This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
TITLE "The Circum-Baltic Languages: Typology and Contact"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
The Circum-Baltic Languages
Studies in Language Companion Series (SLCS) The SLCS series has been established as a companion series to Studies in Language, International Journal, sponsored by the Foundation “Foundations of language”.
Series Editors Werner Abraham
Michael Noonan
Universities of Groningen, Berkeley and Vienna
University of Wisconsin-Milwaukee USA
Editorial Board
Joan Bybee
Christian Lehmann
University of New Mexvico
University of Erfurt
Ulrike Claudi
Robert Longacre
University of Cologne
University of Texas, Arlington
Bernard Comrie
Brian MacWhinney
Max Planck Institute, Leipzig
Carnegie-Mellon University
Marianne Mithun William Croft
University of California, Santa Barbara
University Manchester
Edith Moravcsik
Östen Dahl
University of Wisconsin, Milwaukee
University of Stockholm
Masayoshi Shibatani
Gerrit Dimmendaal
Kobe University
University of Leiden
Russell Tomlin
Martin Haspelmath
University of Oregon
Max Planck Institute, Leipzig
John Verhaar
Ekkehard König
The Hague
Free University of Berlin
Volume 54 The Circum-Baltic Languages: Typology and Contact Edited by Östen Dahl and Maria Koptjevskaja-Tamm
The Circum-Baltic Languages Typology and Contact
Edited by Östen Dahl Maria Koptjevskaja-Tamm Stockholm University
John Benjamins Publishing Company Amsterdam/Philadelphia
8
TM
The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.
Library of Congress Cataloging-in-Publication Data Circum-Baltic languages : Typology and contact / edited by Östen Dahl and Maria Koptjevskaja-Tamm. p. cm. (Studies in Language Companion Series, issn 0165–7763 ; v. 54–55) Includes bibliographical references and index. Contents: v. 1. Past and present -- v. 2. Grammar and typology. 1. Baltic Sea Region--Languages. I. Dahl, Östen. II. Koptjevskaja-Tamm, Maria. III. Series. P381.B33 C57 2001 491’.9--dc21 isbn 90 272 3057 9 (Eur.) / 1 58811 020 6 (US) (Hb, v. 1; alk. paper) isbn 90 272 3059 5 (Eur.) / 1 58811 042 7 (US) (Hb, v. 2; alk. paper)
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Table of contents
List of abbreviations List of contributors
vii x
Introduction The Circum-Baltic Languages: Introduction to the volume Östen Dahl and Maria Koptjevskaja-Tamm
xv
Part 1 Survey of selected Circum-Baltic languages and language varieties The Latvian language and its dialects Laimute Balode and Axel Holvoet The Lithuanian language and its dialects Laimute Balode and Axel Holvoet Russian varieties in the southeastern Baltic area: Urban Russian of the 19th century Valeriy Cˇekmonas
3
41
81
Russian varieties in the southeastern Baltic area: Rural dialects Valeriy Cˇekmonas
101
Swedish dialects around the Baltic Sea Anne-Charlotte Rendahl
137
The Finnic languages Johanna Laakso
179
Part 2 Early history of the Circum-Baltic languages The origin of the Scandinavian languages Östen Dahl
215
Baltic influence on Finnic languages Lars-Gunnar Larsson
Part 3 Contact phenomena in minor Circum-Baltic languages The role of language contact in the formation of Karelian, past and present Stefan M. Pugh
257
Syntactic code-copying in Karaim Éva Ágnes Csató
271
Yiddish in the Baltic region Neil G. Jacobs
285
The North Russian Romani dialect: Interference and Code Switching Aleksandr Yu. Rusakov
313
On some Circum-Baltic features of the Pskov-Novgorod (Northwestern Central Russian) dialect Valeriy Cˇekmonas
339
Name index
i 1
Language index
i 9
Subject index
i 15
AUTHOR ""
TITLE "List of abbreviations"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
List of abbreviations
Abbreviations used in glosses abess abl abs acc acnnr act adess adj adjr adv advr all aux com comm comp cond conneg conv cop dat deb def dem det dim dir du dur elat
