or request" (162). Uncertainty of fulfillment can attach to a desire about something present as well as something future. Poutsma gives (329) the example I hope you're thinking about me. Please, be thinking about me, which is probably intended to mean a hope that something has already started and is going on now. In any event, such imperatives are normal enough. If a man and his wife are caught after committing a theft, and the wife is uncertain whether the husband has the loot on his person or has already disposed of it, she may say to herself, Please, don't have that money on you! A person holding a lottery ticket not yet examined, and hearing the announcement of the winning number, might ls One may not wilfully choose an action, but still wilfully choose to resist it or not to resist it. In Oh, come on; be taken in just once — it isn't going to hurt you; do you think you're perfect? the command is not to resist (a commoner form of this is let yourself be). We have more occasion to command resistance than sufferance, and negative passives are correspondingly more frequent: Don't be frightened (don't let yourself be frightened) by anything he says. Most passive commands with be are otherwise unacceptable: *George, be taken to church by your sister. 13 K-P seem a bit uncertain of their own criterion, for they admit as grammatical the sentence I request that you forget the whole matter in spite of the fact that one can hardly turn the memory on and off at will.

THE IMPERATIVE IN ENGLISH

349

say, before turning the ticket over, Please be the right number! A mother, hearing of some possible danger to her child if he is still on his way to school, might say the prayer Please be at school already! Someone approaching a fallen and hated enemy might say Be dead, damn you! Of course these are all Irish futures: 'What'll he be thinking of me? It'll be ten o'clock already and I'm still here' — the action is present and the verification is future. But it can hardly be maintained that the speaker is thinking about the future if he uses words like now and already. So much for imperatives referring to present time. What of imperatives referring to the past? The handbooks are almost unanimous in citing Have done as no more than a curious relic, and on the face of it one might imagine that a past command is impossible. (Begone looks to be a similar form. Imaginary situations are always possible, as when an actor-impresario in a home-made drama says I've got it! Be born in 1898. That will make our time sequence work out). Yet it ought to be possible to express "what is merely a wish" concerning a past act, if one has not yet verified whether the act has been carried out. English has two perfectives which can be used to show relevance of a past act to the present, which is obviously what is called for here. One has the form He has the work done {already), the other the form He has done the work. The first, being superficially a present ('He at this moment has the work done') would seem to lend itself more readily to an imperative, in so far as we have seen that imperatives can apply to the present. But the two perfectives are not interchangeable. The first is limited to transitive verbs, and is even further restricted to constructions in which the literal sense of having the thing in such-and-such a condition is uppermost. Thus They have planned their dinner and They have their dinner (all) planned are both normal; but while He has discovered the hiding place is normal, *He has the hiding place discovered is not. We are accordingly left with two questions: (1) Does the have the work done perfective occur more readily than the have done the work perfective, as would appear from our dead reckoning?; and (2) Given a situation where the first of these cannot be said, i.e., where the have done the work type is required, is it still possible to have an imperative? If the case of inventing examples is any indication, the answer to the first question is yes. In a telephone conversation, A invites B to come over and share the contents of a bottle. B says I'm on my way, and then adds, to himself, in an undertone, Only for God's sake don't have three-fourths of it drunk already, knowing the condition in which his friend will be once that amount has taken effect. It would be much more difficult to express this as Don't have drunk three-fourths of it already. It is easy to show that the FORM, and not the pastness, is what makes these constructions difficult, for they can be rigged to refer to the future. A guest might say on receiving a sudden invitation by telephone, Just don't have all those good things eaten before we get there, which is normal. Don't have eaten all those good things before we get there is as difficult referring to future as to past. (For one thing, a perfective referring to action rather than state is unmotivated here, for there is too little

350

DWIGHT BOLINGER

contrast between it and a straightforward future: Don't eat all those good things before we get there). The answer to the second question is also yes, given a situation in which there is a strong wish for a past action that has not been verified. I submitted the following passage to twelve native speakers of English, allowing one minute to read it and two minutes to answer the questions that appear below it, with further reading if necessary: Aline looked out the window of the taxi which moved by inches toward the next traffic signal and the next and the next. She gave a frightened glance at her watch. Five o'clock! Neale would be leaving the office. If only she had not written that note. That stupid, reckless, incoherent note, scrawled in three minutes and left on the mantelpiece. How cowardly to confess in writing. He would never understand! The taxi broke free of the chain of traffic and sped along the boulevard. Five-fifteen. He would be home by now. Her lips formed an agonized wish that was a prayer: Please, Neale, don't have read it yet! Let me be there to explain when you do. Let me tell you what was really in my heart. They drew up to the curb. She fumbled in her purse, gave the driver a dollar, and stepped out. If you mark 3, ignore 2 1. The wording of the passage seems normal. 2. The wording of the passage has one or more places that struck me as unusual, but not un-English. (Underline them.) 3. The wording of the passage has one or more places that struck me as incorrect English. (Underline them.) Only two persons marked anything as 'incorrect English': one thought that out rather than out of was wrong, and another questioned that chain of traffic and agonized wish that was a prayer. All the rest felt that the wording was unusual in spots, but not un-English. Their choices were "window of the taxi" (1), "moved by inches" (5), "and left" (1), one or both halves of "broke free of the chain of traffic" (4), "agonized wish that was a prayer" (1), and "don't have read it yet" (3). In other words, only three of the twelve noticed the imperative at all, and no one regarded it as unEnglish. A later comment: "If that is the idea you want to express, how else can you express it?" The restriction, if there is one, does not relate to time but — I think — to accent. The auxiliary have is accented in definite situations, e.g., as a carrier of affirmationnegation : Have you done it? ('I want to know the truth'); He has done it (contrary to what they have said).14 These situations do not include the imperative, which is one reason why we recognize have done as a fossil: have is stressed. But the imperative needs the accent, which can be attached to a do or a don't. Do strikes me as more unusual, though I would not find the following impossible: Please, do have made that call by six o'clock; 14

Also as a stressed modal and as a stressed tense. See Lila R. Gleitman, "Coordinating conjunctions in English", Language, 41 (1965), 260-293, p. 287.

THE IMPERATIVE IN ENGLISH

351

Please, do have made the effort at least once!; Do have given some thought to the question, once you've decided to discuss it. Don't, which is more frequent over-all than do, gets by more easily. An added and highly acceptable dare fits in well here: someone who was looking forward to participating in a job and appears on the scene uncertain whether he still has the chance, might say to his supposed collaborator, Don't you dare have finished that work yet!; this is another unverified wish, but is unlike the others in that it can be said directly to the person concerned, and not as a form of absentee apostrophizing. The other perfective is less restricted — the have can be accented: Please, have that work finished for once, so it won't all fall on me; Don't you dare have that work finished yet! It also admits other complements than past participles: Have the returns in already. Have the buns hot already.

IV. THE IMPERATIVE AS BARE INFINITIVE

It was Jespersen, the favorite traditional grammarian among generativists, who emphasized the futurity of imperatives (IV, 7.4 (1), besides the reference already given), the relevance of tagged will (IV, 15.7 (5)), and the need, even with an expressed somebody as subject, to infer a you (V, 24.1 4 ; Essentials, 15.42). But Jespersen makes two appeals to intuition, and they are not entirely consistent. The first is the appeal to will: "Indeed, it is possible that to the actual speech instinct the imperative is nothing but a kind of abbreviated will sentence: Have an egg = Will you have an egg?" (IV, 15.75). The second is an appeal to the form of the imperative: "As the imperative is formally identical with the infinitive, it may by the actual speech instinct be felt as such" (V,24.2X). Curiously, he uses the same evidence a second time: the tag question in Take a seat, will you? His two ideas could have been separated more clearly had he used an example like Have a chocolate. — Oh, may I? Thanks! Kruisinga goes farther. He lumps the imperative together with other examples of the 'plain verb stem' used predicatively (§§ 171-185), including types like Why refuse to come? and Gerry Palliser swim! Of course he can't (§ 183), and says "The predicative plain stem can express a command or entreaty" (§ 172). Curme points out another analogy with the infinitive (§ 44 II 3a, p. 419): "In reporting indirectly a command we never employ the imperative, but an optative — the volitive — subjunctive or an infinitive with the force of a volitive subjunctive: (direct) 'Come at once', (indirect) 'He said I should come at once' (or / was to come at once), or 'He told me to come at once.'" The kinship is closer than this implies, for the types He said I should come at once and He said I was to come at once only rather loosely report the meaning, not the form, of the imperative. The direct form of the first is You should come at once, of the second You are to come at once; it would be as accurate to say that He pointed out the necessity of my coming at once reports Come at once. The virtually unambiguous reporting of an imperative is with an infinitive: He told me to come at once, He bade me come at once, and, with

352

DWIGHT BOLINGER

the subject deleted exactly as in the direct command, He said to come at once.u The important thing to note here is that finites (statements and questions) are reported with finites (It's a nice day, He says [that] it's a nice day; Is it a nice day?, He asks if it is a nice day), and non-finites (Come at once) with non-finites (He says to come at once).19 This analogy between direct and indirect discourse points, I think, to a tie, within direct discourse, between imperatives and infinitives. So our three traditional grammarians present us with two intuitions as to the kinship of the imperative: one, Jespersen, a kinship with will sentences; all three, including Jespersen in a different mood, a kinship with the infinitive. I want to see how far the latter idea can be carried, i.e., the extent to which the two can be lumped together. V. THE UNDIFFERENTIATED INFINITIVE

If we follow Kruisinga we will recognize a "plain stem used predicatively" and not fall back on "ellipsis" to account for things like Why bother?, How make them understand?, John be rude with anyone? If the imperative is to be dignified by being recognized as a structure in its own right, these deserve the same attention. The question then boils down to looking for interrelationships among these formally similar structures, to establish an identity among them. The prima facie evidence favors this: the imperative and the bare infinitive are identical in form. It remains to make a case for such a close relatedness in distribution that the distinction between 'imperative' and 'infinitive' is reduced to the same kind of context that distinguishes May I have the sugar? spoken at table and May I go with you tomorrow?, where a mere yes is a satisfactory answer for the second but not for the first. (1) What Jespersen calls "characteristic clauses of indifference with a preposed verb in the crude (stem) form" (IV, 15.4 (7)) would be felt, at one end of the gradient, as imperatives by most native speakers of English, if the subject is you. Yet they blend imperceptibly with older instances where the subject is not you. In fact, you has asserted itself to the point where other subjects would nowadays generally be felt a bit odd. Of Jespersen's two examples Go whither he will, Vie be none of his followers and Look where you will, immeasurable Obscurantism is girdling this fair France, only the second (after eliminating the other quaint features and reducing the first to Go where he will or Go wherever he pleases and the second to Look wherever you please) would be usual in spoken English today. In the following, the first three are surviving fossils: 15 This last formula both reports and conveys an imperative. If A is asked Why are you in such a hurry? and replies B said to come at once, he reports B's command Come at once. If A says to B Tell C to come at once, B may give the command directly, Come at once — A's orders, or indirectly, A says to come at once. 16 I refer, of course, to the verba dicendi. The verb to claim may report with an infinitive (He claims to be the heir) but not with the change of subject that causes no trouble with the imperative (*He claims John to be the heir).

THE IMPERATIVE IN ENGLISH

1. 2. 3. 4. 5. 6.

353

Come what may, we shall not fear. Be that as it may, my argument still stands. Try as she will, she can't do it. Say what you will, they won't believe you. Lie all you want, it can't hurt me. Go where you please; I don't give a damn.

All it takes to make 4 and 5 unmistakable commands is a conclusive intonation, like that of 6. But that is all it takes to make the same change of meaning in He can eat anything he likes, it won't bother me ('No matter what he eats, it won't bother me') vs. He can eat anything he likes; it won't bother me, and I doubt that we would think of classifying He can eat anything he likes in two ways in order to take care of that semantic contrast. It is hard to insist on any clear-cut separation between imperatives and non-imperatives in utterances like the ones I have listed. (2) With conditions we have already seen at one extreme a kind of sentence that could not be a command (e.g., Kruisinga's example, § 182, Give him a fact, he loaded you with thanks), at the other extreme a kind that contained some element of condition but was obviously meant as a command {Here, take this pill and you'll feel better), and in between and covering the extremes a tendency toward mutual accommodation: (1) The more hypothetical the condition the better (a characteristic, as we shall see, of the bare infinitive): a preference for indefinite subjects and a requirement of intrinsic consequence. (2) A restriction of the verb to be to its bare infinitive form. (3) The inclusion of dares, which are a kind of perverted command, in the scheme of condition-consequence. The accommodation has reached the point where a sentence like Join the Navy and see the world can be intended both ways. Usually when two different constructions overlap (produce homonymy) and are intended both ways, we recognize a pun: I can drink a Venetian blind. But with command-conditions we do not sense that any pun was intended. This suggests that there is no pun, that the structures are felt to be identical, and that the common ground is the bare infinitive with a meaning of hypothesis. (3) With predisposing context, I find another gradient. I believe that any speaker would feel that the first examples below contain imperatives, yet I find identical utterances with merely a change of person perfectly normal: 1. How can I convince them? — Tell them the truth. 2. What can I do to smoothe things over? — Just be more polite in the future. 3. What must I do to get ahead? — Work harder. 4. What should I do? — Write that letter. 5. How can she convince them? — Tell them the truth. 6. What can they do to smoothe things over? Just be more polite in the future. 7. What should he do? — Write that letter.

354

DWIGHT BOLINGER

Again, I have no sense of punning when I view the answer in No. 1 as either Tell them the truth, a command, or You can tell them the truth, an 'ellipsis'. It happens that all these examples can be viewed as replacements after the auxiliary: What can I do? — You can be more polite-, What must I do? — You must work harder, etc. Also in Should they go now? — No, see what happens and if necessary go later it is possible, but a trifle far-fetched, to insist that there is an ellipsis of No, they should see what happens and if necessary (they should) go later. Farther out on the gradient the argument for ellipsis gets weaker. In the example The poor fellow doesn't know which way to turn, he's got himself in such a complicated mess. — Make a fresh start. That's the only way for him to put himself in the clear there is no obvious frame. 17 In the first and second of the following examples, the "imperative" may be blended with indirect discourse (Vd say, "Let yourself be arrested"), but the third is not: She wants to know the best career for a woman? I'll tell her: get herself a husband. "He's afraid the police are going to hang something on him." — "In that case, I'd say let himself be arrested and then sue for false arrest." "Why can't he support his family?" "He doesn't earn enough on this job". "Then get himself another one. I don't see why we should be expected to support him". A final predisposing context is alignment with a you imperative. In the following example from Poutsma (201), it is not necessary for George to be within earshot: No; you get the paper and pencil and the catalogue, and George write down, and I'll do the work. The speaker might add, Go call him, will you, and tell him we need him and what he's supposed to do. The examples to this point have all been either what would be obvious as con17

One could argue for a blend with The poor fellow doesn't know what to do... Make a fresh start. That's the only thing for him to do, where the infinitive can be regarded as a nominalized predicate: 'What to do (the only thing for him to do) is make a fresh start.' I purposely avoided this wording in the example, and chose way rather than thing to create one obstacle to the nominalization of the bare infinitive: way, unlike thing or what, requires to with the infinitive: The way for him to put himself in the clear is to make a fresh start. But it really is not necessary, I think, to lean so far backward in order to find a context that is utterly free of a grammatically enveloping frame. When we consider the following set, Would you advise me what to do? — Try harder. Would you advise him what to do? — Try harder. D o you know what happened last night? — The whole East underwent a power failure. Would you explain why he doesn't like it there? — They don't feed him enough. if Try harder is an embedded construction because it fits into the position of the nominal laid for it by what precedes, then so is They doitt feed him enough (is why). Any answer that is directly responsive to an interrogative-word question may be regarded as embedded.

THE IMPERATIVE IN ENGLISH

355

ventional imperatives or are so close that the sharpness of the conventional notion is called into question. For a more comprehensive view of the bare infinitive it is necessary to look at forms that lie clearly outside the realm of imperatives or apparent imperatives, and to ask whether there are nevertheless such fundamental resemblances that the differences can be relegated to intersection with some other system, like what happens when the statement They've left already is turned into a question They've left already? with only a change of intonation. The evidence is mostly semantic, with certain distributional ties that can be fairly clearly stated. I start with those. (1) Restrictions on tense are the same as with imperatives. The question types Why not go there myself? (Poutsma, 436), How make them understand?, Where find anyone more agreeable? occur freely without have, but I find *How have it done sooner? as difficult as the corresponding command, while the other perfective, How have it done sooner?, is possible. (2) Restrictions on statives and non-statives resemble those with imperatives. Thus John be tall? is doubtful, but in "Your brother practically insulted me" "John be rude to a friend? That's certainly not like him" non-stative be rude is normal. The omission of the copula is normal with statives: John sick?, An American a traitor?1* (3) Omission of the subject is largely optional, as with the imperative: "If old Doc Jones doesn't collect some of those bills, he's going to go bankrupt." "(Old Doc Jones) dun his patients? He wouldn't have the heart". (4) The bare infinitive readily occurs in an answer that refers to something unaccomplished, but not otherwise: 18

Otherwise, e.g., when the reference is to a fact rather than a hypothesis, omission of the copula is more in the nature of a sentence fragment. In the following exchanges, "I hear that Henry is staying at a hotel and Georgette has gone home to Mother." "Henry and Georgette mad at each other? Well, I can't say that I'm surprised." "I predict that those two are going to get on each other's nerves in no time at all. I'll bet that within six months Henry will be staying at a hotel and Georgette will be home with Mother." "Henry and Georgette be mad at each other? You're kidding." the form without be might do for the second but the form with be will not do for the first. In the first a fragment like Henry and Georgette fighting? would be equally normal. In the second, if the copula is omitted one has the sensation of two strings rather than one: Henry and Georgette? Mad at each other?

356

DWIGHT BOLINGER

What are you going to do? — Make the most of it. What did you do? — *Make the most of it. The one who asks the question, however, may readily combine these: What did you do, make the most of it? — here the matter of accomplishment is still up in the air. The same restriction shows up when the response is someone else's proposed future action (in this case we have the borderline with commands again): If he won't believe them, what do they do to convince him? ( = 'What are they supposed to do?'). — Tell him the truth. I marvel at their always getting him on their side. Just what do they do to convince him? — *Tell him the truth. I have starred the latter, though it could occur as a truncated finite (cf. Got home early last night, went to bed, didn't wake till ten). Again, the person asking the question can include the suggested but unconfirmed answer in his question: What do they do to convince him, tell him the truth? In tag questions, the un-Auxed negative may occur when the action is viewed as unconfirmed, but not otherwise. The intonation of course differs also: Not sell it did

('He failed to sell it, did he?')

(He) didn't sell i t did he? This is the usual form taken by the negative with hypothetical bare infinitives. It is diagnostic in another way also, which will be taken up later. Difficulty of the past and unaccomplishment of the action suggest that bare infinitive forms have a semantic element in common: that of hypotheticalness. This is obvious in the type already cited, Why wait?, Why not call them up now?, How believe in such a fantastic story? It is also obvious in the imperative, where lack of accomplishment or confirmation was seen to be the common semantic thread. Other common types are the following: (1) Turn my father against me — that's what she wants to do. The action is potential. *Turn my father against me — that's what she tried to do, for something past, is ungrammatical. (2) Come over and visit us this evening. — And leave my husband to fix his own supper? I wouldn't dare. The hypothetical imperative meshes with a hypothetical consequence with no more sense of a grammatical change than would be felt in combining two finites: She came over to visit us last evening — And left her husband to fix his own supper, I daresay. (3) It seems to me that you could at least be as generous as your predecessors were. — Oh, sure: make you one concession and find ourselves making you another and another; there would be no end to it. This is an inference, where the infinitive might be introduced by you mean, that is to say, etc.

THE IMPERATIVE IN ENGLISH

357

(4) Buy a few shares of Hayden and Smith. The price is right. — Hmm. Buy Hayden and Smith. Not a bad idea. The action is something to be mulled over. (5) Poutsma's example (437), My nephew marry a tragedy queen! may refer to something that has already happened, but which the speaker figuratively refuses to accept. English has one verb, to suppose, which makes quite explicit the relationship between hypotheticals in general and commands as hypotheses: it is used for both. First, as equivalent to the hypothetical question what if?: "I order you to do this". "Suppose I refuse". "Then I'll just have to make you". Second, in conditions: Suppose I show you how, will you do it? (If suppose is replaced by supposing, the intonational break is reduced). Third, in proposing a hypothetical line of action: Suppose I go in while you wait out here. Sentences like these can be interpreted more and more in the direction of commands, as in the following gradient: Suppose you go in while I wait out here. Suppose you take care of it. Right now. Suppose you go in and ask him, will you? Now suppose you just let me be the judge of that, please. As with other requests for commands, when the first person is involved the tagged auxiliary is shall: Suppose I go in and ask him, shall I? Suppose we both try, shall we? While suppose has been stereotyped in this use (and imagine, assume, and probably most other verbs would sound strange) it is not exclusive. The same meaning of hypotheticalness is present in the following: Let's just say you march down there right now. I'm getting a little tired of waiting. The infinitive, in this perspective, is a general hypothetical that includes the imperative as its most frequent manifestation. The action of the imperative is hypothetical — the command may or may not be carried out. The rest of the identification comes from the context. The mandatoriness is from the intonation: Let m e

have

a

^

spoken with the normal straight fall, which distinguishes commands from other utterances (any one of which can be made more or less mandatory by using this intonation), e.g.,

358

DWIGHT BOLINGER

How about

. a little hght

a 1Jtt1

o v e r h e r e ? i9

If spoken with a C accent, a native speaker of English would probably say "That's a request, not a command": LCt m e h a v e

a nicke1

and might even hesitate about how to punctuate it, a problem which would be settled in favor of a question mark in Le, Me have a nick=" which is closest to Jespersen's (not K-P's) notion of the imperative as a 'will' sentence: Will you let me have a nickel? As for the subject, it comes from the speaker-hearer situation, where it is commonplace to omit you: Smell that soup? Going there tomorrow? Like a chocolate? Other instances of the bare infinitive also leave the subject to the situation: the persons are reversed in Why pick on me? — I haven't done him any harm and Why pick on him? — he hasn't done me any harm; and Why do that? can mean 'Why should you, we, anyone do that?' By limiting their examples to utterances in which there is no preceding verbal context, K-P and others have unwittingly IMPOSED the only element of context which is left, that of A addressing B — the vocativeness is all there is to fall back on. It is no wonder then that utterances like Go now, Call mama, Read this book, Hand over all you have, etc. lead to the conclusion that everything of this kind has to have a you subject and be directed toward compliance. Under the circumstances, nothing else could be read in. Stripping down syntactic samples to minimal sentences in the hope of excluding everything irrelevant may have the effect of creating a vacuum which unexpected irrelevancies rush in to fill. The bare infinitives constitute a single system that derives its variety from intersections with other systems. If we insist on a special derivation for the imperative mode, consistency would require that we do the same with declaratives when they are used as commands: Nobody moves until I give the word, and that means you; You don't speak first, he does; First they come in and then you tell them what to do; You sit here and he sits over there; Please, I hope you're not going to insist; It's time for your medicine, dear. Or with questions: Do you mind stepping to one side?; Why 19

The international dependence of commands is illustrated in Your

na

me

vs.

Your nam

e?

each responded to in the same way, but the first felt as a command and the second a question ('Tell me your name'; 'What is your name?'). The bare infinitive is not the only thing intersecting with intonation and a predisposing context to make a command. Any contextually appropriate noun, adjective, or adverb, especially with quantitative modifiers, is equally suitable: Easy now; Not so fast; Faster!; Quiet!; More light down here; No monkey business, now; Harder!; Careful!; Arms front; Head back; Forward!; One side!; Up on top!

THE IMPERATIVE IN ENGLISH

359

don't you be more careful?; Would you kindly leave me alone? A similar case would be the insistence on a vocative element in the imperative where again to be consistent it would be necessary to add a vocative to Does the master wish to sit here? addressed to the master. This is not to say that we should no longer talk about the imperative mode. There must be more instances of the bare infinitive used for commands than there are of all other uses combined; for convenience we should call them imperatives. My only aim here has been to try to show that there are better reasons for calling the imperative an infinitive than for calling it a will sentence.20 20 If the suggestion of the will sentence is rejected, what other ways of handling the imperative generatively are there? One is suggested by K-P themselves, which is to make the starting point a sentence with a 'request' type of verb: I Verbrequest that you will Main Verb (note 9, p. 149). This is to avoid assuming an Imperative morpheme, which becomes unnecessary, since request has the same restrictions. (I am perplexed by the retention of will even here. The result is ungrammatical, and does not appear in any of the examples cited.) A solution that is perhaps less clumsy is the now unfortunately obsolescent verb to bid, which also has the same restrictions and calls for fewer deletions: I Verb bid you Main Verb. Poutsma lists beg and pray as similar verbs used with the infinitive without to. Pray is especially suggestive because of what happened after it became fossilized: Prithee go. Since imperatives were actually detached from this construction, it makes a good source. Imperatives were also detached from constructions with please e.g. (May it) please (you to) go, but the machinery is more awkward. Another is suggested by an oddity in the distribution of shall, which is, first, the auxiliary that is used in calling for a command, and, second, is the only auxiliary that is not repeatable in the same sense in which it appears in the question: Shall I let him in? — Yes, do or Yes, let him in (not *Yes, you shall). Shall we go now? — Yes, let's (not *Yes, we shall). Shall they report at the same time tomorrow? — Yes, have them report then (not *Yes, they shall report etc.). Contrast these with the readiness with which other auxiliaries are repeated in the same sense: Must I do it? — Yes, you must. Could it happen here? — Yes, it could. Mightn't it be better to ask first? — Yes, I guess it might. (Might is irregular in this respect. In diffident requests, it is not repeatable: Might I have a look? — * Yes, you might. But this can be taken care of by saying that what is not repeated is the diffident use of the past tense. The interlocutor replies with the "same" verb, may. The same is true, though less strictly so, of could-can. Need is marginal as an auxiliary: Need I say more? Do I need to say more?— Yes, you do, *Yes, you need). This at least is true of American English as I know it, and suggests that shall rather than will is the auxiliary that belongs to the imperative. So perhaps the imperative could take off from You shall do it Do it. But there is a difficulty here in that shall sentences in the same dialects that use shall in the way I have described limit it to commands exerted on the elements, not on the hearer. If I say You shall have it I mean that no outward circumstance will prevent your having it, and You shall go there tomorrow would not be used at all unless the person spoken to had expressed a wish to go. A third solution, and the one that I prefer, is to put the imperative in the kernel, expressing it with a zero Aux, which in question form is transformed to shall. Shall as a question form of the imperative is just as specific to commands as do is to certain finites. (One might even transform the zero Aux to do, but there is an ambiguity: Well, do I get it or don't I? — Yes, you get it, where the question implies a request, or Yes, get it, where the question solicits a command). Ways of incorporating let's (Let's sit down) and have (Have them sit down), which as noted above also pair with shall, would need to be devised; note also in tags: Let's go, shall we? Zeroing the Aux would simplify bringing in do as an emphasis morpheme and don't for negation. One advantage that this has over

360

DWIGHT BOLINGER VI. THE IMPERATIVE DIFFERENTIATED FROM THE INFINITIVE

Where the imperative and the hypothetical bare infinitive diverge somewhat is in the use of do. In the negative this divergence can be seen in the contrast between don't and not, the latter being normal with the h y p o t h e t i c a l : You are not to see him again. — Not see him again? That would be unbearable. (It is worth noting that the sentence fragment Not to see him again? would be equally normal here, and is evidence, of sorts, that Not see him again? is not a fragment). Whereas a don't can be parroted, e.g. after Don't ask questions either Don't ask questions? — that would be contrary to my nature or Not ask questions? — that would be contrary to my nature is appropriate, don't is otherwise sometimes doubtful: Stay home. — Hmm. Not take my usual walk; I don't like that idea. ?Hmm. Don't take my usual walk, etc. The negative imperative, of course, is with don't.21 But the situation with don't is by n o means clear. It seems to be gaining ground at the expense of not. If there is the slightest possibility of viewing a negative hypothetical in a way that would justify don't, we tend to use it. I find What should they do? — Not write that letter uncomfortably formal; I prefer Don't write that letter,

a derivation from a clause is that it is less like putting the cart before the horse. There is a resemblance between I order that he do it and I order, "Do it," on the one hand, and I ask whether he is going and I ask, "Is he going?" on the other. I would not say that it is impossible to derive Is he going? from whether he is going (whether could be regarded as the yes-no question morpheme, as it actually was at one time), but it seems like going at things backwards, and the same is true of Do it from that he do it. 21 But the asymmetry in the negative is no more a disadvantage in coupling the imperative with the bare infinitive than in deriving it from a Verbrequest structure, for the same problem is found there. The English subjunctive is in such a state of collapse that this does not show up consistently, but it is present to a large degree in some dialects and probably to a slight degree in most. The best Verbrequest with which to illustrate is insist, in third-person singular. (I realize that in some dialects examples 2, 4, and 6 are ambiguous): 1. I insist that he take the medicine ('He must take the medicine') 2. I insist that he takes the medicine ('I know that he does') ~f 3. I insist that he do take the medicine (emphatic order) ' 4. I insist that he does take the medicine (emphatic statement) 5. I insist that he not take the medicine ('He must not take it') 6. I insist that he doesn't take the medicine ('I know that he doesn't') 7. I insist that he do not take the medicine (emphatic order) 8. *I insist that he don't take the medicine (a victory for the puristic battle against he don't, which has engaged this construction even though irrelevant to it) Example 5 shows the traditional negative with a subjunctive, exactly the same as with the bare infinitive. In order to get a negative imperative out of it transformationally, the same change is required as to get it from the infinitive. Of course, other examples of Verbrequest have dropped the bars to some extent, just as we have seen don't spread in the bare infinitive form to situations where not formerly prevailed. So I request that you not do it sounds a bit formal nowadays beside I request that you don't do it. The near-quotative nature of this, 'I request, "Don't do it"', naturally facilitates the step. With third-person singular there is no satisfactory solution: I request that he not do it, too formal; I request that he doesn't do it, too obviously indicative; I request that he don't do it, in trouble with the he don't taboo.

THE IMPERATIVE IN ENGLISH

361

despite the third person. There is enough of a suggestion of a command for it to get by. Similarly in Please tell him what to do to avoid those emotional upsets. — Just don't (just not) work himself into a lather every time someone criticizes him a bit — here there is a hint of indirect discourse (Tell him "Just don't work yourself..."). The extension of don't at the expense of not is clearest in the substitution, in some dialects, of don't let's for let's not. Don't

has another point of infiltration via 'imperative' conditions.

We have

seen some evidence of the possible aphetic origin of these. Don't corroborates it. In all conditions where negative don't would accord with the number and person of the deleted subject, I find it more or less acceptable: ([If] I) don't work, (and) I don't get paid. ([If] they) don't fire him, (and) they won't have to pay him. ([If] we) don't hang onto our jobs, (and) how do we eat?22 ([If] you) don't keep up appearances, (and) it's hard to get ahead. But not in third singular: *([If] she) doesn't work, (and) she doesn't get paid. Even with don't it is no good: *Don't work, and she doesn't get paid, and removing the apparent conflict in person and number does not help: *Don't work and she won't get paid. It seems clear that the don't in these sentences is from a finite, not from a hypothetical. The latter normally calls for not, and is ungrammatical in conditions: *Not work and you don't get paid. Nevertheless, the more the condition resembles a command, the more acceptable don't is in third singular. Thus in Don't take more than her share, and she'll be OK the condition is not a gnomic generalization but refers, at least potentially, to a single act of refraining from taking more than her share, and as such is something that can be commanded. Similarly Don't let herself be fooled, and she won't have to suffer the consequences,23 The verb let has such strong ties with commands that an otherwise unacceptable third singular seems to be able to ride in on it: Don't ever let anyone help her, and how can she get along?2* So don't is more or less generalized in conditions on the strength of imperatives. Whereas don't seems to be spreading, do is still specific to imperatives. 25 It is not This does not have to be inclusive we. Compare Never take more than her share, and she'll be O K . ?Never take more than her share, and she can't make ends meet. The first is potentially a command. The acceptability of the first and the doubtfulness of the second shows up even more clearly if never is replaced by don't ever. 24 Similarly with one, which we saw earlier was passable where he and she were not. In the negative, one has more trouble getting by except with let. I find affirmative Deny oneself the help of others, and how can one get along? better than negative ?Don't allow oneself the help of others, and how can one get along? or ?Don't permit oneself to be helped, etc., but not better than a negative with let: Don't let oneself be helped by others, and how can one get along? 28 More than this, to imperatives with a deleted you subject. While I find Don't anybody say a word grammatical, I reject any instance of do with subject before the verb (including the instances of do 23 23

362

DWIGHT BOLINGER

used in hypotheticals, and is unlikely in conditions: in Do come tomorrow, and you'll see our new house the union of the two clauses is forced; it is an afterthought, and nothing like that of Come tomorrow and you'll see our new house, where the first clause may be both a command and a condition. To this extent imperatives are distinct from general hypotheticals. But if we put do in the same category as please (p. 7), I beg you, now {Eat it, now! with fall-rise on now), or other elements of context that show how the infinitive is to be interpreted, this exception is less serious. When we regard imperatives as hypothetical infinitives another rather general trait of commands falls into place: the use of courtesy devices. These include forms of indirection (It's time for...), a change of person (Please, we must be more attentive), an explicit formulation of a wish (I hope...), and posing the order as a question (Do you mind...1). They also include posing the order as a hypothesis: Suppose..., Let's just say..., and the hypothetical bare infinitive. Generatively I would single out two sources, with much overlapping. The first would account for conditions by basing them on finites. The second would account for the imperative by putting it, along with other bare infinitives, in the kernel, or by deriving it from the infinitive. HARVARD UNIVERSITY

you cited by Jespersen, III, 11.84S, though I would accept these as possible literary forms), and this implies that only second-person reflexives can appear: Do help yourself somebody, and let's get started. Do is thus more explicitly imperative in the conventional sense than is don't, and this shows up in one other way. Do cannot even be parroted: Do bring your sister, I so want to see her. — Bring my sister? vs. *Do bring my sister? The latter, if possible at all, would be understood to repeat words more than meaning.

G. B O N F A N T E

ITALIA E GRECIA

N

' E L volume I degli Studii in onore di A. Schiaffini (Roma, 1965), 164ss., ò pubblicato la mia conferenza di Atene (20 nov. 1964) 'TX,coaaiKaì

ayéaeiq neta^ù 'EXXàòoq Kaì 'ItaXiag", in cui ò ripreso ampliàndolo il tema proposto in Hommages à Léon Herrmann (Brusselle-Berchem, 1960), 171 ss.1 In ambedue questi lavori dicevo che il problema è amplissimo e poco esplorato. Vengo qui ora ad onorare il caro amico Roman Jakobson con alcune giunte che mi pàiono di qualche peso, ma che non pretèndono certo di esaurire l'argomento. [La mia conferenza di Atene è stata poi stampata anche nella 'E7tiaTr)|ioviKfi èitexripiq xoO riavsTiiaTTiiiioo A0T]vò)v, 1965-66, 417ss.]. 1. Già da tempo (Kretschmer, Skutsch) si è cercato di connèttere la limitazione greca dell'accento indoeuropeo alla terzùltima sillaba con la stessa limitazione in latino;2 e in particolare osserva il Leumann (Lat. gr., 183s.; v. anche 189) che il latino si accosta3 all'eòlico (lèsbico) : (a) nell'eliminare completamente l'accento libero indoeuropeo; (b) nel limitare l'accento alla penùltima e terzùltima sillaba, eliminando l'accento sull'ultima (anche il Meillet, MSL., 20, 1918, 171 ammette qualche affinità). 1 Fermo restando quanto scrissi negli Studii Schiaffini nr. 24, 177, che cioè esiste un rapporto fra greco e latino nel passaggio da un accento musicale a un accento intensivo, mi domando ora se questa innovazione (per ovvii motivi di prestigio culturale) non sia partita dal greco, e precisamente dall'attico, in cui tale trapasso cominciò forse nella lingua volgare già alla fine del IV sèc. a. C. (v. Altheim, op. cit. nel testo 324 con rinvìi; un po' genèrico Schwyzer, Griech. gr., 390ss.). V. anche quanto dico più sotto delle differenze quantitative in greco (e in latino). 2 Questa limitazione dell'accento alla terzùltima sillaba non appare in nessun' altra lingua indoeuropea: né in cèltico, né in germànico, né in islavo, né in bàltico, né in vèdico (né in sànscrito). E se il Pisani (Storia delle lett. antiche dell' India2, Milano, 1959, 9) scrive "l'accento va posto [...] sulla terzultima" se la penùltima è breve, si tratta certo solo di una distrazione dell'illustre indòlogo; cfr. J. Wackernagel, Altind. gramm., I, (Gottinga), 1896, 296s. Nel sànscrito 'clàssico' l'accento risale fino alla QUARTULTIMA quando la penùltima e la terzùltima sono brevi. 3 Già Ateneo X, 425 A accostava l'accentazione lèsbica a quella latina; e forse per questo il latino era considerato dagli Antichi come un dialetto eòlico : cfr. Varrone framm. 295 p. 311 Funaioli e ora Leumann-Hofmann-Szantyr, Lat. gramm., II, 23* (alla fine del volume) e l'accurato studio di E. Gabba nella Miscellanea Rostagni (Torino, 1963), 188s. Oltre alla baritonesi pensàvano forse all' a per i| e al digamma conservato (uideó somigliava più a FISCÓV che a ìScbv, p. es. ; e fama somigliava più a (panò che a (plinti ; fàgus più a cpctyóg che a 8evi xpôrccp KaxaK>avo|iévr|v, aùxf)v èX,eu9epo7tpeJtv

è^À.à|iv|/£cov àSpavcôç è^aaSsvoOaav, Suvaxœç È7ti xô 9eo|ii|ir|xov àvayonévr)v oùk àno^eÎTtoucrav èauxfjç à v a v ô p i a xf)v 9eoei8fj K i v r ) a i v à X X ' àppe7iœç àcpopôkrav eiç

xriv ÙTtepoÛCTiov K a i ôuvajxojioiàv Sûvajiiv K a i xaûxr]ç eùcôva 8uva|ioeiSfj K a x à x ô èipiKxàv yivo(iévT]v K a i rcpôç |ièv aùxf]v cbç à p x i S û v a j i o v Suvaxrôç è7teaxpa|4iévr|v, Ttpàç 8è x à S s C x e p a SuvanoSôxcoç K a i 9eoei8côç

rcpoïoOaav

xrçv 8è xwv àyicov

èÇouauSv, xf|v ô n o x a y f ) xcôv 9eicov Kupioxrixcov K a i 5uvâ|iecûv, xï|v eûkoctuov K a i àaûnva|ioç, ôuvaxœç, 8ova|ioSôxa)ç ; à propos des è^ouCTÎai on a èÇouaiorcoiôç, êÇouaiapxia, è^ouoiaaxiKÔç. Denys a eu peut-être ici recours au rcapr|ynévov, c'est-à-dire à un procédé d'expression en partie tautologique, 9 pour suggérer l'idée que l'être qui assume la fonction imitatrice constitue déjà l'image d'une propriété de Dieu. 10 Ainsi, les Kupiôxr|xeç ont comme but de d'identifier à ce qui est ôvxcoç Kopiôxriç et Kupicoç ôv, les Suvàjxsiç à ce qui est 5uvap.orcoiôç et àpxiSûvajxoç Sôvaniç, les èÇoocriai à ce qui est èÇooaioTioiôç è^ouciapxia. En outre, les quatre derniers participes dans A et B se groupent en paires dont l'une dénote l'imitation comme aspiration et l'autre comme réalisation: on classera sous la première rubrique les participes èq>ie|iévr|v-è7i£(TXpa|A|iévr|v, àcpopc&aavè7reaxpan|iévr)v; sous la seconde, les participes Siarc^àxxouCTav (êauxf|v xe Kai xà |iex' aùxriv) -yivonévr|v, yivonévr|v- JtpoïoOaav (rcpôç xà Ssûxepa). Cette classification s'avère plausible par le fait que les participes des premières paires sont associés à des syntagmes ou des composés dont les éléments se conjoignent en sens inverse: xfjç ôvxœç Kupiôxr|xoç- xô Kupicoç ôv, Suva|iorcoiôv (Sûvajiiv)àpxi8ûva|xov (Suva^iv); par le fait aussi que les participes des secondes paires s'associent à des composés dont le second élément est -eiô-: àya9oei8c5ç avec 8ia9

10

Cf. R. Jakobson, Novejsaja russkaja poèzija (Praha, 1921), 43.

Denys écrit dans son De divinis nominibus, c. VIII, 4: èÇ aùifK (se. xf|Ç toO 0eoO ëuvàjiEcoç) eiatv ai 3soei8sîç TCÙV àyy6XiKcôv SUXKCKTUCÛV 8uvà(isiç- èt, aÛTfjç Kai tô eïvai ànsxajtTWTCOç ëxoutji Kai nàaaç aûmw xàç voepàç Kai à9avàxouç àeiKivriCTÎaç ... aÙTfjç ètpieicjriç aùxoïç xô SùvaaSai Kai xô slvai xaOxa, Kai è(piea9ai àei elvai, Kai aùxô xô Sùvao9ai Èv; xi|v sî)Kocr|iov ( + àaôn(pupxov) eùxaÇiav + TÔ xexaynévov xfjç Ö7repKoo|iiou ( + voepâç) èi;oucnôxT|xoç ; où xupavviKcôç ... djioKexptinévriç + àKpaxfixcoç ... àvayo|iévr|ç ( + àvayoùariç); Kai Ttpôç xf|v ... àcponoiounévriç + Kai xaûxr|v àva^anîtoûariç. On peut signaler aussi les correspondances: où xupavviKÔç èrci xà %eipco- àKpaxfixcoç êni xà 9eîa, C

Agreement between 2 and 3 5

Result 6

O —

marked

+

14 Cf., at least, his article "Zur Satzperspektive im modernen Englisch", Archiv für das Studium der modernen Sprachen und Literaturen, LXXXIV (1929), Bd. 155, 200-210. — Some Russian scholars use the term 'aktual'noe (smyslovoe) ölenenie' (B. A. Il'jis, K. G. Krusel'nickaja, I. P. Raspopov.) 15 See Hockett's article quoted in footnote 5. 14 Cf. J. Firbas, "From comparative word-order studies. (Thoughts on V. Mathesius' conception o f the word-order system in English compared with that of Czech)", Brno Studies in English, IV (1964), 111-128; "Non-thematic subjects in contemporary English", TLP, 2 (1966), 239ff.