abessive ablative absolutive accusative action nominalizer active adessive adjective adjectiv(al)izer adverb adverbializer allative auxiliary comitative common (gender) complementizer conditional form of the verb used with negation converb copula dative debitive definite (article) demonstrative determiner diminutive direct dual durative elative
emph erg ess fem freq fut gen ger hab hon ill imp impf impr indef indir iness inf inst instruc int intr ipfv ips loc masc minf n neg nfin nom
nonpast nonpast nr nominalizer neut neuter obj object obl oblique opt optative part participle pass passive pf perfect pfv perfective pl plural po partial object poss possessive pot potential pp perfective particle ppa present participle active ppp past/preterit passive participle
pref prep pret pres prtv past refl sub sg subj suff sup term tinf to trnsl zerik
prefix preposition preterite present partitive past reflexive subject singular subjunctive suffix superlative terminative Estonian ta-infinitive total object translative Basque “zerik”-case
FSwd Gal Grg Grk Grm HGrm Hng Ice Ing Ir It Jat Kar Kom Krm Ksh Lat LGrm Lith
Finland Swedish Galindian Georgian Greek German High German Hungarian Icelandic Ingrian Irish Italian Jatvingian Karelian Komi Karaim Kashubian Latin Low German Lithuanian
Additional abbreviations Abx Arm Balt Blg BY Bylr CourlY Cur CY Cz Dal Dan Dut Eng ErzaMrd Est EY Fin Fr
Abkhaz Armenian Baltic Bulgarian Baltic Yiddish Belarusian Curonian Yiddish Curonian Central Yiddish Czech Dalecarlian Danish Dutch English Erza Mordvin Estonian: Eastern Yiddish Finnish French
List of abbreviations
LivK LivSal
Livonian in Curonia Salis-Livonian, Livonian in Vidzeme (near Salis). LRmn Latvian Romani Ltg LatgalianLtv Latvian Lud Ludian Mar Mari MarEast Eastern Mari Mrd Mordvin NEst Northern Estonian NEY Northeastern Yiddish NRRD North Russian Romani dialects Nsam North Sami OHGrm Old High German OLith Old Lithuanian Olo Olonetsian ONrs Old Norse OPrs Old Prussian Oss Ossete PC partitive nominal construction Plb Polabian
Pol PPC Rmn Rus SAE Sam SCr SEst SEY Slve Spn StY Swd SY Ttr Udm Ukr Vot Vps WY Yid ZY
Polish pseudo-partitive nominal construction Romani Russian Standard Average European Sami Serbian/Croatian South Estonian, Southeastern Yiddish Slovenian Spanish Standard Yiddish Swedish Southern Yiddish Tatar Udmurt Ukrainian Votian Veps Western Yiddish Yiddish zameter yidish, Samogitian Yiddish
ix
AUTHOR ""
TITLE "List of contributors"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
List of contributors Östen Dahl, [email protected] Dept. of Linguistics, Stockholm University, SE-106 91, Stockholm, Sweden Maria Koptjevskaja-Tamm, [email protected] Dept. of Linguistics, Stockholm University, SE-106 91, Stockholm, Sweden Laimute Balode, laimute.balode@helsinki.fi Dept. of Slavic and Baltic Languages and Literatures, P. O.Box 4, Vuorikatu 5B, FIN-00014 Helsinki, Finland Axel Holvoet, [email protected] Lietuviu kalbos institutas, Antakalnio 6, 2055 Vilnius, Lithuania Valeriy Cˇekmonas, [email protected] Dept. of Slavic languages and literatures, Vilnius University, Krokuvos 1–24, LT-2005 Vilnius 5; Lithuania Anne-Charlotte Rendahl, [email protected] Dept. of Linguistics, Stockholm University, SE-106 91, Stockholm, Sweden Johanna Laakso, [email protected] Institut für Finno-Ugristik der Universität Wien, Spitalgasse 2–4, Hof 7, A-1090 Wien, Austria Lars-Gunnar Larsson, Lars-Gunnar.Larsson@finugr.uu.se Dept. of Finno-Ugrian languages, University of Uppsala, Box 256, SE-751 05, Uppsala, Sweden Stefan Pugh, [email protected] Dept. of Slavic Languages and Literatures, Duke University, Durham, NC 27 706, UK Éva Á. Csató Johanson, [email protected] Department of Asian and African Languages, Uppsala University, Box 527, SE-751 20 Uppsala, Sweden
List of contributors