506

FRANTISEK DANES

From the matrix we learn that the usual order on the semantic level has been changed under the impact of the functional needs of the contextual bipartition. The resulting utterance links up fully with the given context (and in this sense it is not marked, but normal), and yet, if valued from the point of view of the linguistic system (i.e., as found in an isolated sentence), its word order will be experienced as marked. 1.44. Thus, by means of our matrix we can define a 'neutral sentence' as one which has no 'minus' sign in the column 5. The linguistic interpretation of this statement is, clearly, that the functional needs of the T-C bipartition do not — in the case of the 'neutral sentence' — lead to the change (inversion) of a usual order. In general, the explanation of the notion of 'neutral order' ought to be sought in the interaction (interrelations) of patterns on the three different syntactic levels. In patterns, some of their elements (or all of them) may be positionally bound, by a strong, or by a weak rule. If, in the matrix of an utterance, all three levels are 'in agreement' (coordinated), such an utterance has a neutral word order. (The necessary condition is, of course, that at least one pattern contains a bound element. In utterances that do not fulfil this condition, the distinction between the neutral and the marked order is irrelevant.) The notion of 'marked word order', on the other hand, implies the solution of a conflict between levels. The solution involves the existence of a hierarchy of levels and of some specific linguistic devices. There are two hierarchies: (a) hierarchy of different orders: (1) strong rules (i.e., grammaticalized order and fixed order), (2) weak rules (usual order), (3) free rules (labile order); (b) hierarchy of levels: (1) T-C level, (2) semantic level, (3) grammatical level. — The means for solving conflicts are: (a) "inversion", in the case of weak rules; (b) sentence intonation; (c) particles, articles, lexical means, specific grammatical constructions; (d) selection of a different pattern. 1.5. At this moment we must answer two essential questions: (1) What happens when, in a matrix of the above type, a grammaticalized or fixed order appears? (2) Is the inversion C -» T also possible (and if so, what would be the consequences)? 1.51. Question (1) may be demonstrated by the English sentence "John hates Mary". Its matrix contains the grammatical pattern S => V => O with grammaticalized order. This utterance fits in with a consituation where the topic of the discourse is 'John'. But in a different consituation, where the topic would be 'Mary' and the Comment "the hatred of John for her", the order of sentence elements should be changed. Column 2 (actual sequence) in the respective matrix would be:

o

V

S

G

Ac

Ag

T

C

17 The components T-C coincide, in different utterances, with different sentence elements or groups of them. While the former are only two in number (and often without a precise dividing line), the latter rank from one to a theoretically unlimited number.

507

ORDER OF ELEMENTS A N D SENTENCE INTONATION

But the first (grammatical) line OVS is incompatible with the pattern S => V => O, showing the grammaticalized order. There are, generally, two ways of solving this problem: either to relinquish the possibility of rendering the T - C bipartition (in fact, there are differences in languages as to their 'sensitiveness' for contextual needs), or to use a different, more suitable grammatical construction. In the case of our English example, we can make use of the passive construction (in which words or phrases denoting Ag and G, respectively, are mutually replaced). The passive construction Mary is hated by John has the following matrix (in contrast to the active construction, the difference appears on the semantic level): PASSIVE

ACTIVE

S

Vpass

O

S

V

o

G

Ac

Ag

Ag

Ac

G

T

C

T

C

(Of course, we do not claim that in English the passive construction is the only possible solution of all cases. There are other means, such as the definite and indefinite article, sentence intonation, and others.) 1.52. In answering question (2), we shall consider the English sentence "John is writing to his father". If we take it as an answer to the question 18 "What is John doing?", we should assign the Topic value to "John' and the rest of the sentence would form the Comment of the utterance. This is the normal, neutral case, with T -* C order. Now, imagine the same sentence as an answer to another question, viz. "Who is writing to his father?" In this consituation, 'John' is felt as Comment and the rest of the sentence as the Topic of the utterance. Thus, the order of components would be inverse, viz. C - T , and the utterance experienced as marked, with emphatic coloring. The explanation of this shift must be sought for in the fact that in the respective matrix, in the last line of column 5, the sign 'minus' appears, due to the inversion of the usual order T -> C. At this point, the conclusion is not surprising, as it follows from our definition of 'neutrality'. But one must ask whether there is any formal device by means of which the shift in the contextual structure of the given sentence would be manifested (word order being, in this case, out of the question). In the domain of the WRITTEN language, no such device would be employed and the respective alternative of the contextual structure of the sentence should be inferred from the consituation only. The SPOKEN language, however, avails itself of a specific means for signalling the Comment of the utterance, viz. the sentence intonation. This assertion seems to be of great importance and in the next section of our paper we shall outline the principal regularities between sentence intonation and T - C bipartition of utterance. 19

The questions are used here in order to elicite a proper consituation.

508

FRANTI§EK DANES

2. The functional diapason of SENTENCE INTONATION (or, more correctly, of the intonation of utterance), as one of prosodic features, is very wide and, as has recently been very aptly pointed out by D. Crystal and R. Quirk,19 the expressions 'prosodic' and 'paralinguistic' denote "a scale which has at its 'most prosodic' end systems of features (for example, intonation contours) which can fairly easily be integrated with other aspects of linguistic structure, while at the 'most paralinguistic' end there are the features most obviously remote from the possibility of integration with the linguistic structure proper " The systemic functions of intonation contours are several,20 one of them being simply to signal the T-C structure of utterance. 2.11. It appears that in many languages (perhaps in most, but we dare not, for the present, make this a universally valid statement), the Comment of the utterance would be associated with the center (nucleus) of the (terminal) intonation contour.21 This means that in languages where, as a rule, the Comment is placed towards the end of the utterance (cf. Hockett's semiuniversal), the centre of the terminal intonation contour (CI) should be located on the last stress-unit of the utterance. E.g.: The train has come. John hates Mary. — German: Der Zug ist gekommen. — Russian: Prisel poezd. 2.12. But as we have mentioned above, the grammatical rules of the English word order do not fully allow rearrangement of the sentence constituents according to the needs of the consituation (i.e., in accordance with the T-C structure). Thus, in the sentence There were some pictures on the walls, the phrase 'on the walls' does not evidently belong to the Comment (the definite article signals the respective notion as already known); consequently, the CI would not be placed on the last stress-unit of the utterance, but on the last stress-unit of the Comment of the utterance, particularly on the word pictures. A contrastive comparison with languages having the so-called 'free' word order appears to be very illuminating. Let us consider the following example (cf. D. Worth, op. cit., p. 50): RUSSIAN

ENGLISH

in consituation (a):

Krovati stojali v jego kömnate

The beds were in his room

in consituation (b):

V jego komnate stojali krovati

There were beds in his room

Thus in languages like English, the CI would be often placed on non-terminal elements of the utterance, while in others, as for instance in Slavic languages (conspicuously in Czech), the normal suprasegmental (prosodic) shape of the utterance (i.e., of the non19

D. Crystal and R. Quirk, Systems of Prosodic and Paralinguistic Features in English (The Hague, 1964), 12. 20 Cf. F. Danes, "Sentence intonation from a functional point of view", Word, XVI (1960), 34-54. " Some authors call this center 'the sentence stress' or 'the logical stress'.

ORDER OF ELEMENTS A N D SENTENCE INTONATION

509

emphatic utterance in which no word is emphasized for contrast) is that with the CI in the last stress-unit. The supposition that the English sentence is only to a small degree sensitive to the needs of the context turns out to be not fully valid; the differences between languages lie mainly in the different means that are employed (as has been ingeniously pointed out by J. Firbas in connection with the use of articles). I will adduce some other examples, in which the Comment occupies a position very remote from the end of the utterance (so that a relatively very long terminal intonation contour arises): (1) "[They went to the river, but the water was so swift that the monkey was afraid. 'Get on my neck', said the elephant, 'I shall carry you...] I am not afraid to swim across a swift river.'" 22 (The phrase 'to swim across a swift river', although standing at the end, conveys the notion which has been explicitly mentioned in the previous context and consequently belongs to the Topic, not to the Comment, of the utterance). — (2) "[They've been cleaning them up the whole morning.] I've never seen such energy." 23 (The notion 'such energy' is implied in the preceding utterance.) 2.13. To sum up: While in Slavonic (esp. in Czech) the variability of word order is compensated for by a rather uniform (automatic) location of the terminal CI, in English, on the other hand, the highly fixed word order is compensated for by a great variety of the possible positions of the CI in the utterance. In other words: in English it is rather the suprasegmental phonological structure that signals the 'functional perspective of utterance', i.e., the points of the highest communicative dynamism. Thus, we may conclude that the functional load of the two linguistic devices is different in various languages. (It is worth mentioning that the relatively high functional load of intonation in English shows itself in relation to other linguistic functions as well.24) 2.2. Utterances conveying emphasis are governed by special rules. At least two classes of emphatic utterances should be distinguished: (1) emphatic utterances proper, (2) utterances with emphasis for contrast. In class (1) the emphatic feature characterizes the utterance as a whole, while in class (2) this feature is associated with one particular element of the utterance only. 2.21. The emphatic utterances proper are characterized by the inverse order on the contextual level, i.e., by the order C-T, and, consequently, by the onset position of the CI (this being located on the initial stress-unit of the utterance) on the suprasegmental phonological level. E.g.: The train has come! (in contradistinction to 'normal' The train has come). In Russian: Poezd prisel! X Poezd prisel; and also (due to the 'free' word order) Prisel poezd! X Prisel poezd. In German: Der Ziig ist gekommen! X Der Zug ist gekommen. 22

Adopted from W. Jassem's book Fonetyka fezyka Angielskego (Warszawa, 1954). Adopted from W. R. Lee's An English Intonation Reader (London, 1960). 24 Cf., e.g., the following statement of M. A. K. Halliday, A. Mcintosh, P. Strevens: "... it is important to realize that spoken English makes extensive use of intonation to carry grammatical meaning" (The Linguistic Sciences and Language Teaching, London, 1965, 53). 23

510

FRANTISEK DANE§

2.22. The emphasis for contrast is intended "to show that a word is contrasted with another (either implied or previously expressed), or that a word introduces a new and unexpected idea". 25 D. L. Bolinger has called such utterances 'sentences of the second instance'. 26 Contrastive emphasis may be rendered — according to circumstances — by a set of means: (a) word order, (b) a shift of the CI (it would be displaced from its 'automatic' (neutral) position), (c) a specific phonological form of the intonation contour. It is clear that in English the possibilities (b) and especially (c) are the most common. E.g. : I sent a book to hër (and not to someone else), the contrast having been achieved by means of the shift of the CI. (The normal form would be I sent a book to her, according to the rule that in sentences ending with a preposition and a pronoun, the final pronouns are not stressed.) In addition to it, the contrast may be pointed out or modified by means of special intonation contours. — Other examples : [Look out, here comes a cär.] It looks like oür car. — What are you going to do? — It was extremely cold this year. — I nearly did forget it (the emphasis is foregrounded by means of the construction with 'do' as well). — John loves Mary (and not George) = It is John who loves Mary (with a special construction). — I saw the man coming along the road (reassuring the person spoken to ; a very long intonation contour). — I met her father (and not her mother or brother...); in this case the contrastive emphasis is rendered by means of a specific (emphatic) contour only, as the contrasting word is in the final position, so that the placement of the CI cannot function as a distinctive feature. Many analogous examples might be easily adduced from other languages. German : Ihr Bruder hat länge auf Sie gewartet. // Ihr Bruder hat lange auf Sie gewartet. / Ihr Brüder hat lange auf Sie gewartet. (The normal form : Ihr Bruder hat lange auf Sie gewärtet.27) Some Russian examples may be found in the book by J. E. J. Buning and C. H. van Schooneveld,28 Czech examples in the book on Czech sentence intonation by the present author. 29 It is exactly this contrastive emphasis that is often referred to as 'logical (sentence) stress'. 2.23. In addition, it is possible to use a specific non-terminal intonation contour for singling out the Topic of the utterance. E.g., in English: My brother | went to Brighton (implying: 'as for my brother, he...'). If it succeeds | I shall make... ; in Russian: Esli tebe grüstno |, ...; in German: Geheime Sorgen | sind eine schwere Läst. 2.3. It has often been suggested that sentence intonation performs proper grammatical functions and lexical (semantic) functions as well. Undoubtedly, such examples, as: 26

D. Jones, An Outline of English Phonetics (London, 1956), 277. D. L. Bolinger, "Linear Modification", PMLA, 1952, 1117ff. 17 Cf. O. von Essen, Grundzüge der Hochdeutschen Satzintonation (Ratingen/Düsseldorf, 1956). 29 J. E. Jürgens Buning and C. H. van Schooneveld, The Sentence Intonation of Contemporary Standard Russian as a Linguistic Structure (The Hague, 1961). Cf. also C. L. Ebeling "Subject and Predicate, especially in Russian", in Dutch Contributions to the Fourth International Congress of Slavists (The Hague, 1958). 2 " F. DaneS, Intonace a vita ve spisovné destiné [The Sentence Intonation in Standard Czech] (Praha, 1957). 26

ORDER OF ELEMENTS AND SENTENCE INTONATION

(1) (a) (b) (2) (a) (b) (3) (a) (b)

511

I didn't visit the doctor | because I was ill. / didn't visit the doctor because I was ill.30 Please wire | if I am to come, Please wire if I am to come. It is the country | that suits my wife best, It is the country that suits my wife best.31

evidently show the relevance of the suprasegmental phonological structure for the grammatical and semantic interpretation of such sentences. Nevertheless, it seems to us that the different grammatical and semantic interpretations are based on two different underlying T - C structures of the respective utterances. In other words: rather than saying that the intonation here works as a grammatical device (distinguishing, e.g., and object clause from an adverbial one, or determining the function of the conjunction), we should rather say that this is an accidental effect of two possible T - C structures of the given utterance. (E.g., in (2a) the 'if'-clause is marked as Comment, whereas in (2b) it is the verb 'wire' that is signalled as an emphatic Comment, while the 'if'-clause conveys a thing that is already known from the consituation.) An analogous situation may be found in cases where sentence intonation seems to determine the semantic meaning of a word. E.g.: (4) (a) He also v / j i W Prague. ('He visited some other places and Prague additionally') (b) He also visited Prague. ('Others visited it, and he did so as well') Fairly analogous examples may be adduced from other languages. But in all of them we shall find that the two different semantic interpretations of also (German auch, Russian toie, Czech tez, etc.) are determined by the fact that in (4a) the adverb brings out the subsequent utterance portion as a new fact (Comment), in addition to facts already known, whereas in (4b), with also bearing the CI, the situation is inverted. Negative clauses seem to be a field especially favorable for the 'semantic intonation' in many languages. Cf. the following Czech utterances and their English counterparts: (5) (a) S kaidym nemluvi. — He does not speak to anybody (with a falling-rising contour) (b) S kaidym nemluvi. — He does not speak to anybody (with a falling contour). The utterances (a) have the meaning 'he speaks only to some people', while (b) means, in an appropriate situation, 'he speaks to no one'. It is worth mentioning, however, that the meaning (a) is, in Czech, associated with the normal (automatic) form of the utterance, but the meaning (b) is signalled by de-automation, while in English the situation is reversed (it is the meaning (a) that is associated with a marked contour). Nevertheless, both languages have another feature in common: the semantic differenso S1

Cf. W. R. Lee, op. cit. Cf. M. Schubiger, The Role of Intonation in Spoken English (Cambridge, 1935).

512

FRANTISEK DANE§

ces (a) and (b), rendered by intonation, depend, to a certain extent, on a favorable consituation; the respective intonational forms alone do not fully ensure that the utterance will be interpreted only in one of the two possible meanings, or, in other words, such utterances are not entirely unambiguous. But, at the same time, both languages have a pronoun, namely nikdo, no one {nobody), which, in contrast to kaidy, anybody, is unambiguous, having the negative meaning (a) only (in accordance with the presence of the explicitly negative morpheme ni- and no, respectively).32 3. It is self-evident that any systemic linguistic description, whether generative, or not, has to take into account all relevant facts about the most elementary and common linguistic devices, viz., the order of elements and the utterance intonation.33 And it seems to me that the generative scheme suggested by the M.I.T. group is, in its recent form, scarcely able to account for them in a satisfactory way, mainly because its phonological component does not include the sentence intonation; besides, the position of the T-C organization, the systemic character of which is hardly to be denied, is yet to be stated. (Its considerable stylistic import does not, in essence, contradict the undeniable appurtenance of the T-C principle to la langue.)3i CZECHOSLOVAK ACADEMY OF SCIENCES PRAGUE

®a Cf. J. Vachek, "Obecny zapor v anglictine a v cestine" (with an extensive English summary "Universal Negation in English and Czech"), Prague Studies in English, VI (Praha, 1947), 7-73. — E. S. Klima, however, does not mention these facts in his paper "Negation in English" (in: The Structure of Language, Englewood Cliffs, 1964, 246-323). — A more detailed analysis of such negative clauses in Czech may be found in F. Danes, "Pfispevek k rozboru vyznamove vystavby vypov£di [Towards the analysis of the semantic structure of utterance]", Studie a prace lingvisticke, I (Praha, 1954), 215 n. "3 Cf. Halliday, Mcintosh, Strevens, op. cit., 73. " Some remarks in Chomsky's book Aspects of the Theory of Syntax (Cambridge, Mass., 1965) as well as a personal communication by the same author seem to inaugurate a possible attempt to incorporate the said facts into the new generative framework. As regards sentence intonation, an earlier study by R. P. Stockwell, "The place of intonation in a generative grammar of English", Language, 36 (1960), 360-367, contains some valuable suggestions.

PETAR D I N E K O V

A TRAVERS LES ÉCHANGES LITTÉRAIRES BULGARO-POLONAIS: PENCO SLAVEJKOV ET A D A M M I C K I E W I C Z

L

A littérature bulgare connaît bien peu de jeunes écrivains possédant comme Penco Slavejkov (1866-1912) au début de sa carrière une culture aussi riche et •à aussi variée. Après avoir reçu une excellente éducation au sein de sa famille (son père est le célèbre écrivain Petko R. Slavejkov), le jeune homme enrichit encore sa culture à Leipzig et plus tard, dans le cercle des écrivains les plus éminents gravitant autour de la revue "Mis'l". Pendant son séjour à Leipzig (1892-1898), Penco Slavejkov fréquente assidûment les cours des plus grands savants et philosophes allemands et ses nombreuses lectures portent non pas tant sur les œuvres littéraires contemporaines en vogue, que sur les grandes œuvres classiques. Et bien qu'il affectât, sur le tard de sa vie, de prétendre avoir gaspillé son temps en Allemagne, c'est sans contredit à Leipzig que Penôo Slavejkov avait posé les fondements de son immense culture et de cette érudition remarquable qui se manifestent tant dans ses articles, dans ses traductions, dans les anthologies qu'il compose, que dans l'ensemble de son œuvre poétique. Penco Slavejkov est un type bien déterminé d'écrivain dans la littérature bulgare, essentiellement cérébral et cultivé. Il suffit pour s'en convaincre d'analyser ses nombreux articles traitant de questions littéraires ou culturelles, où brillent l'ampleur de ses conceptions et de ses vues, la profondeur et la mobilité de sa pensée. Cet 'intellectualisme' pénètre la trame même de son œuvre lyrique; Penôo Slavejkov introduit dans la poésie bulgare des thèmes et des problèmes nouveaux et il s'efforce d'orienter la pensée créatrice bulgare vers les grandes recherches idéologiques de la littérature européenne et mondiale, car il conçoit aussi sa grande mission de créateur et de penseur. C'est dans cette lumière qu'il faut envisager l'attitude de Penôo Slavejkov envers Adam Mickiewicz, tant sous le rapport de sa profonde admiration pour l'œuvre du grand poète polonais, que sous celui des influences de celle-ci sur la poésie de Penôo Slavejkov. Il est plus que probable que Penco Slavejkov avait eu des connaissances fragmentaires de l'œuvre de Mickiewicz avant son départ pour Leipzig, pour avoir sans doute lu des extraits de cette œuvre parue en traduction dans la Chrestomathie bulgare,

514

PETAR DINEKOV

ouvrage fort populaire à cette époque (1884) et composé par Iv. Vazov et K. Velickov, qui contenait des extraits de Pan Tadeusz et de Konrad Wallenrod, ainsi que quelques sonnets tirés du recueil Sonety Krymskie (tous traduits du russe par Iv. Vazov). Cinq années plus tard, le recueil composé par St. Kostov et D. Misev, XpucmoMamun no U3ynaeane na CAoeecmcmma (1889), publie des nouveaux extraits de Pan

Tadeusz

et de Konrad Wallenrod, suivis d'un extrait de Dziady (traduit du polonais par Chr. Kessyakov). Paraissent également tout un ensemble d'articles et d'essais consacrés à l'œuvre de Mickiewicz. Il est fort possible que Penco Slavejkov ait également eu connaissance à cette époque de certaines traductions en russe des œuvres du poète polonais, qui lui eussent permis d'avoir une idée d'ensemble de son œuvre, mais il n'en reste pas moins que c'est à Leipzig, plusieurs années plus tard, qu'il pénètre toute l'ampleur de la poésie et l'œuvre créatrice profondément originale de Mickiewicz. Les nombreaux témoignages de contemporains le démontrent éloquemment. Jord. Marinopolski, condisciple de Penco Slavejkov à l'Université de Leipzig, écrit à ce sujet: "A cette époque, Penco s'exaltait sur la poésie de Mickiewicz, qu'il comparait à Homère, il lisait et relisait Pan Tadeusz, opposant son auteur aux poètes allemands." 1 Un autre condisciple de Penco Slavejkov, Velilco Jordanov, souligne également le vif intérêt de son ami pour les œuvres de Mickiewicz: Chacun pouvait voir que Penco ne fréquentait pas l'Université dans le seul but d'y faire des études systématiques, comme tous les autres étudiants, pour apprendre une profession. Il cherchait au contraire à rénover son âme en la pénétrant de la poésie et de la culture allemande, de la littérature allemande, pour se préparer à atteindre les hautes cimes où se tiennent les grands maîtres de la littérature — Goethe, Schiller, Heine, et certains autres. Ayant déjà assimilé l'esprit créateur slave, dans les œuvres des classiques russes, dans celles de Mickiewicz et de Slowacki, et il devait maintenant accéder aux sommets du Walhalla germanique.2

A Leipzig, Slavejkov fréquente non seulement ses condisciples allemands, mais aussi les étudiants venus de tous les pays slaves, qui fraternisent bien vite et constituent l'Association des Etudiants slaves, dont il fut plusieurs années durant le président. Là il eut tout le loisir de connaître à fond l'œuvre de Mickiewicz soit directement, soit dans les cours de ses professeurs et par les traductions allemandes du poète polonais. Penôo Slavejkov avait d'ailleurs appris assez de polonais pour lire les œuvres littéraires polonaises dans l'original. Il avait suivi les cours de slavistique du professeur A. Leskien et avait choisi pour thèse scientifique Heine en Russie. Il avait à ce propos l'intention de passer un semestre à Moscou pour préparer cette thèse. L'intérêt de Penco Slavejkov envers l'œuvre de Mickiewicz continuait d'être aussi vif après son retour de Leipzig. Un éloquent témoignage de cette constante admiration se retrouve dans le travail de Penco Slavejkov sur son poème "Kï>pBaBa n e c e H " ("La chanson sanglante"). Sans prétendre épuiser toutes les données aui abondent dans 1

Ilemo C/iaeeÛKoe, 17. K. fleopos, II. K). Todopoe e cnoMemime m cbspeMeHHUifume eu (Sofia,

1963), 53.