Neil G. Jacobs, [email protected] Department of Germanic Languages and Literatures. The Ohio State University, 314 Cunz Hall, 1841 Millikin Road, Columbus, OH 43210, USA Alexandr Rusakov, [email protected] Dept. of General Linguistics, St. Petersburg University, Lenin str. 11, apt. 4, 197136, St.Petersburg Russia Vytautas Ambrazas, [email protected] Kraziu˛ g. 7–3, 2001 Vilnius, Lithuania Bernhard Wälchli, [email protected] Dept of Linguistics, Stockholm University; and Inst. für Allgemeine Sprachwissenschaft, Universität Bern, Länggassstr. 11, CH-3012 Bern Helle Metslang, [email protected] Tallinn Pedagogical University, Chair of Estonian, Narva Road 29, 10120 Tallinn, Estonia Baiba Metuzale-Kangere, [email protected] Dept. of Baltic Studies, Stockholm University, SE-106 91, Stockholm, Sweden Simon Christen, [email protected] Stadtbachstr. 42a, CH-3012 Bern, Switzerland Leon Stassen, [email protected] Instituut voor Algemene Taalwetenschap, Erasmusplein 1 (k. 9–12), NL-6525 GG Nijmegen, The Netherlands Thomas Stolz, [email protected] Universität Bremen, Fachbereich 10: Linguistik Postfach 33 04 40, D-28334, Bremen, Germany Kersti Boiko, [email protected] Faculty of Modern Languages, Finno-Ugric Programme, University of Latvia, Visvalza 4a, Riga, LV-1050, Latvia
xi
DOCINFO
AUTHOR ""
TITLE "Introduction"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Introduction
AUTHOR "Östen Dahl and Maria Koptjevskaja-Tamm"
TITLE "The Circum-Baltic Languages"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
The Circum-Baltic Languages* Introduction to the volume Östen Dahl and Maria Koptjevskaja-Tamm
Areal linguistics and typology have multiple connections. The use of typology has long been recognized in areal linguistics: typological considerations are an important tool in evaluating suggested isoglosses, i.e. if typologically marked (crosslinguistically infrequent, idiosyncratic) features are attested in several neighbouring languages, this similarity can hardly be attributed to an accident. Typologists, on the other hand, have shown an increasing interest in areal generalizations: while some typologists try to minimize the risk for a potential areal (and genetic) bias in a sample by various sophisticated sampling procedures, others find uneven areal distributions of features a fascinating object of study: the most important recent contributions to areal typology include Dryer’s work on word order (1989), Nichols’ work on correlations among marking type and various other language parameters (1992), Dahl’s work on tense and aspect (1995), Stassen’s book on intransitive predication (1997). In practice, however, the contacts between areal linguistics and typology, as well as communication between experts in these two fields, often are weak. One basic problem is, of course, that both types of research require hard work with a huge amount of data, but with completely different focuses. For areal linguists, the desideratum is a complete documentation — both synchronically and diachronically — of linguistic properties in a restricted area, including minimal variation among closely related language varieties. Typologists, who sometimes find this preoccupation with details boring, dispense with a large portion of them — much to the irritation of areal linguists, who find this attitude superficial and suspicious. Practical difficulties are also an obstacle in contacts between areal linguistics and typology. Areal studies require enormous knowledge — of the languages spoken in the area and of the linguistic literature concerning their synchronic state and history. One can hardly expect areal linguists to have a good orientation in linguistic typology. On the other hand, one cannot expect that typologists will be able to have a good orientation in descriptions of various phenomena which appear or have appeared in small publications at various “obscure” places, are written in “smaller” languages and are directed towards a narrow circle of specialists.