« Ibidem, 48.

LES ÉCHANGES LITTÉRAIRES BULGARO-POLONAIS

515

ce sens, nous nous bornerons de citer certains moments. En 1909 le poète, qui est de passage à Varsovie, fait part de ses impressions sur le monument de Mickiewicz dans une lettre adressée à Mara Belceva : Mickiewicz! J'ai réglé le cocher du fiacre et me suis campé devant le colosse de bronze. Tu l'as bien vu sur la carte postale, n'est-ce pas? Je n'aime pas ces grands monuments, surtout lorsqu'ils ont ce style romantique pompeux. Ils l'ont dressé, Dieu sait pourquoi, à cette hauteur, pour qu'on n'aperçoive pas son visage; ils l'ont dressé ainsi dans le seul but de permettre aux somptueuses draperies de son manteau de fourrure de bien se déployer. Ce genre de monuments sert de prétexte aux sculpteurs de faire l'étalage de leur habileté dans l'art de traiter les accessoires, tandis que l'homme perdu dans cette profusion de détails inutiles reste au second plan. C'est là le monument du manteau de fourrure d'un grand poète. Et en fait, les Polonais (même à présent) ne voient en Mickiewicz que son manteau, c'est-àdire son revêtement politique. Ils voient ce qu'il y avaint en lui de temporaire, et non l'éternel, non pas le poète et l'homme, non pas le represéntant du génie polonais devant l'humanité, mais un des tambours-majors de l'armée des combattants pour la liberté et la conscience politique. Un jour de l'année, un mardi ou un vendredi quelconques! Pense un peu que jusqu'à présent dans toute la littérature polonaise il n'y a pas de vraie 'ästhetische und menschliche Würdigung' de cet homme qui a enduré toutes les souffrances, de ce chantre de Dieu en l'homme! Je suis resté longtemps en contemplation devant ce titan à l'âme si fière vaincue par la vie, que le sculpteur n'a pas su rendre (ou moins c'est ce qui me semble), et j'ai éprouvé devant lui, ce justicier de la vie, le désir de dire moi-aussi les paroles que j'avais au cœur. Sitôt rentré à l'hôtel, j'ai écrit une ode que je lui dédie et que je te lirai, une fois corrigée, dès que je serai de retour. Il en est sorti une espèce de profession de foi, tout le contraire de ce que je devais écrire quelques jours plus tard à Moscou devant le monument de Gogol' ("Bogoborec")! Comme tu le vois, je n'ai que poésie dans l'âme, et tristesse de tout ce que je vois et j'entends. Dans l'une de ces odes je m'écrie "résigne-toi!", et dans l'autre "lutte!". Telles étaient mes impressions sur place et peu m'en chaut que dans ce cas précis il y ait des contradictions entre elles (ou en moi).3 Ces paroles contiennent une confession des conceptions littéraires et esthétiques de Penco Slavejkov et montrent tout aussi bien son attitude envers Mickiewicz. Une indication particulièrement intéressante est contenue dans l'allusion à une ode dédiée au poète polonais composée à Varsovie, qu'il caractérise en l'opposant au poème "Eoroôopeu" ("Impie"), composé quelques jours plus tard à Moscou devant le monument de Gogol'. Dans l'œuvre littéraire de Penöo Slavejkov il n'existe pas de poème directement lié au nom de Mickiewicz. Or, il est parfaitement improbable que le poète ait détruit cette ode écrite à Varsovie. Considérant le fait que le poème "Eoroôopeu" est compris dans le recueil Ocmpoebm na ÔAaMcenume ("L'Ile des Bienheureux"), publié sous un nom d'emprunt (Ivo Dolja) dans un but de mystification, il serait logique d'y retrouver aussi l'ode à Mickiewicz. C'est ce qu'a très judicieusement fait Angel Todorov, qui voit dans le poème ">Kpeu Ha acHBOTa" (Le prêtre de la Vie), dont le manuscrit est daté du 7 mai 1909 à Varsovie, 4 cette ode dédiée au poète polonais. Ce poème est compris dans le cycle d'Ivo Dolja. Dans la caractéristique de l'œuvre d'Ivo Dolja, l'on fait à maintes reprises ressortir que sa 3

rieHHO OiaBeÄKOB,

4

n e m o CjiaBefiKOB,

Cbôpauu cmuiienun, 8 (Sofia, 1959), Cbôpauu cwuHenun, 2 (Sofia, 1958),

118-119. 318.

516

PETAR DINEKOV

particularité la plus saillante en est le caractère contradictoire. Par sa conception, le poème ">Kpeii Ha acHBOTa" se rapproche des traits caractéristiques de l'ode de Mickiewicz fournis par Penco Slavejkov lui-même dans la lettre citée, qui oppose cette pièce au poème "BoroGopeii", par lequel débute le cycle d'Ivo Dolja. Le nom d'Adam Mickiewicz revient à plusieurs reprises encore dans la correspondance de Penco Slavejkov. Citons plus particulièrement une lettre adressée de Rome à l'éditeur Al. Paskalev, en date du 9 décembre 1911: La traduction de "Pan Tadeusz" que tu m'as envoyée ne vaut pas un liard. Celui qui l'a faite ne connaît même pas la langue bulgare ordinaire, tandis que du polonais, il ne connaît guère que la langue courante, et n'a aucune notion de la langue poétique, avec ses tournures particulières, ses arabesques et ses nuances. Et par conséquent, il n'y a rien à en tirer.5 De quelle traduction s'agissait-il? Cette question n'est pas encore élucidée. Il est fort possible que Penco Slavejkov ait visé la traduction de Xr. Kesjakov, qui travaillait depuis assez longtemps déjà sur une traduction du texte intégral de Pan Tadeusz à partir de l'original polonais. Il en avait publié quelques extraits en 1889, alors que le texte entier fut présenté au public en 1918-1923 et 1934.6 Bien que Kesjakov eût travaillé à partir de l'original, sa traduction, tout en étant sous bien des rapports meilleure que celle de E. Karanov (publiée en 1901), n'en possédait pas moins les imperfections indiquées dans la lettre de Penco Slavejkov. L'intérêt de Penco Slavejkov envers l'œuvre de Mickiewicz est corroboré par de nombreuses données tirées des archives du poète bulgare. Penco Slavejkov avait, en outre, une connaissance approfondie de toute la poésie de Mickiewicz et ceci ressort parfaitement à la lecture du long article consacré à l'œuvre du poète polonais, publié en 1899, ainsi que d'une analyse de la genèse du poème "KtpBaBa neceH" ("Chanson sanglante"). L'article de Slavejkov fut publié à l'occasion du 100e anniversaire de la naissance du grand poète polonais, commémorée par la presse bulgare. Bien qu'avec un certain retard, la revue littéraire MUO>A, la plus importante à l'époque, rend un vibrant hommage à Mickiewicz et publie avec l'article de Slavejkov cinq extraits de l'œuvre du grand poète, traduits par Kiril Xristov. Une analyse de cet article nous montre que Slavejkov connaissait à fond l'œuvre de Mickiewicz qu'il tenait en très haute estime. Slavejkov détermine la place de Mickiewicz non seulement dans le développement de la littérature polonaise, mais à l'échelle de la littérature mondiale. Il le place à l'égal de Dante, Homère, Goethe, Byron et Tolstoj et écrit: "Pour les Polonais, les œuvres de Mickiewicz sont la Bible." Si "l'antique Hellade, disparue depuis des siècles, vit encore dans les œuvres géniales de ses fils, surtout dans l'Iliade", on pourrait dire de même pour les Polonais, "que s'ils venaient à disparaître comme les anciens Hellènes de la vue spirituelle de l'humanité, leur vie, la vie d'un grand peuple, subsistera dans les œuvres de ses 6

IleHTO OiaBeîiKOB, Cbôpanu cbumenua, 8 (Sofia, 1959), 248. Cf. IL XlHHeKOB, "Bi.jirapcKHTe npeBOflH Ha 'IlaH Taaeyin'", E3UK0eedcK0-emH0zpaificKu u3CAedeanuH e nauem ?ia anad. Cm. POMOHCKU (Sofia, 1960), 917-921. 6

LES ÉCHANGES LITTÉRAIRES BULGARO-POLONAIS

517

créateurs de génie, et surtout dans 'Pan Tadeusz'." Penco Slavejkov souligne que "Mickiewicz n'est pas seulement un grand poète, il est un héros de tout un peuple. Sa vie, comme son œuvre, sont l'incarnation des idéaux d'un peuple, condamné par une cruelle destinée à vivre avec ses seuls idéaux." Plus loin il écrit: "Nul des grands poètes mondiaux, sauf Dante, ne pourrait à plus juste titre dire de soi-même: Mon âme est incarnée dans mon pays natal — dans mon corps palpute l'âme de ma Patrie. Mon pays natal et moi sommes un tout!" Partant de cette conception du caractère profondément populaire de l'oeuvre de Mickiewicz et de son immense portée pour la nation, Penôo Slavejkov s'arrête avant tout sur les deux œuvres maîtresses du poète polonais, Dziady et Pan Tadeusz. Slavejkov connaissait fort bien également l'œuvre lyrique de Mickiewicz, ainsi que ses merveilleuses ballades, mais il était guidé par une prédilection personnelle pour les grandes compositions épico-dramatiques de Mickiewicz, en raison du fait qu'il travaillait lui-même à cette époque sur le poème "Ki>pBaBa neceH" ("Chanson sanglante"). Slavejkov se livre à une analyse minutieuse autant que passionnée de ces deux grandes fresques du poète polonais ; il exalte la grandeur et la puissance des personnages et cherche les superlatifs de leur comparaison : pour Dziady — le Faust de Goethe et le Manfred de Byron, pour Pan Tadeusz — Y Iliade d'Homère, La Divine Comédie de Dante, et La Guerre et la Paix de Tolstoj. En parlant de Pan Tadeusz, Slavejkov répète les paroles de G. Brandès, qui estime que c'est le seul grand poème épique créé au XIXe s. Slavejkov ajoute: "On pourrait penser que l'on trouverait difficilement des thèmes épiques, dans le sens caractéristique de ce terme, dans la vie des peuples européens d'aujourd'hui, et qu'il n'y a plus de poètes possédant assez de nerf pour entreprendre de recréer l'ensemble de la vie culturelle de tout un peuple, avec tous ces détails qui justement caractérisent et rendent intéressante et vivante une telle recréation. Au cours de notre siècle, seuls deux poètes ont réussi: Mickiewicz avec 'Pan Tadeusz' et Lev Tolstoj avec 'La Guerre et la Paix'." Lors de l'analyse de Dziady, Slavejkov s'emploie surtout à faire ressortir le caractère romantique de cette œuvre, le vol infini de l'imagination et la profondeur des sentiments du poète ("Les cris d'amour de Pétrarque, de Heine et de Musset pâlissent devant l'incoercible embrasement et la sublime extase d'amour d'un Mickiewicz"), la force extraordinaire de l'inspiration poétique, qui se manifeste au plus haut degré dans l'improvisation. Slavejkov incline à expliquer ces particularités par "le système nerveux d'une impressionnabilité presque douloureuse" du poète polonais. Analysant plus loin le poème Pan Tadeusz, en qui il voit le chef-d'œuvre de la poésie polonaise, Slavejkov insiste surtout sur le caractère épique de l'œuvre, sa puissance expressive et sa composition. Cette dernière question l'intéresse plus particulièrement (il l'étudié également à propos de Dziady), car elle est en étroite relation avec son propre travail sur le poème " K i p B a B a n e c e H " ("Chanson sanglante"). Par opposition à Dziady, Slavejkov souligne la simplicité de la forme et le naturel de Pan Tadeusz, la force du sentiment. Slavejkov estime que c'est là "la seule œuvre concise et rigoureuse" de toute la littérature polonaise de la première moitié du siècle passé et qu'à ce titre,

518

PETAR DINEKOV

"Mickiewicz se présente comme le digne représentant du génie polonais devant l'humanité toute entière". En s'efforçant de mettre en relief les particularités littéraires les plus caractéristiques de Dziady et de Pan Tadeusz, Slavejkov résolvait les problèmes essentiels liés à sa propre création. Il avait abouti à ces conclusions non seulement par son analyse et sa connaissance de ces deux grandes œuvres de la littérature polonaise, mais aussi en étudiant les ouvrages de Brandès et de son professeur à l'Université de Leipzig, le professeur Volkelt (il cite d'ailleurs ces deux auteurs dans son article). Ce vif intérêt envers l'œuvre de Mickiewicz si clairement et si largement manifesté par Penôo Slavejkov nous incite tout naturellement à rechercher les liens créateurs qui ont pu s'établir entre les deux poètes, qu'il s'agisse de rapports d'un caractère typologique, soit d'une influence directe de l'œuvre du grand poète polonais sur Slavejkov. La première question n'est pas encore étudiée. On pourrait établir un parallèle typologique entre les deux auteurs sous le rapport de l'utilisation du folklore dans leurs œuvres. C'est là un trait qui très certainement rapproche les deux poètes et constitue un fait important dans l'histoire des littératures polonaise et bulgare, mais il importe cependant de bien faire ressortir que ces deux auteurs mettent à contribution la création poétique populaire au sein d'une ambiance historique fort différente, partant de positions et de conceptions différentes et poursuivant jusqu'à un certain point, des desseins purement littéraires foncièrement différents. Au point de vue historico-littéraire, les influences de Pan Tadeusz sur le poème de Slavejkov "RtpBaBa neceH" présentent un bien plus faible intérêt. L'existence de telles influences a été établie par tout un ensemble d'auteurs (A. Jensen, Malco Nikolov, Iv. Lekov, Wl. Bobek, etc.).' Mais en ce qui concerne le caractère de cette influence, il y a bien des choses non encore élucidées, qui nécessitent des lumières et des interprétation nouvelles. Pour déterminer les rapports existant entre Pan Tadeusz et "KtpBaBa neceH" nous possédons une base solide, constituée par l'article de Slavejkov sur Mickiewicz, dans lequel l'analyse de Pan Tadeusz tient la plus grande place. Il existe, en outre, toute une série de preuves directes corroborant le fait que Penôo Slavejkov s'était inspiré de ce poème lors de la création de "KipBaBa neceH". Dans un de ses cahiers de notes tenu en relation avec son travail sur "KipBaBa neceH", où Slavejkov indiquait fort souvent ses sources, on peut lire la remarque suivante: "Cela, il faudrait essayer de l'améliorer, peut être par une interruption ou en l'abrégeant. Revoir ce passage chez P[an] T[adeusz], p. 136, vers 590-604."8 Lorsqu'on parle des affinités existant entre "KtpBaBa neceH" et Pan Tadeusz, il importe avant toute chose de rejeter toute pensée de dépendance servile ou de plate ' M a j i i o HHKOJIOB, " ' I ï a H T a i t e y i n ' H ' K i p B a B a n e c e H ' " , ynu/iuufeH npeeAed, 37 (1928), 376-393; A . MeHceH, ' " I l a H Tafleym' H 'ICBPBABA neceH'", Jlucmonad, X I I I (1931), 11-17; HB. JleicoB, "IleHHO CNABESKOB H AAAM MHHKCBHH", Jlumepamypen ZMC, I V (1932), 159; WL. Bobek, "Kilka

uwag o symbolicznym znaczeniu gtôwnych postaci w 'Krwawej piesni' P. Slaveikova", CSopuw e necm m npotf). JT. Mujiemun (1933), 645-648. 8 Cf. les notes accompagnant "KbpBaBa neceH" sous la rédaction de P. Dinekov (1949), 329. Dans les archives de Penco Slavejkov il existe d'autres notes analogue également.