xvi The Circum-Baltic Languages
The area around the Baltic sea provides ample illustration of these difficulties. Not only are there many languages and many dialects spoken here, but there are also several strong local linguistic traditions, some of them with fairly old roots. Making the information accumulated within those traditions more easily accessible to linguists outside them is a highly desirable but not an easily attainable goal. Recent developments in linguistics have brought about more intensive contacts with typologists and specialists in particular languages. Thus, the combined efforts of numerous linguists of different denominations involved in the programme “EUROTYP” resulted in many interesting generalizations about Europe as a linguistic area. The end of the cold war and new media of scientific communication paved a way for new fruitful dialogues among various specialists. There are still seemingly unsurmountable difficulties. General typologists and specialists in Baltic, Finno-Ugric, Slavic and Germanic languages will all perceive the goals of the areal study of CB-languages differently. The present pair of volumes represents the first major attempt to reconcile those differences. Our topic, then, is Circum-Baltic languages — the languages spoken around the Baltic Sea. Obviously, the delimitation of this set of languages will have to remain vague, for several reasons. The first source of vagueness resides in the
Northern Sami
Inari Sami
Skolt Sami
Lule Sami Pite Sami Ume Sami
Karelian
Southern Sami
Finnish Olonetsian
Dalecarlian
Norwegian
Swedish Estonian Livonian
Latvian Danish Northern Frisian Low German
Ludian Veps
Ingrian Votian
Russian
Lithuanian Kashubian
Karaim
Belarusian Polish High German
Map 1.The Circum-Baltic languages. Non-territorial languages not shown: Romani varieties, Yiddish, Tatar.
The Circum-Baltic Languages xvii
preposition “around”. If a language is spoken on the coast of the Baltic, the case is clear, but how far from there should we go? Then, some languages may not have native speakers in the area but are still important for the study of language contacts, for instance, Latin and French. Should they be included? Finally, to make a list of the Circum-Baltic languages we need to draw the borderline between languages and dialects or varieties, a notoriously hopeless task. The list of Circum-Baltic languages given below, and the map of the area (Map 1), are therefore somewhat arbitrary, and should be taken only as a starting-point for the discussion.
Organization of the volumes The first of the two volumes — Circum-Baltic Languages: Volume 1 — Past and Present — surveys important sub-groups in the present-day Circum-Baltic languages, placing them in their geographical, historical and societal setting and discussing specific contact situations. The second volume — Circum-Baltic Languages: Volume 2 — Grammar and typology — focuses on grammatical phenomena in the Circum-Baltic languages, relating them to the larger typological perspective. Each of the volumes contains three sections. The first section of the first volume contains overviews of four subsets of Circum-Baltic languages and language varieties, representing all the major languages families in the area. Latvian and Lithuanian and their dialects are presented in two chapters by Laimute Halmode and Axel Holvoet. Johanna Laakso and AnneCharlotte Rendahl give surveys of the Finnic languages and the CB Swedish dialects, respectively. Finally, Valeriy Cˇekmonas discusses Russian dialects in the CB area in two chapters. The second section is devoted to the early history of the CB languages. Östen Dahl discusses the origin of the Scandinavian languages and Lars-Gunnar Larsson the influence of the Baltic languages on the Baltic Finnic languages. The third and last section of the first volume treats contact phenomena in some of the minor (in terms of number of speakers) CB languages and language varieties. Karaim, a Turkic language spoken by a small group in Lithuania, is treated by Éva Ágnes Csató. The formation of Karelian, a Finnic language spoken in the Karelian Republic (Russia) is discussed by Stefan M. Pugh. Neil Jacobs surveys the varieties of Yiddish in the CB region and Aleksandr Yu. Rusakov discusses interference and code switching in the variety of Romani spoken in Northern Russia, and Valeriy Cˇekmonas looks for contact-induced phenomena in the Pskov-Novgorod dialect of Russian. The first and largest section of the second volume comprises six chapters, which all treat grammatical phenomena in the languages east of the Baltic from the point of view of diachronic development and areal influence. Three of them focus on
xviii The Circum-Baltic Languages
nominal case: Simon Christen discusses different syntactic positions in which the genitive case may appear in the Baltic and Finnic mentioned and Baiba Metuza¯leKangere and Kersti Boiko compare the case systems of Latvian and Estonian. Vytautas Ambrazas concentrates on a more specific diachronic development: how the use of the nominative for object marking arose in the eastern CB area. The contributions of Helle Metslang and Bernhard Wälchli both treat the historical development of the use of verb particles for aspect or Aktionsart marking in Estonian, Latvian and Livonian. The three chapters in the following section also treat grammatical phenomena, but from a more explicitly typological point of view. A shared focal point of the chapters is the role of nominal case in various syntactic constructions: Leon Stassen’s chapter with the role of cases such as the instrumental, essive and translative in nonverbal predication, Maria Koptjevskaja-Tamm’s with partitive and pseudo-partitive constructions, and the chapter by Thomas Stolz with the expression of comitative and instrumental roles. In the concluding chapter of the second volume, Maria Koptjevskaja-Tamm and Bernhard Wälchli survey a number of important features of CB languages, arguing that although the notion of a Sprachbund is not satisfactory for characterizing the linguistic situation in the CB area, the study of the CB languages from an areal-typological point of view reveals a linguistic landscape with many interesting properties of its own.