LES ÉCHANGES LITTÉRAIRES BULGARO-POLONAIS

519

imitation de la part de Slavejkov. Entre ces deux œuvres, il existe des différences fondamentales que tous les chercheurs (et surtout A. Jensen et Wl. Bobek) ont fait ressortir de façon on ne peut plus catégorique. Ces différences résident non seulement dans les détails, dans l'ampleur et la portée du sujet, dans les caractères des principaux personnages ou les descriptions du genre de vie et de la nature, mais aussi et surtout dans l'ambiance spirituelle et la trame littéraire de ces deux poèmes. Le critique littéraire polonais Wl. Bobek fait fort judicieusement remarquer que les personnages de "Ki>pBaBa neceH" ne sont pas seulement "des créations littéraires prenant part à l'action du poème, mais ils sont porteurs de certaines idées et possèdent à ce titre la signification de symboles" ... "Il est notoire que Slavejkov a écrit sa 'KtpBaBa neceH' sous l'impression et l'influence de 'Pan Tadeusz'. Mais dans l'épopée de Mickiewicz aucun des personnages n'est revêtu de cette signification symbolique; ce sont des personnages vivant d'une intense réalité artistique, ne donnant prise à nulle interprétation symbolique" (loc. cit., 645). Bien que les deux poèmes soient également inspirés par un profond amour de la Patrie, consacrés aux destinées des deux pays et saturés de pensées sur le sort tragique du peuple, ils n'en diffèrent pas moins essentiellement dans sous le rapport de la conception littéraire, que sous celui du sujet. Pan Tadeusz s'attache à décrire avant toute chose la vie quotidienne du Polonais, dans toute sa diversité et toute sa richesse folklorique, dans ce qu'elle a de sublime et d'humble à la fois, avec une vitalité profondément contagieuse. D'où sa simplicité, son naturel, le charme souverain de l'expression littéraire. Slavejkov construit son œuvre sur un drame national avec des touches drues et puissantes: deux événements historiques marquants (l'Insurrection d'Avril 1876 et la guerre de Libération de 1877-1878) lui servent de toile de fond pour brosser les destinées historiques du peuple. Les divers événements, les personnages principaux et même les tableaux de la nature sont mués en symboles de ces destinées. De là le caractère très particulier de l'expression — modelée très souvent sur une base contemplative, et partant, dénuée de relief vital et de couleurs éclatantes. Pour toutes ces raisons les deux poèmes dans leur essence-même revêtent une physionomie littéraire foncièrement différente. Il est bien évident que sous le rapport de la physionomie littéraire, il est très difficile de trouver des affinités entre " K t p B a B a neceH" et Pan Tadeusz. Wl. Bobek cherche des ressemblances entre "KbpBaBa neceH" et d'autres œuvres de la littérature romantique polonaise, notamment Krôl-Duch et Anhelli de J. Slowacki. Mais il estime que ces affinités ne sauraient provenir d'une influence directe, car selon toute probabilité Slavejkov ne connaissait pas Slowacki. Personnellement je pense aussi que Slavejkov ne devait pas avoir connu de près l'œuvre de Slowacki. Je serai plutôt enclin, compte tenu de la fonction du symbolique, de chercher une telle influence dans une autre œuvre de Mickiewicz, Dziady, que Slavejkov connaissait parfaitement vien. Bien que le poète bulgare ait relevé les sérieuses imperfections de la composition du poème, il l'appréciait hautement: "Malgré ce défaut, ce poème occupe grâce à ses grandes qualités poétiques, une place de choix parmi toutes ces œuvres poétiques où

520

PETAR DINEKOV

l'on sent palpiter le sourire sublime du génie. 'Dziady' s'apparente au 'Faust' de Goethe et au 'Manfred' de Byron." Slavejkov écrit plus loin: "Je renonce à résumer la troisième partie de 'Dziady', car je sais qu'on ne résume pas de telles œuvres. J'envie celui qui, par la simple lecture, réussit à pénétrer ces œuvres et qui possède un esprit capable de suivre dans son vol la brillante imagination de Konrad jusqu'aux cimes vertigineuses ou se forgent les destinées humaines — 'au-delà des étoiles, au-delà de la ronde des sphères, là ou communient le Créateur avec la Création'." Il importe de faire remarquer que lors de son analyse de Dziady Slavejkov fait à maintes reprises expressément ressortir ce procédé de Mickiewicz qui consiste à donner une signification symbolique à des images et à des événements connus. Parlant de Gustaw, il écrit: "Il nous intéresse non pas en tant que héros d'un poème en qui s'incarnent des traits typiques d'un personnage vivant, mais bien plutôt en tant que figure dont se sert l'auteur du poème pour exprimer en des paroles inspirées ses propres sentiments." Analysant la troisième partie de Dziady Slavejkov souligne: "Konrad incarne les souffrances du peuple polonais. ... Dans cet événement (l'arrestation et le sort des Philarètes) le poète a voulu symboliser le destin de toute la Pologne." Il est bien évident qu'il n'est point nécessaire de nous tourner vers Slowacki, comme le fait Bobek, pour comprendre comment le poète bulgare parvient à l'idée de la fonction symbolique des personnages et des événements dans la composition épique — l'influence de Dziady sous ce rapport est parfaitement admissible. Mais dans ce cas en quoi se manifeste l'influence de Pan Tadeusz! Indiscutablement, et en premier lieu en cela que Pan Tadeusz incite fortement Slavejkov à entreprendre une œuvre d'une ampleur analogue, l'épopée de son peuple. La première idée du poète naît au début des années 90 du XIXe siècle. Dans un des cahiers conservé dans ses archives, où il rédigeait son journal, Slavejkov dès 1895 développe le plan de "KtpBafla neceH". Il note au début qu'en 1891 il avait écrit un 'chant' (plus exactement un poème), intitulé "MjiaaeH 3aropeua", probablement inspiré par le destin de Stara Zagora à la veille de la Libération. Ce cahier-journal de 1895 commence par la note suivante: "MjiafleH 3aropeua" n'y était pas. Je vois maintenant que le plan en était enfantin et qu'avant d'entreprendre sa réalisation, je devrais remanier radicalement le plan lui-même. — J'ai décidé: détruire ce chant écrit en 1891. Abandonner le plan ancien. Attendre que prenne corps ce que je sens sourdre en moi, avant de me mettre au travail. Mon journal est destiné à moi-même; il doit me tracer la voie que je dois suivre afin d'éviter toute confusion à l'avenir et de ne pas revenir sur mes pas comme jusqu'à présent je ne l'ai que trop souvent fait. L'ouvrage qui m'attend est beaucoup plus important et exige une pleine lucidité." Que c'était-il passé au juste? En 1892-1898 Slavejkov est à Leipzig, et il a apporté avec soi le manuscrit de "Mjia/ieH 3aropeua". C'est alors qu'il prend connaissance de Pan Tadeusz qui lui œuvre de nouvelles perspectives créatrices. Le poème de Mickiewicz produit une très forte impression sur Slavejkov qui devait plus tard en parler en des termes dithyrambiques. C'est tout d'abord par le professeur Volkelt

LES ÉCHANGES LITTÉRAIRES BULGARO-POLONAIS

521

que Slavejkov entend vanter les grandes qualités littéraires de Pan Tadeusz et il se met à lire ce poème avec passion, ainsi qu'en témoignent ses condisciples à l'Université de Leipzig. Après avoir lu Pan Tadeusz Slavejkov dresse fiévreusement des plans nouveaux et abandonne "Mjia^eH 3aropeiia" qui, selon ses propres termes, lui semble maintenant excessivement simpliste, 'enfantin', et d'une conception par trop étriquée. Il est captivé par une grande et ambitieuse idée: créer une épopée nationale sur la base de ce grand événement historique que fut l'Insurrection d'Avril 1876. C'est ainsi que naissent la conception et le plan de "KtpBaBa neceH". "L'ouvrage qui m'attend est beaucoup plus important et exige une grande lucidité", note-t-il dans son journal. C'est là que réside le rôle bénéfique de Pan Tadeusz pour la genèse de "KtpBaBa neceH". Slavejkov voit dans le poème polonais, tout comme dans l'Iliade, un modèle pour son œuvre. Il s'inspire de son style épique, il étudie ce style et s'en sert, en intégrant ses particularités dans son poème. De la conception de Mickiewicz, Slavejkov admire le plus l'incarnation de l'idée de la lutte de libération nationale. Dans son analyse de Pan Tadeusz, Slavejkov estime que cette lutte, en tant que motif fondamental, occupe le premier plan ; de son côté, il fait lui-aussi de cette lutte le motif fondamental de son poème et, mieux encore, il le développe avec bien plus d'ampleur que Mickiewicz en faisant de ce motif le centre idéologique de l'ensemble de la composition de l'œuvre. Lorsqu'on parle des affinités existant entre Pan Tadeusz et "KtpBaBa neceH" un autre fait s'impose à notre attention : les analogies que l'on retrouve dans toute une série de détails. Ces analogies ont déjà été relevées par la plupart des chercheurs (surtout par A. Jensen et M. Nikolov) : figuration de certains personnages, de la vie, des paysages de la nature, certains problèmes stylistiques. Tel est le cas, par exemple, vers la fin du premier chant dans les deux poèmes, où il est question de l'agitation révolutionnaire et des émissaires secrets arrivés en Lituanie, ainsi que des réunions clandestines et des nombreux apôtres parcourant les villages et les villes de Bulgarie. L'adresse de la fin du premier chant chez Slavejkov rappelle incontestablement la célèbre tirade lyrique du Xle chant de Pan Tadeusz: O HyflHH BpeMeHa! EpoaceHbe HenpHKpHTO — OT Hero Bi.3flyxa 6e CHKarn 3apa3eH. B MeHTa H noMHCJiH 3a ô t ^ e m n a ,neH 6e HacTOHiuHH 3a6paBeH ... Ben HacraHa H UBeraa npojieT; eft HaMeraa caM EajiKaHa 3e;ieHa .nojxaMa; paôoTHH aparn AHH flofifloxa, ajia ce Hap^Ko B paBHHHH opaHi>T oxojieH BecTHBame c opajio H KaTo HHKora 6e rjiyxo, 3anycT5Dio no m.THma, npeflH Tyâ niyMHH B npojierra ...

Et dans Pan Tadeusz :

522

PETAR DINEKOV 0 roku 6w! kto ciebie widzial w naszym kraju! Kiedy pierwszy raz bydlo wygnano na wiosnç, uwazano, ze chociaz zgiodniale i chude, nie bieglo na ruri, co juz umaila grudç, lecz kladto siç na rolç, i schiliwszy gtowy, ryczalo, albo zulo swôj pokarm zimowy. J wiesniacy, ci^gn^cy na jarzynç phigi, nie ciesz^ siç, jak zwykle, z konca zimy dlugiej, jakby nie pamiçtali na zasiew i zniwo. Co krok wstrzymuj^ woly i podjezdki w bronie, 1 pogl^daj^ z trwogq ku zachodniej stronie.

De tels exemples d'analogies dans les détails, ne sont pas rares dans "KtpBaBa neceH". Mais ces ressemblances entre l'œuvre de Slavejkov et le poème de Mickiewicz, ne déterminent nullement le lien entre les deux œuvres ,pas plus qu'elles ne nous fondent d'affirmer que Slavejkov a copié servilement son illustre prédécesseur. Ce sont des cas isolés. Ce qui est important pour nous est que Pan Tadeusz a. àonné une forte impulsion à la création de "KtpBaBa neceH" et que grâce à cette œuvre Slavejkov a atteint le style épique de l'épopée nationale. Il a appris de Mickiewicz certains procédés stylistiques, s'est inspiré de certains détails — cela ne fait aucun doute — mais il a réussi à créer sa propre conception idéologique et littéraire et à incarner celle-ci dans une œuvre profondément originale et d'une grande portée nationale. Il est évident que le grand poète polonais a joué un rôle important dans le développement créateur de Penco Slavejkov. La question des affinités entre les deux grands écrivains slaves rentrent dans le cadre d'un problème dont l'étude revêt une grande importance et témoigne une fois encore de la fécondité et de la grande signification des rapports mutuels entre les cultures nationales différentes. SOFIA

STOYAN D J O U D J E F F

ESQUISSE D ' U N E MÉTHODE M U S I C O L O G I Q U E P O U R L'ÉTUDE DES VERS POPULAIRES

A musicologie est sœur cadette de la philologie et de la linguistique. Science relativement nouvelle, elle est moins avancés dans son développement. Pour cette raison elle profite souvent de l'expérience, des recherches et des constatations faites par les sciences aînées, en empruntant d'elles des procédés et des méthodes de recherche, des solutions et des découvertes, de l'expérience accumulée. C'est ainsi qu'elle a emprunté à la science du langage maints lois et principes concernant la genèse et l'évolution des formes linguistiques, les ayant appliqués à l'étude de la genèse et de l'évolution des formes musicales. Tenant compte, bien entendu, de la matière spécifique du langage et de la musique, ce procédé s'est avéré très opportun et riche en possibilités. Mais il y a des cas où la philologie et la linguistique peuvent, à leur tour, tirer profit de la science musicale, en utilisant ses méthodes spécifiques, ses acquisitions et ses réalisations. Les recherches dans le domaine des sensations sonores, de l'esthétique et de la psychologie musicales, des zones auditives en ce qui concerne l'intonation, le rythme et le timbre, peuvent trouver une application théorique et pratique dans certaines branches de la science du langage, en particulier dans le domaine de la phonologie, de la psychologie linguistique, surtout en poétique et en particulier dans la science du vers (prosodie, métrique etc.). J'appelle "méthode musicologique" ou simplement "méthode musicale" un procédé de recherche et d'analyse fondé sur la conception qu'il existe une interdépendance étroite entre la parole et la musique et qui se propose de mettre en évidence les propriétés rythmo-mélodiques des textes verbaux. Cette méthode n'est pas nouvelle. Elle a été utilisée déjà dans bien des cas par les théoriciens de la versification ancienne et moderne. Elle a été suivie par O. Riemann et H. Schiller dans l'étude de la métrique classique,1 par Milos Weingart dans les recherches sur le langage parlé moderne, 2 par Rud. Westphal dans son traité sur la 1

Cf. O. Riemann, "Notions élémentaires de musique appliquée à la métrique", dans l'ouvrage de H. Schiller, Mètres lyriques d'Horace (Paris, 1883). 2 Cf. Milos Weingart, "Étude du langage parlé suivi du point de vue musical", dans la collection des Travaux du Cercle linguistique de Prague, I, 1929, 170-242.

524

STOYAN DJOUDJEFF 3

métrique grecque, par Léonide Sabanei'ev dans son étude esthétique sur la musique de la parole,4 par Maurice Emmanuel dans son essai sur l'évolution du rythme,5 etc. L'utilisation de la méthode musicologique en matière de métrique poétique trouve sa justification et sa raison d'être dans deux circonstances principales: 1° dans le fait que la matière sonore des paroles et, en particulier, du langage poétique montre plusieurs éléments en commun avec celle du langage musical; et 2° dans le fait que même de nos jours, la poésie n'a pas encore rompu tout à fait ses liens avec la musique et la danse, qu'elle n'a pas perdu définitivement son caractère syncrétique (au sens donné à ce mot par Alexandre Veselovskij) — bien que les poètes contemporains ne chantent plus leurs vers, ne les accompagnent plus aux sons d'un instrument de musique et ne les soumettent pas au rythme des mouvements corporals comme les poètes et dramaturges de l'antiquité. Néanmoins ils ont gardé la réminiscence du syncrétisme, lorsqu'ils déclarent, avec Verlaine, que la poésie est "de la musique avant toute chose", ou lorsqu'ils affirment avec Edgar Allan Poe: "Contenting myself with the certainty that Music, in its various modes of metre, rhythm and rhyme, is of so vast a moment in Poetry as never to be wisely rejected — is so vitally important an adjunct, that he is simply silly who declines its assistance, I will not now pause to maintain its absolute essentiality. It is in Music, perhaps, that the soul most nearly attains the great end for which, when inspired by the Poetic Sentiment, it struggles — the creation of supernal Beauty. ... And thus there can be little doubt, that in the union of Poetry with Music in its popular sense, we shall find the widest field for the poetic development. The old Bards and Minnesingers had advantages which we do not possess — and Thomas Moore, singing his own songs, was, in the most legitimate manner, perfecting them as poems" (Edgar A. Poe, "The poetic principle", The Works of E. A. Poe, vol. I, p. 173-174, L. 1927). Mais pour l'utilisation de la méthode "musicale", il y a, de plus, des considérations d'ordre phonologique. Le langage parlé, qu'il soit en vers ou en prose, possède presque tous les éléments et toutes les propriétés du langage musical notamment : 1. Le rythme, c.-à-d. la succession régulière de valeurs longues et brèves, qu'on peut noter avec des signes musicaux. Par exemple, le vers de l'Hymne delphique: àtnôç êç "OX,u|X7tov àvcnaSvaxai Xiyú 5è Xaxôoq Ppé|xœv présente un rythme péonique identique au rythme musical J J J J |

cJJJJlcJJcJlJJJdiJJJd

2. Le mètre resp. la mesure, c.-à-d. la succession de valeurs isochrones servant d'unités de mesure pour les valeurs et les groupes rythmiques. C'est ainsi que le distique de Baudelaire: Pour l'enfant / amoureux / de cartes / et d'estam/pes // L'univers / est égal / à son vaste / apétit / — // 8

Cf. Rud. Westphal, Allgemeine griechische Metrik (Leipzig, 1865). Leonid Sabaneiev, Muzyka reci, èsteticeskoe issledovattie (Moscou, 1923). 6 Maurice Emmanuel, "Le rythme d'Euripide à Debussy" (Dans les matériaux du Premier congrès du rythme, Genève, 1926), 103-146.

4

ESQUISSE D'UNE MÉTHODOLOGIE MUSICOLOGIQUE

525

se conçoit comme une série de mesures à trois temps:

3. L'intonation, c.-à-d. l'inflexion de la voix dans son mouvement ascendant ou descendant déterminé par les syllabes accentuées ou atones, par la ponctuation etc. La mélodie de la parole peut être représentée graphiquement par une courbe dont les abscisses indiquent le temps (les moments successifs) et les ordonnées — les hauteurs musicales; ou bien on peut disposer cette courbe sur une portée musicale, ce qui revient au même (le sens droit des 5 lignes horizontales remplissant la fonction d'abscisse) et le sens vertical — d'ordonnée. (Voir les exemples musicaux, N° 1.) 4. L'harmonie ou l'euphonie, c.-à-d. le rapport de sonorités (identiques, semblables ou analogues) entre différents sons resp. voyelles, syllabes ou groupes de syllabes comme dans les cas des rimes, des allitérations et des assonances, ainsi que dans les cas de la synharmonie ou de l'harmonie vocalique et consonantique de certaines langues finno-ougriennes et ouralo-altaïques, le hongrois, le turc etc. P. ex. seretem, mais kôsônôm ou scokolom en hongrois ; cf. goljamo, mais golemi (assimilation sous l'influence de la désinence) en bulgare etc. 5. La dynamique ou l'expression, c.-à-d. la différence de l'intensité par laquelle se profèrent les syllabes et les groupes de syllabes (mots, membres de phrases, propositions, etc.) en vue de la déclamation plus ou moins expressive. Les signes spéciaux de la dynamique et de l'expression musicales—forte ( f ), mezzoforte {fin), fortissimo ( ff ), piano (p), pianissimo (pp), crescendo (cresc.), diminuendo ( d i m . ) , decrescendo (decresc.) — ont trouvé une application pratique dans l'art théatral, oratoire et déclamatoire. 6. Le coloris ou timbre, c.-à-d. la nuance ou la sonorité spécifique des sons vocaux — voyelles, consonnes, syllabes ou groupes de syllabes — que certaines personnes douées de synesthésie conçoivent toujours comme liée à une couleur correspondante. 6 7. L'élément agogique, c.-à-d. la vitesse par laquelle se succèdent les valeurs isochrones servant d'unités de mesure pour les valeurs et groupes rythmiques. La pratique musicale a adopté pour les différents degrés de vitesse ou "mouvements" des termes spéciaux qui trouvent une vaste application dans l'art dramatique, oratoire et déclamatoire (on peut parler andante, adagio, lento, allegro, prestissimo et, en parlant, faire des accelerandi, des calandi, des rubati etc.). Il y a encore bien d'autres éléments communs du langage parlé et de la musique. Ceux que nous venons de mettre en évidence suffisent à nous convaincre qu'il est extrêmement difficile de tracer une ligne de démarcation entre le langage parlé et le langage musical, entre la musique et la parole. Dans la pratique musicale toutes ces propriétés rythmiques, métriques, mélodiques, harmoniques, dynamiques, agogiques et chromatiques des sons et de leurs successions sont mises en évidence et montrées avec plus de précision que dans la pratique orale. Pour ses fins esthétiques la musique a élaboré un système de mensuration et de notation plus exacte que celui de l'écriture 8

Voir Arthur Rimbaud, Une saison en enfer, Sonnet des voyelles (Paris, 1873).

526

STOYAN DJOUDJEFF

et de la mensuration du langage parlé. C'est ainsi que l'écriture musicale a créé des signes spéciaux et pour le rythme (valeur des notes) et pour le mètre (barres demesure) et pour la hauteur mélodique (portée musicale), et pour la dynamique (signes d'expression musicale) et pour l'agogique (signes de mouvement), tandis que l'écriture ordinaire du langage a tout à fait négligé les éléments musicaux de la parole. De plus, la pratique musicale a procédé à une mensuration plus ou moins exacte de ces qualité en y introduisant des unités de mesure pour la hauteur, la durée, la vitesse, l'intensité etc. L'écriture ordinaire du langage ne tient pas compte des valeurs quantitatives de ces éléments et c'est à cause de cela qu'elle est moins précise que l'écriture musicale. Pour cette raison, un des principes de la méthode musicologique en matière de métrique poétique consiste à appliquer le système de mensuration et de notation musicales aux phénomènes du langage parlé. Cela nous amènera à traiter la matière sonore du langage de la même manière dont les musiciens traitent la matière sonore de la musique. Connaissant mieux les langues et le folklore des peuples bakaniques et plus spécialement des Slaves du Sud, nous nous proposons d'utiliser la méthode musicologique pour l'étude de la métrique des chants populaires macédoniens, serbes, surtout bulgares. Mais cette méthode peut trouver d'application à l'étude des systèmes métriques appartenant à des familles de langues très différentes. Lorsqu'on étudie la poésie et la musique vocale des peuples balkaniques — grecs, albanais, roumains, et en particulier celle des macédoniens, des serbes et des bulgares — le principe le plus général qu'on relève et qui se jette dès prime abord à l'esprit de l'investigateur, peut s'énoncer ainsi: D'une manière générale les textes verbaux ("les paroles") de toutes les chansons populaires sont mis "en vers", c.-à-d. ils suivent très fidèlement la règle du nombre et de la mesure, du rythme et de la cadence. Il n'y a presque pas de mélodies vocales avec des paroles en prose. Et vice-versa : tout texte poétique conçu en vers est destiné à être chanté et possède en général une mélodie. Cette mélodie est ou bien encore vivante, c.-à-d. qu'elle est encore en circulation sur un territoire ethnique plus ou moins étendu ; ou bien elle est déjà oubliée et retirée de circulation, c.-à-d. qu'elle est effacée de la mémoire musicale du peuple; ou bien elle est encore vivante et en circulation, mais n'a pas encore été récuperée par l'investigateur-folkloriste. Mais même dans le cas où elle est complètement disparue de la conscience musicale du peuple, des vestiges de son rythme et de son mélos se retrouvent dans le mètre et dans l'intonation de son texte verbal rythmiquement articulé, qui garde le souvenir de cette forme déjà disparue, où la musique et la poésie ont coexisté à l'état syncrétique. Et, en suivant une méthode rationelle et spécifique, en tenant compte du nombre et de la mesure, des coupes et des césures, de l'accentuation et de la forme du texte verbal, on pourrait restaurer, dans des traits les plus généraux, le rythme, le mouvement mélodique et la forme du texte musical disparu. Une telle méthode a été suivie par le musicologue soviétique Kusnarev pour la restitution des melodies perdues de l'ancienne musique arménienne. 7 ' Cf. H. S. Kusnarev, Voprosy

istoriiiteoriiarmjanskojmonodiceskojmuzyki(L.,

1958), 18-19,27 et suiv.

ESQUISSE D'UNE MÉTHODOLOGIE MUSICOLOGIQUE

527

D'autre part, le phénomène inverse est souvent observé: au lieu de la mélodie, c'est le texte verbal de la chanson populaire qui se retire de circulation, tandis que son texte musical continue à circuler sous la forme d'une mélodie purement instrumentale. Et, d'une manière analogue, on peut retrouver dans le rythme, dans les clausules, les cadences, la mesure et le mouvement de telles mélodies instrumentales les vestiges de leur texte verbal oublié. On peut facilement y reconnaître les mètres, les cola et les strophes que formaient jadis les paroles disparues, on peut même déterminer le nombre des syllabes dans les vers et les hémistiches, trouver les places des césures, des coupes et des diérèses. Les mélodies instrumentales gardent donc, elles aussi, dans leur structure rythmique et musicale, le souvenir du stade de leur état syncrétique. Et de nos jours, lorsqu'on écoute de telles mélodies exécutées par d'habiles joueurs du kaval ou de la gâdulka, on a l'impression que l'instrument parle, qu'il récite des vers dont la structure, la forme et la mesure sont si nettement traduites par l'exécutant, que l'auditeur peut à l'instant même reconnaître le type du vers, le nombre des syllabes dans ses subdivisions et le nombre de vers dans les strophes. En analysant une mélodie instrumentale qui a survécu jusqu'à nos jours et dont le texte verbal a disparu, il est donc possible de restaurer, sinon les paroles concrètes, du moins le schéma métrique de ses vers, tout comme il est possible de rétablir en traits généraux les éléments musicaux d'une mélodie disparue, en analysant son texte verbal qui lui a survécu. Pendant ce travail de restauration, il faut pourtant avoir toujours présent à l'esprit un phénomène très fréquent en matière de folklore: l'interférence musicale. Plusieurs textes verbaux très différents peuvent se superposer successivement sur une seule mélodie; ou vice-versa: un seul texte verbal peut supporter l'agglomération de plusieurs mélodies successives. Ces phénomènes d'interférence de textes ou de mélodies peuvent expliquer les exceptions et les écarts fréquents de la règle générale, qui doivent toujours être éclaircis et commentés par l'investigateur. Abstraction faite des superpositions tardives et artificielles de textes sur l'œvre folklorique, il s'est établi dans les chansons populaires originales et authentiques, pendant la longue symbiose de l'élément musical et poétique, un parallélisme, coordonné dans les détails, qui lie très étroitement le texte verbal à la mélodie — et où le problème de priorité ou d'antériorité chronologique de celui-ci ou de celle-là n'a pas pour l'instant d'importance méthodologique. On peut dès lors appeler en aide ce parallélisme, pour la restitution des éléments perdus dans tous les cas où l'une ou l'autre des deux faces de l'œuvre syncrétique fait défaut ou présente certaines incertitudes ou lacunes. Essayons maintenant d'appliquer ces principes à l'étude des vers populaires. La subdivision du texte poétique en strophes, couplets, vers, hémistiches et segments de vers s'effectue par des arrêts ou pauses régulières — césures, coupes, diérèses — qui jouent dans le langage versifié un rôle métrique de première importance. Ces pauses se situent dans le texte à des places rigoureusement constantes, qui ont l'air d'être fixées d'avance, et apparaissent périodiquement et régulièrement, c.-à-d. après

528

STOYAN DJOUDJEFF

un et même nombre de syllabes. C'est grâce à elles que peut se faire le discernement entre un texte mis en vers et un texte à rythme libre ("en prose"). Dans le langage non versifié leur place est arbitraire, c.-à-d. qu'elle est déterminée par des facteurs extramusicaux: la grammaire, la syntaxe, la structure et la ponctuation de la phrase. Dans le langage versifié, la place des pauses (arrêts, suspens) n'est pas libre. Elle n'est déterminée ni par les exigences de la grammaire et de la syntaxe, ni par celles de la ponctuation, mais par une règle ou une norme esthétique qui est de nature purement musicale. A cette règle esthétique que nous appelons conventionnellement "nombre musical", sont subordonnés tous les autres facteurs. Ce nombre est une constante. C'est le nombre des syllabes dans chaque vers et dans chacune de ses subdivisions. Dans le langage versifié il y a donc un certain nombre de lieux fixes ("points morts") qui sont marqués par la fin d'un mot ou d'un syntagme (dans le sens que Ferd. de Saussure attribue à ces terme) et le commencement d'un autre. L'aède est donc obligé de faire le choix des mots de telle manière qu'à chaque lieu fixe ou point mort du texte apparaisse une fin de mot ou une fin de syntagme et un commencement d'un autre mot ou d'un autre syntagme, c.-à-d. que les césures ne coupent pas les mots ou les syntagmes, mais qu'elles se situent entre deux mots, resp. syntagmes différents. Dans la poésie populaire bulgare, macédonienne, serbe, croate, monténegraine, grecque, albanaise et roumaine cette règle est de rigueur pour les césures et les coupes séparant les strophes, les vers, même les héuistiches, et n'admet pas d'exception. Ainsi, dans les décasyllabes suivants: Stàpil Dobri // na bjal mermer kamàk // Da si meri // siva gurgulica // Ne umeri // sive gurgulica // Naj umeri // kletoto si sârce // la césure apparaît après la quatrième et la dixième syllabe de chaque décasyllabe, et chaque césure est précédée par la syllabe finale d'un mot (Dobri, kamâk, meri, gurgulica, umeri, gurgulica, umeri, sârce) et suivie par la syllabe initiale du mot suivant (na bjal, da si, siva, ne umeri, s/va, naj umeri, kletoto si). De même dans les hendécasyllabes suivants, où les proclitiques et les enclitiques forment avec le mot principal des unités phonético-semantiques indivisibles: Dimitro^le // rusokoso // momice, /// idi kazi // n a ^ m a j k a w s i , // Dimitro, /// Da n e w r a z d a // drugo ôedo // kato teb /// ... Et dans des formations métriques plus étendues (vers de 13 syllabes): Sànja sânuvah, // mane mo // Sânja bàlnuvah /// V sabota vecer // mane mo // srestu nedelja /// Ce sme imali // mane mo // kâsti visoki /// Kâsti visoki // mane mo // dvori siroki /// Vàv dvorovete // mane mo / / d o tri karâka /// Do tri karâka // mane mo // cvete nasjato /// ...

ESQUISSE D'UNE MÉTHODOLOGIE MUSICOLOGIQUE

529

Pour les diérèses, séparant des groupes ("mesures") de moins de 4 syllabes, la règle est encore de rigueur, quoiqu'elle présente bon nombre d'exceptions. Ces exceptions s'expliquent, cependant, par l'impossibilité pour l'auteur de trouver toujours des mots de deux ou de trois syllabes pour exprimer ses idées. Malgré cette difficulté, l'effort de l'aède d'y obvier se manifeste très nettement dans les octosyllabes suivants : Ot carja / haber / doftasa // ot carja / i ot / vizirja // Koj ima / sina / da prati // koj njama / samsi / da ide // Na carja / sluiba / da sluzi // na carja / i na / vizirja // Nano mi / Nano / bajraktar // toj njama / sina / da prati // Toj ima / devet / dàSteri // où les "points morts" (lieux fixes) apparaissent infailliblement après la troisième, la cinquième et la huitième syllabe de chaque octosyllabe et à tous ces endroits il y a une fin de mot et commencement d'un autre mot; c.-à-d. la troisième, la cinquième et la huitième syllabes de chaque vers sont des syllabes finales de quelque mot, tandis que la première, la quatrième et la sixième sont des sllabes initiales d'autres mots. Il est vrai, qu'il peut avoir des fins de mots aussi à d'autres lieux du texte versifié — mais ces lieux ne sont pas des lieux constants comme ceux que nous venons de relever. Dans les octosyllabes symétriques suivants : Boli / Jana // belo / gàrlo /// lele / Jane // mome / Jane /// Ne e / gàrlo // kato / gârlo /// lele / Jane // mome / Jane /// Naj e / gàrlo // ljuta / rana /// lele / Jane // mome / Jane /// l'aède, pour rester fidèle à la règle musicale, a été amené à n'employer que des mots mono- et bisyllabiques. Dans ce cas, les "points morts" (pauses fixes) apparaissent périodiquement après chaque groupe de deux syllabes, dont la première est initiale et la seconde — finale. Le nombre musical fixé d'avance pour ce texte peut s'exprimer par la mesure à \ ou | = Les vers suivants à 13 syllabes n'admettent pas de mots supérieurs à 3 syllabes: Snosti / dojde // ôu2di / junak // u nas / na konak /// Sedja / dva dni // sedja / tri dni // sedja / nedelja /// Sutrin / rano // ôuzdi / junak // konôe / si stjaga /// Ja si / konôe // stjaga / Nano // iska / da bjaga /// Ici également les fins des mots, paraissant toujours à des lieux fixes, déterminent des pauses (suspens de mouvement) plus ou moins perceptibles à l'oreille qui se succèdent d'une manière régulière, formant des groupes de deux et de trois syllabes, et qui, eux-aussi, alternent périodiquement. Pourtant toutes ces pauses n'ont pas la même valeur: les unes sont plus brèves, à peine perceptibles, les autres — plus longues. Les mieux marquées sont celles qui séparent deux strophes ou deux couplets voisins, viennent ensuite celles qui sont situées entre deux vers, puis — celles entre les hémistiches et les segments à l'intérieur

530

STOYAN DJOUDJEFF

d'un vers, et enfin les arrêts imperceptibles entre les "phalanges" (groupes dissyllabes et trissyllabes) à l'intérieur d'un hémistiche ou d'un segment de vers. Nous proposons les termes conventionnels "repos", "cesure", "coupe" et "diérèse "pour la dénomination respective de chacune de ces quatre espèces de pauses — abstraction faite de la signification traditionnelle de ces termes. Ainsi nous entendrons par le terme "repos" la pause qui marque la limite entre deux strophes ou couplets ; par "césure" — celle qui marque la limite entre deux vers voisins; par "coupe" — la subdivision du vers en hémistiches et segments; et par "diérèse" — la pause à peine perceptible qui sépare deux "phalanges" (groupes dissyllabiques et trissyllabiques) dans l'intérieur d'un hémistiche ou segment de vers. Les repos, les césures et les coupes sont des pauses réelles qui se situent toujours entre des mots différents, même entre des syntagmes différents, ne coupent jamais les mots et ne séparent pas la clitique du mot principal ; tandis que les diérèses sont le plus souvent des divisions imperceptibles et même purement imaginaires qui se situent aussi bien entre deux mots différents, qu'entre les parties d'un et même mot phonétique ou grammatical, coupant très souvent la clitique du mot principal. Nous indiquerons arbitrairement les repos par quatre traits verticaux (////), les césures — par trois traits (///), les coupes — par deux traits (//), et les diérèses — par un trait (/). Exemple

strophe

diérèse coupe diérèse coupe diérèse césure Ogàn / gori // na pla- / nina // viídam go / gledam go /// Dajmi / mamo // veren / drugar // da ida / da vida //// (repos) f Na ti / dâste // veren / baSta // da ides / da vidis /// (césure) [ Ako j / veren // neka j / veren // ne Sta go / ne sta go //// (repos) / da vidis /// (césure) i Na ti / dàste // veren / bratec // da ides [ A k o j / veren // n e k a j / veren / / n e sta go / ne sta go ////(repos) 2 s. 2 s. 2 s. 2 s. 3 s. 3 s. phal.

phal.

segment *

phal.

phal.

segment „

phal.

phal.

segment '

vers Le texte ci-dessus est formé par trois distiques isométriques séparés l'un de l'autre par des "repos" (////). Les deux vers à 14 syllabes de chaque distique sont délimités par une cesure (///). Chaque vers se subdivise en trois segments inégaux séparés par des coupes (//): deux tétrasyllabes et un hexasyllabe. Les segments tétrasyllabiques sont formés de deux "phalanges" bisyllabiques, tandis que le segment hexasyllabique est formé par deux phalanges trisyllabiques. Entre les phalanges (groupes bisyllabiques ou trisyllabiques) sont situées les diérèses (/). La coïncidence exacte des fins et des commencements des mots avec les lieux fixes

ESQUISSE D'UNE MÉTHODOLOGIE MUSICOLOGIQUE

531

des repos, des césures, des coupes et des diérèses, ou plutôt la succession régulière et périodique de ces fins des mots, alternant toujours avec un nombre constant de syllabes, constitue une preuve suffisante pour l'existence objective de ces pauses telles que nous venons de les définir. Il est évident que les fins et les commencements des mots tombant sur des lieux fixes dans les textes des chansons, ont une fonction phonologique, morphologique, semantique et métrique très importante, dès que l'aède tient tellement à les mettre en évidence et à les disposer sur les points cardinaux de la forme poétique comme l'a très justement observé Antoine Meillet en ce qui concerne la métrique des langues grecque et indo-européenne. 8 La nature des repos, césures, coupes et diérèses, considérées exclusivement du point de vue linguistique, n'est pas purement phonétique. Elle relève plutôt de la semantique. Car ces pauses sont engendrées non seulement et non autant par la fin du mot considéré comme unité phonétique, mais aussi et surtout par la fin du mot considéré comme unité sémantique. Pour cette raison les recherches de la phonétique expérimentale ne s'avèrent pas toujours suffisantes pour la délimitation des strophes, des vers, des hémistiches, des segments et des phalanges des vers, aussi bien que pour l'éclaircissement de leur structure. Phonéticiens et folkloristes, versés dans les sciences physiques et mathématiques, déploient, depuis quelques dizaines d'années, un zèle assidu à faire des mesures et des évaluations exactes des valeurs rythmiques (durées) et des intervalles mélodiques du langage parlé ou chanté. Nous n'avons aucun dessin à sous-estimer leur travail scientifique ni leur contribution à la théorie du vers. Nous voudrions seulement faire observer que les principes de la métrique poétique relèvent non seulement de la nature phonétique et physique des textes, mais aussi et surtout de leur nature sémantique; que la parole et le chant sont des phénomènes psychophysiologiques où il y a des moments subjectifs et irrationnels que les évaluations méchaniques et mathématiques les plus exactes ne peuvent pas expliquer; et que les chanteurs populaires, qui sont le plus souvent des gens modestes et sans préparation scientifique, n'ont jamais songé à faire des mesures et des évaluations mathématiques. Malgré cela, ils composent et exécutent leurs textes avec une justesse musicale, quoique non mathématique. D'autre part, quoique les pauses d'articulation — repos, césures, coupes et diérèses — sont de nature et d'origine phonético-sémantique, cependant la règle qui gouverne leur distribution dans le texte versifié et y détermine leurs places, n'est ni d'ordre phonétique, ni d'ordre sémantique. Elle est de nature et d'origine purement esthétique et relève des lois de la rythmique et de la métrique musicales. C'est ainsi que les syllabes forment des groupes d'un certain nombre d'unités, groupes que nous considérons comme des "mots métriques" et qui, surtout pour les petites divisions, ne coïncident pas toujours avec les mots considérés comme des unités phonétiques et sémantiques. Cependant dans le langage versifié on voit la tendance très nette à adapter et faire coïncider les subdivisions (mots et syntagmes) phonético-sémantiques 8 Cf. A. Meillet, Les origines indo-européennes La fin du mot en indo-européen (Paris, 1913).

des mètres grecs (Paris, 1923). Voir aussi Gauthiot,

532

STOYAN DJOUDJEFF

avec les subdivisions métrico-esthétiques. Cela se fait par la distribution des fins des mots d'après les modèles musicaux qui résident dans la mémoire collective du peuple depuis des temps immémoriaux. Ainsi l'apparition périodique et régulière des cesures, coupes etc., distribuées d'après ces modèles, morcelle le texte versifié en groupes de syllabes isochrones qui constituent de véritables mesures musicales. Le bulgare moderne ne possédant pas de quantité prosodique, toutes les syllabes sont censées être brèves et par conséquent isochrones. Il suffit donc de donner à chaque syllabe la valeur arbitraire d'une croche (J*), pour obtenir de n'importe quel texte une série de valeurs égales. Par exemple, la série des tétrasyllabes de la forme — / — a pour prototype musical la mesure à § ou § et les diérèses, resp. les coupes, coïncident avec les barres de mesure distribuées dans la série de valeurs isochrones.

n\ri\n\n\ n\n\n\n\n\

Padna -/ la e //tamna/mitgla//ne e/ bilo //fâmna/mâgla //naj e/biIo//tïmen/ oblak, etc. (La collection Sredna-severna Bàlgarija, N° 49). Les pentasyllabes de la forme / - - sont calqués d'après la mesure musicale à s = J> Stojan/ i* w zaborfclja//pedese/ i* m j) i

J> hiljadi Ji i*///J>J>i

i pet/

Ici outre les césures et les coupes, toutes les diérèses sont projetées dans la mélodie par l'allongement de la valeur précédente, ce qui donne en résultat la différenciation de chaque hexamètre, de chaque hémistiche (octossyllabe) et de chaque phalange — aussi bien dissyllabique que trissyllabique — dans le schéma rythmique. Ainsi, par la dérogation de l'isochronisme primitif à la troisième, cinquième, huitième, onzième, treizième et seizième syllabes, le rythme de la mélodie a formé quatre anapestes (w ^ - ) et deux iambes (^ - ) hémioliques. Les anapestes ont déterminé quatre mesures à ^ tandis que les iambes — deux mesures à (Dans la mélodie l'anapeste est traduit par le rythme J" j ¿ . à ^ et l'iambe — par le rythme J" J". à Cette circonstance a amené le poète bulgare Asen Razcvetnikov à l'hypothèse que l'octosyllabe asymétrique /— / n'est qu'un des deux hémistiches d'un vers de seize syllabes dérivé de l'hexamètre antique par la perte de la quantité prosodique. Lorsque la différence entre les syllabes longues et brèves s'est effacée, l'hexamètre s'est transformé en un vers de seize syllabes, divisé par la coupe en deux octosyllabes isochrones.9 "Il suffit", déclare Razcvetnikov, "de jeter un coup d'œil attentif sur l'octosyllabe épique de nos chansons des haïdouks et nous verrons combien la forme et le rythme de ce vers ressemble à ceux de l"Iliade'. Il suffit d'écrire sur une ligne deux vers d'une quelconque de ces chansons de haïdouks, et nous aurons la forme extérieure approximative de l'hexamètre classique : * Cf. As. Razcvetnikov, Balgarskijat heksametar (Sofia, Biblioteka Veèni izvori, 1942), 8.

538

STOYAN DJOUDJEFF

Grozdanka po dvor hode§e, drebni si sàlzi ronese, Bogdana ljuto kalneSe: 'Bogdane, Bog da te ubij' La carcasse graphique de ces vers est:

Il est clair, que la ressemblance entre l'octosyllabe de notre chanson épique et l'Iliade, le poème épique de l'antiquité, ne peut pas être aléatoire, mais qu'elle s'explique, très probablement, par quelque affinité lointaine voilée dans l'ombre des siècles. Il incombe à notre science la tâche d'explorer et d'expliquer cette affinité."10 L'hypothèse de Razcvetnikov trouve certaine confirmation dans tout un cycle de mélodies populaires du type de la chanson Zadlâznja Stojan, zaborclja dont la carcasse rythmique comporte deux espèces de valeurs, longues et brèves, formant six pieds métriques — quatre anapestes et deux iambes. Ces valeurs longues et brèves se reflètent sur les syllabes correspondantes du texte verbal en leur communiquant ainsi la quantité prosodique respective. On voit dans le rythme de ces mélodies les vestiges de l'hexamètre dactylique, dans lequel, pour une raison ou une autre (que nous ne traiterons pas ici) les dactyles se sont transformés en leurs équivalents, les anapestes, et les trochées — en iambes. Tandis que le double octosyllabe asymétrique de la poésie épique bulgare (aussi serbe et macédonienne) n'a conservé que la carcasse syllabique de l'ancien hexamètre homérique, pendant que son rythme authentique, basé sur l'alternance de syllabes longues et brèves, a été perdu avec la disparition de la quantité prosodique dans les langues modernes, la forme musicale, qui a emprunté ce rythme à la forme poétique, paraît avoir été plus résistante, car elle l'a conservé jusqu'à nos jours — quoique dans un aspect plus ou moins altéré — dans les mélodies populaires des pays balkaniques. C'est encore la méthode musicologique qui nous permet de poursuivre nos recherches dans ce domaine, en vue d'expliquer cette "affinité lointaine" entre l'octosyllabe épique bulgare et la poésie héroïque de l'antiquité. En général, la méthode musicologique s'avère particulièrement fructueuse dans tous les cas où les textes verbaux des chansons présentent des irrégularités, des difformations, des formes abnormales, douteuses ou difficiles à analyser. SOFIA

10

Razcvetnikov, ibid., p. 13.