List of Circum-Baltic languages Germanic West High German (HGrm) Low German (LGrm) Yiddish (Yid) North Danish (Dan) Swedish (Swd) Dalecarlian (Dal) Norwegian (Nrw) Baltic West †Old Prussian (OPrs) ††Curonian (Cur) †Jatvingian (Jat)
The Circum-Baltic Languages xix
Central Lithuanian (Lith) Latvian (Ltv) East ††Galindian (Gal) Slavic West Polish (Pol) Kashubian (Ksh) †Polabian (Plb) East Belarusian (Bylr) Russian (Rus) Ukrainian (Ukr) Indo-Aryan Romani (Rmn) with varieties/sub-languages: Kelderash, Lovari, Kalo, Baltic, North Russian Finno-Ugrian Finnic Veps (Vps) Karelian (Kar) Olonetsian (Olo) Ludian (Lud) Finnish (Fin) Ingrian (Ing) Votian (Vot) Estonian (Est) with varieties/sub-languages: South Estonian, Northern Estonian (NEst) Sami (Sam) with varieties/sub-languages: Southern Sami, Ume Sami, Pite Sami, Lule Sami, Northern Sami, Inari Sami, Skolt Sami Turkic Karaim (Krm) Tatar (Ttr) † = extinct; †† = only onomastic sources and substratum In addition, more or less isolated dialects under strong influence of other languages, e.g. *Leivu (a Hargla Estonian dialect between Aluksne and Gulbene/Latvia), *Krevinian (Votian near Bauska/Latvia), Estonian Swedish, Nehrungskurisch, Latgalian, Russian of the Old Believers in the Baltics, Urban Russian in the Baltics,
xx
The Circum-Baltic Languages
Baltendeutsch, Halbdeutsch.
Note * In 1991, a six-year research program called “Language Typology around the Baltic Sea” was launched by the Faculty of Humanities at Stockholm University, with Maria Koptjevskaja-Tamm as main researcher. A large part of the work behind these volumes was supported financially within this research program. We want to express our thanks here both to our sponsors and to all the people who have contributed to the volumes.
References Dahl, Östen. 1995. “Areal tendencies in tense-aspect systems”. In: Bertinetto, Pier Marco, Bianchi, Valentina, Dahl, Östen & Squartini, Mario (eds.), Temporal Reference, Aspect and Actionality. Vol. 2: Typological perspectives, 11–28. Torino: Rosenberg & Sellier. Dryer, Matthew. 1989. “Large linguistic areas and language sampling”. Studies in Language 62: 808–45. Nichols, Johanna. 1992. Linguistic Diversity in Space and Time. Chicago and London, The University of Chicago Press. Stassen, Leon. 1997. Intransitive predication. Oxford studies in typology and linguistic theory. Oxford: Clarendon Press.