539

ESQUISSE D UNE METHODOLOGIE MUSICOLOGIQUE

Zalip- ca

ru-sa

Fpu-stata rav- na

de- voj-ka za-Ie- ze

Dob-ru-

te - be

so

Do-ve- ie-ra

ta- ja

Ko-ljo

dza(2) tripä-ti jag- ma

de -

A- Ii- ja mo-re

vgu-ra

-

sta- na -

Ii- ja (2) soj o - va

pus- ta

lté ti

ze- le-na

doj -

tarn -

lo

eu-do

bu - ra

da

da

ti ka-za Ko-ljo

es- en-na sla -

na

stu-de- na

sen

bra -

ce ta ne sta.

es- en-na

sla -

na

stu- de - na

Mo Den go bra-la

ma den

Ro -

Id- ti - la,

sa

ro den

go bra-la,

den

la

M- t i - la.

(2)

STOYAN DJOUDJEFF

Bo-

lffp • mi

le-2itmo-

bo-Ien

r

f

-f-f-r

Di-mit- r o le

ko-so

f¡ tri

de-do

é u - bri-òan-ie le,

Co -

-Ma-

ne

mo

mi -

mi-

Slan- c e -

ra

Ka -

Mu-

ra

Mu-

sta-

fa

(2)

sta -

Í

r u - so

ì" a. QA11 fllP / (7 fl—n—fi P o - s a - di

Ka-

re

la

to

Za-dal-ánja

mo-mi- ce,

m . ff P [1 p »• le - í l u - k a

be-lo m o - m i - í e ,

lo

òe -

ma -

tre -

Sto-jau,

do,

mo

i-di

fì ájr

namaj-kasi,

np—fl—p—«Ì— « n J>

cik c i - g u - l i - g u

ja

koj

ed-

p e - re,

ka-ii

iz-ka-Si

5u -

no

baj- no -

Za-bor-zvi

se

T i ßi

Di-mit-ro

J> » J» « }•

tri

le-i

P 1ï

lu-ka/

n a di - v a - n a - t a

ka

na

por -

ti?

lu-

do

mia -

do.

le

Jo

-

v a - ne.

LUBOMfR DOLE2EL

THE TYPOLOGY OF THE NARRATOR: POINT OF VIEW I N FICTION

I

N last decades, the study of literature, has entered the stage of structural analysis. Russian Formalism, Prague structuralism and American New Criticism are, in my opinion, different manifestations of the same effort, the effort to attain an exact description of literary structures and, in this way, to discover specific features of literature as art. Without invalidating various 'external approaches', structural analysis of literature is concentrated on the study of intrinsic, immanent features Which are considered to be the 'literature itself', literariness. In my paper, one special problem of literary theory will be outlined from the structural point of view. This problem can be referred to as 'typology of the narrator' or, using a term which is common in Anglo-American literary criticism — 'point of view in fiction'. It is an ancient problem in the theory of fiction, one especially treated by Percy Lubbock in The Craft of Fiction (1921) and described in its development in N. Friedman's article "Point of View in Fiction" ( P M L A , LXX, 1955). In the Russian tradition, this problem has been analyzed in the framework of the theory of skaz (B. Ejchenbaum, V. V. Vinogradov) or in M. M. Bachtin's typology of the 'epic expression' (epiceskoje slovo).1 In contemporary German criticism, K. F. Stanzel described the basic forms of 'point of view', referring to them as to 'typische Erzahlssituationen'. 2 It is not my aim to give a survey of the previous investigations in this field, one can be easily found in Friedman's or in Stanzel's studies. Nor will I give a detailed description of the particular types with an exhaustive analysis of the literary material. I want rather to suggest a new approach and a new classification in a strictly structural manner. The main emphasis will be placed on the theoretical and methodological part of the problem. Let me, therefore, begin with a brief summary of several principles which, in my opinion, are a kind of 'axioms' of the structural theory of the literary work of art. 1

B. Ejchenbaum, "Kak sdelana 'Sinel' Gogolja", in: Poetika (Petrograd, 1919); V. V. Vinogradov, "Problema skaza v stilistike", in: Poetika (Leningrad, 1926); M. M. Bachtin, Tipy prozaiceskogo slova, reprinted in: Michigan Slavic Materials, No. 2 (Ann Arbor, 1962). 8 F. K. Stanzel, Typische Formen des Romans (Gottingen, 1964).

542

LUBOMIR DOLE2EL

(1) As already indicated, structural analysis is concentrated on the intrinsic, immanent features of the literary work; no notion with an 'external' reference is to be used in the basic notional framework of the theory. In our special problem, we will, therefore, eliminate the notion of 'author' and all other notions connected with it ('author's intent, approach'; 'author's subjectivity, impersonality'; 'telling — showing', etc.). Instead, the notion of narrator, as a purely structural factor of fiction, will be used. The relationship between the narrator and the author is irrelevant for the structural theory. (2) Structural theory has stressed the formal analysis of the literary work of art. I do not intend to discuss at length the controversial dichotomy of content and form, which has been challenged several times by structuralists. What is of consequence for me in this connection is the fact that in a literary work of art the content is controlled by the form. Using a paraphrase of a famous statement, we can say that there is nothing in content that has not been in form. In other words, in a literary work of art everything to be expressed has its specific expression. This tenet, of course, has very important implications for the methodology of literary analysis. Content can be studied through the form and in the form; thus, an application of exact methods in the investigation of literature is possible, the form being much more accessible to such methods than content. With regard to this interpretation of the interrelationship between content and form, we are able to assess the role of linguistic methods in the structural analysis of literature. The form of the literary work of art is always a verbal form (verbal expression). Adequate methods of investigating literary form are, therefore, linguistic methods. I must, however, add immediately that the notion of linguistic methods, as used here, is to be understood in a broad sense: as a rich scale of very different methods used in the study both of the linguistic system and of linguistic "performance", i.e., including those of stylistics, poetics, psycholinguistics, etc. It is not by chance that the history of structural literary analysis offers many examples of close cooperation between literary science and linguistics. Besides the reason just mentioned, there is another one, perhaps even more important: linguistics, dealing with an exceedingly complex structure, has elaborated or applied most progressive methods of investigating structures. The example of linguistics deserves to be followed in a structural analysis of literature. This is not to be interpreted as expression of 'linguistic imperialism', a claim that the study of literature is merely part of linguistics; I am aware that even in the 'intrinsic' investigation of the literary work of art there exist problems which surpass the competence of linguistics and cannot be solved by linguistic methods proper. On the other hand, there is no better evidence of a misleading direction in literary theory than its isolation from linguistics. (3) In the development of the structural literary theory it has been stressed repeatedly that the chief aim of the theory lies in analysis and description rather than in evaluation. In this respect, structural analysis is antithetical to the traditional literary criticism which concentrates on evaluative judgments, often neglecting

THE TYPOLOGY OF THE NARRATOR

543

analysis. Moreover, traditional literary criticism is impressionistic and subjective, basing its evaluative judgments on a priori criteria of subjective taste or prejudice. In investigations of the problem of point of view in fiction these a priori judgments are very common, at one time favoring 'showing' to 'telling' or 'impersonality' to 'subjectivity', at another the reverse. Structural theory, not bound to changing literary fashions, considers the particular types of narration as manifestations of historically determined efforts of literature to attain at various epochs certain specific aims and effects. There are no scientific reasons to prefer one or another type of narration. (4) Structural analysis is not to be regarded as a mere empirical description of structures and their relationships. One of the most progressive and important fields of structural theory is the TYPOLOGY of literary structures with its high degree of generalization. The pivotal place of typology in the framework of structural theory was pointed out by Tynjanov and Jakobson in their notable theses Voprosy izucenija literatury ijazyka (1927): "Analysis of the structural laws of language and literature necessarily leads to fixing a limited number of realized structural types (or of types of development of structures)." 3 Thus, typology of literary structures should be the peak of the structural analysis. One must point out, however, that efforts toward formulating a theory of literary types are not very common (in comparison with investigations in other fields of literary analysis). Only in recent years has the typology of literary structures received an essential stimulus. I have in mind the interpretation of structural types as logical models (or quasi-models), an interpretation which allows the introduction of the most progressive scientific methods, the methods of modelling, into structural analysis of literature. The interpretation of the literary type as an abstract model has been broached by K. F. Stanzel with regard to the novelistic types: "Novelistic types are to be understood as logical constructions which are never perfectly realized in literature, in its particular works, and which need not to be realized". 4 The suggestion of interpreting literary types as logical models, however, is far from being a satisfactorily elaborated methodology. There is no comparison in this respect between contemporary structural linguistics where the use of models is quite common, and contemporary literary analysis where modelling is rather a desideratum. One must add that a typology of literary structures does not mean a return to a positivistic taxonomy, forcing all the varied and complex literary phenomena into a limited number of clean-cut categories. Literary types when grasped as logical constructions do not exclude, but facilitate, the description of various transitional and 'mixed' phenomena. The representation of literary structures by means of modeltypes is dynamic, in contradistinction to the traditional static taxonomy. With regard to the principles of structural literary analysis just outlined, it is not difficult to predict that our analysis of point of view in fiction will start at the level ' Reprinted in: Michigan Slavic Materials, * Op. cit.

No. 2, 102.

544

LUBOMiR DOLEiEL

of formal, linguistic features which are presumed to express the distinctions of particular types of narration. A theoretical framework for the study will be provided by introducing the notions of two opposite types of texts: texts with speaker (S-text) and text without speaker (S-text). These text types, again, are to be understood as logical models which, of course, must be confronted with reality. Texts with speaker are characterized by three basic relationships: the relationship to speaker, the relationship to hearer, and the relationship to referent. This text type can be drafted as follows:

On the contrary, texts without speaker, lacking the relationship both to speaker and to hearer, display only the relationship to referent:

One must point out at the beginning that an important stimulus to the distinction of S-texts and S-texts can be found in Jakobson's classification of utterances with regard to the distinction of 'narrated event' and 'speech event'.5 It is not possible now to compare in detail Jakobson's system with the scheme here suggested. I will merely note, that only S-texts are 'speech events' in the proper sense, whereas S-texts are, in fact, 'speech events' approximating 'narrated events'. The notion of speaker is to be understood in the structural, immanent sense. The speaker is an intrinsic organizing and integrating factor of the text structure. The presence or absence of this organizing and controlling center as reflected in the text structure can be described by a set of binary linguistic features. In a purely logical respect, these features are analogous to the binary distinctive features which have been introduced by Jakobson to the representation of phonemic systems. However, whereas Jakobson's distinctive features form a homogeneous level of the linguistic description (sometimes called the merismatic level), binary features of the text structure belong to different linguistic levels. •

R. Jakobson, Shifters, Verbal Categories and the Russian Verb (1957).

THE TYPOLOGY OF THE NARRATOR

545

For our present purpose, the following binary features are the most important (the description of the features is given here only in a brief summary; for a fuller account see my book 0 stylu moderni ceske prozy, 1960): (1) S-texts display the system of three grammatical persons of verb (and also of pronouns); persons, as typical shifters, are used with reference to the participants of the 'speech event'. On the contrary, S-texts, where the reference both to the speaker and to the listener is nullified, the system of grammatical persons is reduced to only one person, the third. From the functional point of view this is, of course, a neutral 'zero' person. (2) Two important features result from the fact that S-texts are incorporated into the time-space system of the speaker, whereas S-texts are lacking such orientation. Biihler's 'origo', the temporal and spatial position of the speaker, is, of course, the centre of the time-space system in S-texts. (a) As an expression of the reference of the 'narrated event' to the temporal 'origo' of the speaker, a system of three verbal tenses is used in S-texts. On the contrary, in S-texts only the past tense is used as a non-referential tense characterizing the 'narrated event' simply as 'having been'. As far as the so-called historical present is concerned, it can be regarded as a stylistic variant of the narrative preterite, with the same non-referential function. (b) Temporal and local adverbs and demonstratives, such as: now, yesterday, this, here, there etc., are differentiated by their functions in S-texts and S-texts. In the S-type, their main function consists in 'demonstratio ad oculos* (Buhler), i.e., in expressing the reference to the time space position of the speaker. This function can be referred to as deixis. On the contrary, in the S-type, where 'demonstratio ad oculos' is impossible, the means of the above mentioned type refer to a certain 'intrinsic' temporal or spatial centre of the 'narrated event', determining, e.g., the temporal succession of motifs in an episode or the spatial organization of a landscape description. This function will be called elenxis. The situation is, however, complicated due to the fact that in S-texts both deixis and elenxis are possible. This distinctive feature is therefore to be formulated only with regard to the deixis, i.e., as presence or absence of deixis. (3) In S-texts which are characterized by the presence of a potential receiver, allocution is possible, i.e., direct contact with the receiver, such as address, appeal, question, etc., expressed by corresponding grammatical or lexical means. On the other hand, the S-type, lacking the reference to the receiver, is characterized by the absence of allocution and allocutional means. The lack of these means corresponds to the lack of the second grammatical person, as implied by the first feature. (4) The semantic part of the distinction between S-texts and S-texts presents a very complex problem. In general, S-texts are characterized by a subjective semantics, expressing the speaker's subjective relationships to 'narrated events'. In a particular case, with regard to the dramatic text, this subjective semantics has been described by V. V. Vinogradov as follows: "The dramatic word reproduces the reflex

546

LUBOMÎR DOLEÎEL

of reality in the consciousness of various characters. It expresses not only distinctions between things, but also distinctions between different points of view of acting characters". 6 We will speak of a subjective aspect in S-texts. Without going into detail, I will only enumerate the basic components of the subjective aspect: subjective factual evaluation, subjective emotional reactions, subjective modality and subjective causality. These, and perhaps other components of the subjective aspect, make the S-text a compact semantic structure controlled by the speaker's rational and emotional relationship to the 'narrated events'. On the contrary, the lack of the subjective aspect in S-texts leads to an 'objective' semantic structure which is directed only toward the expression of the 'narrated event' itself. (5) S-texts are characterized by use of specific stylistic means representing either speaker's idiolect or his social or local dialect. An idiosyncratic style, emerging from the synthesis of individual and social stylistic elements, and expressing the 'linguistic personality' of the speaker, is a characteristic quality of S-texts. On the contrary, S-texts are not differentiated nor specified in the stylistic respect; they are stylistically 'neutral'. We will refer to this feature as to the presence or absence of a personal style. One can suppose that other linguistic features distinguishing S-texts and S-texts can be introduced; for the present, however, the described set is sufficient to establish a framework for our analysis. In brief, the distinction between the S-type and the S-type can be presented in the following scheme:

S-text S-text

A

B

C

D

E

F

three persons

three tenses

deixis

allocution

subjective aspect

personal style

1 0

1 0

1 0

1 0

1 0

1 0

In this binary system of representation the operation of neutralization will be defined as follows: Neutralization is an operation which consists in mixing positive and negative values of the features in two ways: (1) some features of the set present positive, others negative values; (2) some features of the set present both positive and negative values. The two kinds of neutralization are described in the following scheme:

first kind second kind •

A

B

c

D

E

F

0 0

0 0

0 0

1 0/1

1 0/1

1 0/1

V. V. Vinogradov, Stil' PuSkina (Moskva, 1941).

547

THE TYPOLOGY OF THE NARRATOR

Applying our text types to the analysis of the structure of fiction, we have to make allowance for the fact that there are two kinds of speaker, two sources of verbal activity in the narrative structure: narrator and characters. (The definition of these two basic components of the narrative structure cannot be given within the framework of our scheme, these components belonging to the content level of the structural analysis). That means that we have to distinguish two subsets in the domain of S-texts: ^/S c -texts (characters' 'speech') S-texts" (k IV MeHWyHapOflHOMy Cl>e3Ay CJiaBHCTOB) (Mocraa, 1958), 52-53. " M. Swadesh, "The Phonemic Principle", Language, X (1934), 117-129. 28 G. L. Trager, "The Phonemes of Russian", Language, X (1934), 334-344. " L. Bloomfield, Language (London, 1961), chap. 13. 30 TCLP, VIII, 105-115.

DAS PROBLEM DER MORPHONOLOGIE

563

Morphemen vorkommen, ergibt sich die Notwendigkeit eine theoretische Grundform 3 1 jedes morphologischen Elements zu suchen. Die theoretische Grundform ist keineswegs eine der Realisationen; in Bloomfields Praxis ist es eine solche symbolische Darstellung, aus der es möglich ist durch Anwendung in einer bestimmten Reihefolge geltender Regel jede akustische Realisation des betreffenden Morphems abzuleiten. Die theoretischen Grundformen bestehen bei Bloomfield aus Morphophonemen, ihre aktuellen Realisationen aus Phonemen, wobei diese Phoneme als distinktive Werte aufgefaßt werden, die in der realen Aussprache vorhanden sind (§§ 5, 38). Das Morphonem ist also eine vorausgesetzte (heute würde man sagen: eine postulierte) Invariante auf der phonematischen, die theoretische Grundform auf der morphematischen Sprachebene. Die Idee der theoretischen Grundform übte auf die weitere Entwicklung der linquistischen Theorie einen mächtigen Einfluß aus. Um von den phonischen realisationen zu solch einer morphonologischen Darstellung zu gelangen sind gewisse Prozeduren notwendig. Die eigentliche Realisation (und ebenso die phonetische Transkription) enthält jedoch nicht alle dazu nötigen Informationen. Also zwischen der phonetischen Realisation (bzw. Transkription) und der morphonologischen Darstellung gibt es keine eindeutige Beziehung. Das erwies sich aber mit der bekannten Forderung in Konflikt zu sein, daß zwischen phonischer Realisation (d.h. auch ihrer phonetischen Transkription) und phonematischer Transkription eine ein-eindeutige Beziehung bestehen sollte, damit es möglich wäre nicht nur von der phonematischen Transkription den Wortlaut abzuleiten (was aufgrund Bloomfields theoretischer Grundform durch Anwendung kombinatorischer Regel möglich ist), sondern auch umgekehrt, aufgrund des Wortlautes eindeutig phonematisch zu transkribieren. Bloomfields Auffassung öffnet also den Weg zu so einer Definition der Einheit der phonematischen Sprachebene (gleichgültig ob wir sie Phonem oder Morphonem nennen), die unabhängig von einzelnen phonetischen Eigenschaften ihrer Realisationen wäre. 2.2.3. Aufgrund dieser zwei Konzeptionen — der traditionsmäßigen und der bloomfieldischen morphonologischen — wendete Roman Jakobson in seiner epochebildender Studie "Russian Conjugation" 3i gleichzeitig zwei Transkriptionen an: eine phonologische und eine morphonologische. Die morphonologische Transkription wird aus den verschiedenen Vorkommen desselben Morphems gewonnen und wird nur zur Darstellung bloßer Verbalkomponente ("bare verbal components", S. 155) verwendet, d.h. zur Darstellung isolierter Morpheme (Stämme, Affixe). Umgekehrt, die phonologische Transkription entspricht der Forderung der ein-eindeutigen Beziehung zur phonischen Realität und wird zur Darstellung tatsächlich gebildeter Formen verwendet. Auch durch diesen Unterschied

31

Ibid., 105; vgl. auch in Language, XIII, 9 "an artificial underlying f o r m " oder "a theoretical basis form". 82 Word, IV (1948), 155-167.

564

EUBOMfR DUROVIÖ

in der Transkription soll offensichtlich der Unterschied zwischen postulierten Werten und deren Realisationen angedeutet werden. Zum Unterschied von Bloomfield, für den die Morphonologie bloß eine Untersuchung des 'inneren Sandhi' ist (und die theoretische Grundform ist keinesweg eine der realen Alternanzen), gelangt Jakobson zu seinen Grundformen auf eine ganz andere Weise: wie das offensichtlich auch aus der Beschaffenheit der russischen Sprache hervorgeht (Betonung, Schwundvokale), verfolgt er die Alternationen der untersuchten Morpheme an allen deren Segmenten, also nicht nur im 'inneren Sandhi'. Als Grundform jedes phonemischen Konstituenten betrachtet er diejenige Alternante, die in einer Position steht, in der auch alle übrigen möglich sind. Mit anderen Worten: Jakobson scheidet die Resultate der Positionsveränderungen aus und identifiziert seine Grundform nicht direkt aus dem Signal, sondern aus allen möglichen Realisationen des betreffenden Morphems, wobei für jedes Segment die Position der maximalen Phonemunterscheidung (starke Position) gesucht wird. Daraus ist zu ersehen, daß durch Bloomfields Interpretation die Begriffe und Termini 'Morphonologie' und 'Morphonem' in die amerikanische Linguistik in einer ganz anderen Bedeutung hineingelangten, als sie von Trubetzkoy verwendet und von dem Prager linguistischen Zirkel im "Projet ..." kodifiziert wurden. In Trubetzkoys Benutzung, wie wir gesehen haben, ist die Morphonologie eine Untersuchung der Alternazen in den Allomorphen eines und desselben Morphems. In der amerikanischen Linguistik der Nachkriegszeit stellt Morphonologie (morphophonemics) vor allem eine gewisse Methode der Identifikation minimaler phonologischen Einheiten vor, die als postulierte Werte aufgefaßt werden (in der Regel wird der Terminus 'morphophonemes' oder 'morphonemes' gebraucht). Wie sich M. Halle darüber äußert: "... this is precisely the import of condition /3a/,33 for it sets up a distinction between phonemes and morphophonemes for the sole reason that the former can be identified on the basis of acoustic information alone, whereas the latter require other information as well." Solch ein Unterschied zwischen Phonemen und Morphonemen ist in der gegenwärtigen linguistischen Literatur geläufig geworden. Die funktionelle Unterscheidung dieser, in Jakobsons "Russian Conjugation" verwendeten zwei Vorgänge, macht der weiteren Entwicklung den Weg frei. Wenn es möglich ist, die "bare components" morphonologisch darzustellen, kann die morphonologische Transkription auch dann nicht ausgeschieden werden, wenn diese Komponente kombiniert werden, da jedes aktuelle phonetische Wort als Realisation eines oder mehrer solcher postulierter Werte aufgefaßt werden kann. Da es also möglich ist, nicht nur die "bare components", sondern auch den Morphembestand konkreter Wortformen morphonologisch darzustellen, muß sich dringend die Frage erheben, ob die morphonologische und phonologische Transkription nebeneinander einen Sinn haben, bzw. ob Phoneme und Morpheme nebeneinander zweckmäßig sind. 83

Die Bedingung, daß die phonologische Transkription nur von der im Signal enthaltenen Information ausgehen kann. M. Halle, The Sound Pattern of Russian ('s-Gravenhage, 1959), 24.