DOCINFO
AUTHOR ""
TITLE "Survey of selected Circum-Baltic languages and language varieties"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Part 1
Survey of selected Circum-Baltic languages and language varieties
AUTHOR "Laimute Balode and Axel Holvoet"
TITLE "The Latvian language and its dialects"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
The Latvian language and its dialects* Laimute Balode and Axel Holvoet
General data 0.1 Number of speakers Latvian is now spoken by about 1,340,000 people in Latvia, and by about 110,000 people outside Latvia (though some sources estimate the number of speakers outside Latvia at about 300,000). No figures are available as concerns the number of representatives of the individual dialects. In 1935, Latvians constituted 77% of the population of Latvia. The largest minorities were: Russians (8.8%), Jews (4.9%), Germans (3.3%), Poles (2.5%) and Belarusians (1.4%). In 2000, 57.6% of the population consisted of Latvians, and the largest minorities were: Russians (29.6%), Belarusians (4.1%), Ukrainians (2.7%), Poles (2.5%) and Lithuanians (1.4%). The Livonians, a Finnic population once inhabiting most of the coastal regions of Latvia, have been largely assimilated. At the close of the 19th century, some 3,000 people spoke Livonian, but their number has now dwindled to less than a hundred. 0.2 Territorial division Latvia is traditionally divided into four regions: – – – –
Kurzeme (Courland proper, the western part of the former Duchy of Courland); Vidzeme (the western part of Livonia, former Swedish Livonia); Latgale (Letgalia, the eastern part of Livonia, former Polish Livonia); Zemgale (Semigalia, the eastern part of the former Duchy of Courland, around Jelgava, Germ. Mitau, and further to the east); the easternmost part of Semigalia, stretching in a narrow strip south of the Daugava, is sometimes singled out and referred to as Upper Courland (Augškurzeme) or Selonia (Se¯lija).
4
Laimute Balode and Axel Holvoet
Tamian dialects Livonian dialects
Central dialects High Latvian dialects
Curonian dialects Semigalian dialects
1.
The Latvian language
1.1 Latvian among the Baltic languages When compared to Lithuanian, with which it constitutes the eastern branch of the Baltic languages, Latvian appears to have gone through a more rapid development. For the common Eastern Baltic ancestor language we may reconstruct sound shapes and forms quite similar to those of modern Lithuanian. The most important features which give Latvian its present-day appearance are the following: 1.1.1 The original mobile stress has been replaced by an immobile stress on the first syllable, cf. Lithuanian metù m«etam(e)
Latvian [mjæÁtu] [«mjæ˜tamjæ]
metu metam
meaning [Ámæt˜u] [Ámætam]
‘throw’ (pres.1sg) ‘throw’ (pres.1pl)
All vowels in final syllables, whether open or closed, have undergone some kind of shortening or reduction; this may involve:
The Latvian language and its dialects
a.
frequent loss of original short vowels, usually retained in Lithuanian:
Lithuanian m«eta m«etat(e) naktìs
Latvian [«mjæ˜ta] [«mjæ˜tatjæ]/[«mjæ˜tat] [nakjÁtjis]
b. shortening of long vowels and diphthongs (Lithuanian has retained part of the old word-final long vowels and diphthongs, though shortening occurs here as well): di«enos a«kys linaı˜ lieta«us
Original sequences of the type /V + N/ (V representing any vowel and N any nasal sonorant) have undergone changes before a consonant or word-finally. They have been replaced with long vowels or diphthongs (probably through an intermediary stage of nasalized vowels): kri«nta penkì
[«kjrjin.ta] [pjænjÁkji]
krı¯t pieci
[kri˜t] [Ápietësi]
‘fall’ (pres.3) ‘5’ (nom.masc)
For consonants, the most striking feature is the change of palatalized velars into dental affricates (probably through the intermediary stage of palato-alveolar affricates): kèpti gérti
[Ákjæpjtji] [´gjæ˜rjtji]
cept dzert
[tësept] [dëzert]
‘bake’ (inf) ‘drink’ (inf)
1.1.2 In morphology, Latvian has carried through a number of innovations (the central dialect is the most innovative in this respect, cf. 3.1.2 below): 1.1.2.1 The neuter forms of adjectives and pronouns have been completely lost, whereas Lithuanian has largely retained them (nouns are either masculine or feminine in both languages). 1.1.2.2 Simplification of nominal declension: virtual loss of the instrumental case (which had certain consequences for the government of prepositions, cf. 1.1.3), partial (for adjectives, complete) loss of distinctions between the inherited declension classes.