DAS PROBLEM DER MORPHONOLOGIE

565

2.2.4. Dieses Problem hat M. Halle in seinem Werk The Sound Pattern of Russian angeschnitten. D a er die phonologischen Einheiten als abstrakte theoretische Konstrukte betrachtet, findet er keinen Grund außer ihnen noch andere, durch ausschließlich analytische Operationen gewonnene Einheiten einzuführen. 34 Ausser diesem theoretischen Argument, das für jeden nicht völlig überzeugend sein muß, führt er auch noch praktische Gründe an. Er weist darauf hin, daß die sogenannte phonologische Transkription viel komplizierter ist und — was noch mehr überzeugend ist — daß sie redundant ist, d.h. man kann sie eindeutig von der morphophonematischen Transkription ableiten, während ein umgekehrter Vorgang nicht möglich ist. 35 Halle lehnt so das Phonem völlig ab und in seiner ganzen obenerwähnten Arbeit spricht er nur von den Morphonemen der russischen Sprache. Da sich jedoch in dieser Auffassung der Begriff 'Phonem' völlig verlor, hat auch die Unterscheidung zwischen den Phonemen und Morphonemen keinen Sinn mehr und wenn Halle von den Morphonemen spricht, geht er offensichtlich nur deshalb so vor, um eine terminologische Konfusion zu vermeiden. Ebenso gut könnte er von Phonemen sprechen. 2.3. Die mit Jakobsons (morphophonematischer) und Halles Transkription gewonnen Darstellungen sind im Prinzip identisch mit denen der Moskauer phonologischen Schule in ihrer klassischen Auffassung 36 und zwar ohne Rücksicht darauf, daß diese Schule sie erreichte, ohne daß das Problem der Morphonologie und des Morphonems angeschnitten würde. Im Gegenteil, der Begriff 'morphophonematische Transkription' (für denselben Vorgang, den man in dieser Schule immer 'phonematische Transkription' nannte) wurde hier erst in neueren Arbeiten Avanesovs eingeführt, 37 die ein Zugeständnis dem Sensualismus sind und in denen man als Novum (im Verhältnis z.B. zu dem ziterten Werk Avanesovs und Sidorovs) die sogenannte 'wortphonematische Transkription' einführt, die auf denselben Prinzipien beruht (ein-eindeutige Beziehung zum Signal), die Halle einige Jahre später (vgl. 2.2.4) als überflüssig ablehnt. Warum hat das Problem der Morphonologie für die Moskauer Schule nicht existiert? Obwohl die Moskauer Phonologen zweifellos von der Prager Phonologie inspiriert wurden, besonders von ihrer funktionellen Auffassung der Sprache, betrachteten sie das Phonem als einen Bestandteil des Morphems und identifizierten es nur über das Morphem, ohne Rücksicht auf die durch Neutralisationen hervorgerufene Identität oder Nichtidentität der einzelnen Laute. Die Folgerung dieser Konzeption war, daß hier das Problem der 'kombinatorischen Morphemveränderungen', niemals existiert hat. 84

Ibid., 23-24. M. Halles Besprechung des Buches P. H. ABaHecoB, @OHemum coepeMeHHoeo pyccKoeo AumepamypHozo n3biKa, Word, XVI (1960), 150-160. " Vgl. Fußnote 14 (das Buch von Avanesov und Sidorov). " P. H. ÄBaHecoB, ouemuKa coepeMemozo pyccxoeo jiumepamypHoeo H3um (MocKBa, 1956). 3S

566

EUBOMÌR DUROVIC

Und da die morphematische Einstellung der Moskauer Phonologie die Folgerung einer engen Bindung zwischen der phonematischen und morphologischen Ebene ist, ist es begreiflich, daß man hier die freien Alternationen immer als einen Bestandteil der Morphologie betrachtete. In der zitierten Avanesovs und Sidorovs grundlegenden Arbeit werden z.B. die Phonemalternationen in dem Einführungskapitel in die Morphologie, zusammen mit grammatischen Suffixen und der Betonung, als Mittel zum Signalisieren grammatischer Bedeutungen erörtert. 3. In den letzten Jahren wurde die Morphonologie zum Gegenstand der Kritik zweier bedeutender Sprachwissenschaftler. A. A. Reformatskij schrieb im Jahre 1955 die Studie "O COOTHOUICHHH (J>OHETHKH H RPAMMATHKH (M0paKTbi eme He oöjiaaaioT 3HaHHMocTbK), M o a c e T 6biTb npe/iocTaBjieHa M0p0H0Ji0rnH" (112). Hierher gehören 'historische Alternationen', wie z.B. 6ery/6eacHuib (r/ac) oder coH/cHa (o/0) u.a. Reformatskij sagt wörtlich: "B c a M O M itene, KaKoe fleno r p a M M a T H K e , KaK ynemiio o rpaMMaTHHecKHX 3 H a n e H H 5 i x H cpedcmeax ux ebipaotcenux [kursiviert von mir, II. D.] ao TaKHX £ and the Vedic type vi{.u There are two theories as to the character of this connection. 15 Some scholars (Kurylowicz, Pedersen), elaborating De Saussure's analysis of the Baltic pattern, trace the Slavic mobility back to the Vedic-Greek mobile paradigm with the stress alternating between the penultimate and the last syllable and with an end-stressed nom.sing. (Sanskrit pitá, pitáram, pitúh, pitre)-, others (e.g. Stang) prefer to see the source of Slavic mobility in the Greek-Vedic paradigm with the stress alternating between the first and the last syllable and an initially stressed nom.sing. (Sanskrit púmán, púmámsam, pumsáh). Since Slavic combines the second type of alternation with an end-stressed nom.sing., there must have taken place some kind of contamination. Stang writes: "there are so many characteristic agreements between the Baltic and the Slavonic accentual paradigm that there can be no doubt about their historical identity." 16 I rather suppose that the agreement can be explained by assuming for Baltic and Slavic a similar, but not identical development, starting from the same I.-E. pattern. I accept the main points of Pedersen's theory for Baltic. 17 In the Slavic verb I do not find any trace of I.-E. mobility (in spite of Stang's hypothesis, p. 112). Examples: /o'sa/, acc. /o'sáN/, but /nogá1/, acc. /no'gáN/ /no'sitej,

no'sjió[N],

no'sejp]i/,

aor.

/no'sit/,

-/-part,

/no'silá/,

part.perf.pass.

14 J. Kurylowicz, "Zasadnicze uwagi o kwestii intonacji stowiañskich", Sprawozdania Polskiej Akademii Umiejpnosci, 42/10 (1937), 281-284. 15 For the literature, see Stang, 177 f. 16 P. 64. 17 H. Pedersen, Etudes lituaniennes (Copenhagen, 1933), 21-44. Stang (12f.) rejects this theory. The reason is perhaps that Pedersen insists on calling his principal law "phonetic", which it is certainly not.

580

C. L. EBELING

/ n o ' s j e n a / , b u t /rodi'tej, r o d i o ' [ N ] , rode'i[s]i, rodi't, rod!'la, r o d j e ' n a / /neste'j, neso'[N], nese'[s]i, nese't, nesla', n e s e ' n a /

Stage A l . (a) Law of marginal oxytones. If in one paradigm x x1 and x x1 x, then x x1 x > x x x1. f x s t a n d s f o r all k i n d s o f s y l l a b l e s . ]

Examples: instr.pl. /nogamfi] 1 / (but /o'sam[i]/) instr.sg. A 0 /duHmo'mi/ *dymomb > /duHmomi 1 / /rodjo'fN], rodei[s]i', rodila'/ (but inf. and p.p.p. form separate nominal paradigms: /rodi'tej, rodje'na/) / n e s o ' [ N ] , nese[5]i'/ ( b u t / n e s e ' n a / , n o m . p a r a d i g m )

(b) Law of maximal contrasts. If in one paradigm x x1 and x x x' (but no x' x or x), then x x' > x' x. If in one paradigm x x', x x x' and x x x x' (but no x'...), then x x' > x1 x and x x x1 > x' x x. Examples: gen.sg. Ala/maHtere's/, instr.pl. /maHterim[i]'/ > A l b /ma'Hteres, maHterim[i]'/ /ne'so[N], nese[s]i'/, aor. /ne'set/, -/-part, /ne'sla/, but aor. /piHt/, -/-part. /piHla 1 / As to chronology, I do not find sufficient indications to differentiate (a) and (b). But if a chronological difference exists (as is assumed in the above examples), it is clear that (a) precedes (b). The only exception I have found is the imp.pl. A l b /ne'soite/. The reason of the special development of the imp. may be sought in the fact that the endings of the 1st and 2nd person pi. were in various periods interpreted rather as suffixes. Jakobson's theory of the agglutinative nature of the modern Russian imp. is well-known. Historical symptoms of this phenomenon are: (1) the sing. *xvali is a back-formation based on *xvalite;18 (2) after the phonetic diversification *-ois > *-i, *-oite > *-£te, the uniformity was restored in Russian (HCCH, HecHTe); (3) in R. CTaBbTe ( < *stavite) the first part has developed as if it were a separate word (cf. ind. CTaBHTe < *stavite), and likewise Polish stawmy, stawcie. The treatment of the accent in A l may be a fourth symptom. Prior to A l b the following morphological changes took place. (I) Confusion of nom. and acc. sing, in masc. -o-stems. Since the acc. had inherited a retracted accent, this accent thus became characteristic of the nom.-acc. 18

Chr. S. Stang, Das slavische und baltische Verbum (Oslo, 1942), 241.

HISTORICAL LAWS OF SLAVIC ACCENTUATION

581

(II) Hereupon came the confusion of masc. and neuter -o-stems in the sing. There can be established three types of neuter -o-stems at this time: (a) /pero'N/ ~ pi /pe'ra/, (b) /duo'roN/ ~ pi. /duora 1 /, a n d (c) / p o l i o ] N / ~ pi. /polia'/.

The contrast between sing, and pi. of types (a) and (b) was old. It must have been absent in neuters with the suffix *-io- with originally adjectival or abstract meaning (e.g. *aje 'egg', orig. meaning 'bird's'), and in words where a pi. was hardly used, such as /meNso'N/ *mgso (type c). Type (b) was in the sing, identical with the masc. columnal type, and the nom.-acc. had the same accent as all masc. -o-stems. The accented ending /-O'N/ of (a) and (c), however, was typically neuter. As a result, the sing, of (b) altered its gender, while (a) and (c) replaced their ending by another typically neuter ending, viz. pronominal *-od. Henceforth type (b) lived on as a suppletive paradigm with a masc.sing. and a neuter plural. I prefer this old solution to the one given by Illic-Svityd (pp. 131 ff.), although I admit that the latter's explanation is in some respects more attractive. IlliC-SvityC holds that stressed /-O'N/ developed into -o, unstressed /-ON/ into -6,19 and unstressed [and stressed(?)] /-¿ON/ into -ie. This view has the following undesirable consequences: (1) Phonetically, there are no other examples in the oldest history of Slavic where the place of the stress would be responsible for a different development of a vowel (cf. especially gen.pl. *-on). (2) Chronologically, the theory contains two untenable points. First, IlliC-Svityc is forced (p. 132) to postpone the fusion of the singular of old barytone neuters with masculine nouns until after the passing of the root-stressed non-acute masc. -o-stems to the mobile type (stage A6 below, the first accentual development which did not extend over the whole Slavic territory), for otherwise old neuters of the type *dvon, too, would have joined the mobile type. Second, Il!ic-Svity£ places the liquidation of the final nasal after the change of vowels caused by a preceding j (as in pol'e). (3) Illic-SvityS does not explain the different development of neuters with originally long and short vowel in the root, e.g. *m%so and *pero: he ascribes it simply to quantity (pp. 133 and 152f.), but this runs counter to the other results of his study, according to which oxytone paradigms, both with originally long and short vowels, become mobile in Slavic. The explanation of *mqso, *testo, *pivo, *seno, *zolto as singularia tantum has the advantage that it can also account for *prdso, which behaves alike. In my conception, morphological change (II) presupposes the accentuation /poljo'N/, and is therefore anterior to A l b . 19 The passing of a group of neuters to the masc. gender, caused by the change *-on > proves that the assumption of I.-E. *-o for the nom.acc. of neuter -o-stems (see, e.g., F. V. Mares, "Vznik a rany vyvoj slovanske deklinace", Ceskoslovenske prednasky pro V. Mezindrodni Sjezd Slavistu v Sofii (Prague, 1963), 55) cannot be correct.

582

C. L. EBELING

Stage A2. Law of Hirt/Illi£-Svity£. xH x1 > x1 x The formulation of this law, established by Hirt, was later modified several times, by Pedersen, Mikkola, Lehr-Splawinski, Bonfante. According to the reinterpretation of Illi£-Svityc (pp. 78-81, 156), one must distinguish three types of long syllables: xH(S), i.e. vowel + laryngeal ( + sonant), e.g. e < eH, i < iH, ou < oHu; xSH, i.e. e- or o-grade of vowel + sonant + laryngeal, e.g. ou < ouH (= oua); e/d(S), i.e. lengthened grade of vowel ( + sonant), e.g. e (: e), ou (: oy). Illiii-SvityS assumes that in Baltic and Slavic a root-syllable with a long vowel of the type xH(S) attracted the ictus, e.g. I.-E. *griud(< *griHud) > SI. *griva (acute, columnal). The law does not explain -u-stems, such as *sym. Illii-Svityc, referring to Meillet20 for the fact that -M-stems with I.-E. long root vowel show in Slavic as a rule mobility with metatony, supposes that the type became barytone as a consequence of the law in question, and that the mobility is secondary. However, I fail to see the motivation for this change from barytone to mobile. More important are the difficulties presented by the verb: since Illi£-Svityc obviously wants the law to be phonetic, it should work also in other than nominal forms, whenever the phonetic conditions are present. However, verbs of the type *piH- 'to drink' do not conform to the law, cf. S.-Cr. inf. piti, aor. pih, pi, -I-part. plo, pila. The 3rd p. sg.aor. is especially important: it can not be accounted for as a product of analogy, because there is no model. It must be an old root-aorist. 21 As such, it is a remnant of an older paradigm whose other members were replaced by -i-aorists built upon the infinitive. With a view to such cases I amplify the law as follows: xH attracts the ictus from an immediately following syllable; in paradigms with a phonetic retraction in the majority of the forms, it was analogically extended to the other forms as well, while other paradigms remained mobile. Became columnal: -o-stems, with four trisyllabic cases: instr.sg., dat.loc.pl., dat.dual; -¿-stems, with four trisyllabic cases: dat.instr.loc.pl., dat.dual (instr.sg. had /-aN/).22 Remained mobile: -«-stems, with eight trisyllabic oxytone cases: dat.instr.sg., gen.dat.instr.loc.pl., gen.dat.dual (nom.pl. *synove had probably initial stress); oxytone pres./aor. Examples: phonetically /sunumi 1 /, /gru'zena/, /gruze[s]i'/; analogically /du'momi/, /gri'uamfi]/. 20 21 22

See fn. 5. Stang, ... Verbum (see fn. 18), 65. See, e.g., Mares, o.c. (fn. 19), 59.

HISTORICAL LAWS OF SLAVIC ACCENTUATION

583

A 2 is certainly posterior to A l a , f o r many trisyllabic oxytones appeared only then. That it is also after A l b may be shown by the development A l a gen.sg. /maHtere's/ > A l b /ma'Hteres/ > A 2 /mà'teres/ (columnal paradigm). Stage A3. Quantity > pitch. x1 > x ' ; x ' i > x 'i; x'i > x-'j; x1 = x1 [x = short vowel,

• = quantitatively

neutral, ' = rising, ' = falling,

1

= intona-

tionally neutral stress, j = any second component o f a diphthong.] A s a consequence o f the neutralization as to quantity o f the first components o f diphthongs, the intonations become relevant in stressed syllables. This development does not create long falling monophthongs. The potential opposite o f x', i.e. x\ is supplied only by new formations.

Lehr-Splawinski 23 pointed to the existence o f a

whole group o f feminine words which, surprisingly, in later Slavic are columnally ending-stressed: trava, xvala, krasa, skala, dyra, zara.

Neither starting f r o m I.-E.

mobile-oxytones, nor f r o m barytones, do the rules o f Slavic accentuation produce such forms if the root had originally a long vowel. It follows that Slavic has in these words secondarily lengthened grade, e.g. trava, cf. OChSl. trovç, skélti.2i

skala, cf. Lith.

Accordingly, the list o f complete I.-E./Slavic correspondences collected by

Illiô-Svityc does not contain any noun o f this kind. Especially in roots, lengthened grade was rare in I.-E. 25

Lehr-Splawinski mentions further the masc. columnal

oxytones R. 6mk, ôbixà and n/iauf, tiAaufâ, and he adds that these are certainly new formations (p. 94, fn.). S.-Cr. bak.

The former is represented by a concurrent form

*bbkb,

A s to the latter, Lehr-Splawinski reminds o f Czech plaskylplosky,

Ukr.

naacKuûjmocKm. I conclude that a new type /trâ'uâ/ originated after A 3 .

A s to the ancestor o f

6BIK, it can have got its long vowel only after A 6 , because otherwise the word would have become mobile. A3 is after A2, f o r syllables o f the type x H , if accented, are always acute, and, consequently, were unaffected by A3, which divided diphthongs in acute and circumflexed ones. Examples : /dû'mâN/ (columnal), /su'nuN/ (mobile) A 2 /gôluà1, gô'luâN, borda 1 , bo'rdâN/ > A3 /gâ-luâ', gâ-'luâN, bâ-rdà', bâ-'rdâN/ Stage A4. Homogenization o f diphthongs, x-i > x° T h e diphthongs in /-N, -j, -u/ are homogenized, but retain a feature o f "breaking", which I indicate with

e.g. /ï°, ê°, ç°/. Later stages show traces o f "breaking" only

23 T. Lehr-Spiawinski, "Najstarsze praslowiariskie prawo cofania akcentu", Symbolae grammaticae in honorem I. Rozwadowski, II (Cracow, 1927), 85-100 (p. 93). 24 J. Kurylowicz, "Le degré long en slave", RS, 16 (1950), 1-14. 25 A. Meillet, introduction à l'étude comparative des langues indo-européennes (Paris, 1934'), 154,157.

584

C. L. EBELING

in syllables which were post-tonic at stage A4; therefore I mark it in these positions only. However, after nasal vowels, where it has no relevance, I do not write it. Examples : /na'siti°, na'sjQC), na'si°[s]i/ *nositi, noso, nosisb I have shown elsewhere 26 that this process (there: stages VI and VII1/IX 27 ) is later than the rise of distinctive intonation A3 (there: stage II). Stage A5. (a) Reshuffling of mobile paradigms. If in one paradigm x1 (x) and x x1, then x x1 > x1 x, unless the final accent is motivated because it helps avoiding homonymy. As a result, in the paradigm of *bos- all bisyllabic forms are stressed on the first syllable, with the exception of : /basu 1 / gen.sg.fem., cf. /ba'su 0 / nom.-acc.pl.fem. /bas«?1/ instr.sg.fem., cf. /ba'sç/ acc.sg.fem. /basé 1 / loc.sg.fem., cf. /ba'së 0 / dat.sg.fem. /basa 1 / nom.acc.pl.neuter, cf. /ba'sâ/ gen.sg.masc.neuter /basu 1 / gen.pl., cf. /ba'su/ nom.sg.masc. /basù 1 / instr.pl.masc.neuter, cf. /ba'su 0 / acc.pl.masc., and, possibly, the gen.loc.dual. Cp. the -/-part, /pi'lu, pila'/ *pih, pila. (b) Meillet's law. If in one paradigm x' x and x x1, then x' x > x' x. Examples : /su nu/ (but /du'mu/) /ga-'lvç, ba-'rdç/ (but /va'dç>/ *vodç with neutral pitch) (c) Retraction of the ictus to a preposition from a barytone form of a mobile paradigm. A long syllable gets falling pitch, e.g. /na' vad