5
6
Laimute Balode and Axel Holvoet
1.1.2.3 Simplification of verbal inflection: the distinction of three types of presents (-a-, -a¯- and -i-stems) is virtually abandoned; in most dialects the distinction of two types of preterite stems is abandoned as well. 1.1.2.4 A special inflectional form of the verb was created to express necessity, corresponding to the modal verbs ‘must’, ‘have to’ of other languages: the so-called debitive. The object of a debitive is usually in the nominative, the subject (agent) in the dative: Man ja¯pl¸auj siens. me:dat mow:deb.pres hay:nom ‘I have to mow the hay.’
The debitive is not conjugated for person, and the tenses are formed by addition of the corresponding tense forms of the auxiliary bu¯t ‘be’. The debitive was originally an agglutinative form arising from the coalescence of the relative pronoun with the verb in a non-finite relative clause added to the subject of a possessive construction (cf. 1.1.4). The construction from which the above example arose would have to be rendered literally as ‘I have some hay (which) to mow’. 1.1.3 In morphosyntax, a striking feature of Latvian is the fact that prepositions can govern different cases in the singular and the plural. This is probably due to the loss of the instrumental, which coalesced with the accusative in the singular and with the dative in the plural. As a result, some prepositions seemed to govern the dative in the plural, but the other case form in the singular. This pattern was generalized, and in most of the dialects all prepositions now govern the dative plural, regardless of the case they govern in the singular (though some dialects retain the genitive plural with at least some prepositions). 1.1.4 A peculiarity of Latvian syntax is its possessive construction, which is not based on a verb meaning ‘have’, but on the verb ‘be’, with the object of possession usually in the nominative (though in some dialects it is always in the accusative), and the possessor in the dative (cf. Latin mihi est ‘I have’). Man ir ma¯ja. me:dat be:pres.3 house:nom.sg ‘I have got a house.’
This construction is rare in Lithuanian, which has the verb ture˙ti ‘have’, but it is reminiscent of the possessive constructions used in the Finnic languages and in Russian. This possessive construction formed the basis for a peculiarly Latvian form expressing necessity, cf. 1.1.2.4.
The Latvian language and its dialects
1.2 The formation of the Latvian language At the end of the 12th century (the earliest period covered by the chronicles relating the German missionaries’ activities and the subsequent conquest of Livonia by the Swordbearers or Livonian Knights), the territory of present-day Latvia was inhabited by five tribes: –
–
the Finnic Livonians (Livones), who occupied a coastal area reaching much farther inland than nowadays; they were the first with whom the German merchants, missionaries and knights established contact, and so gave their name to the whole of present-day Latvia and Estonia (German Livland, whence the Latinized form Livonia and, in later times, Polish Inflanty); subsequently this denomination was restricted to the Duchy of Livonia, extending to the north-east of the Daugava; four Baltic (Indo-European) tribes: – the Curonians (Curones) in the western part of Latvia (Courland proper); – the Semigalians (Semigalli) east of the Curonians, to the south of the Bay of Riga; – the Selonians (Selones), east of the Semigalians, in Upper Courland; – the Letgalians (Lethgalli), in the central and eastern parts of Latvia, to the north-east of the Daugava.
The territories of all these Baltic tribes, with the exception of the Letgalians, extended into the territory of present-day Lithuania; they were thus partly assimilated by the Lithuanians. It is now generally recognized that Curonian was a separate dialect of Baltic, displaying a series of features which place it in an intermediate position between Lithuanian and Latvian (some would prefer to say: between Eastern and Western Baltic). Its phonetic features can partly be reconstructed on the basis of place names, which, in turn, make it possible to identify a number of Curonian substratum features in present-day Latvian and Lithuanian, occurring in, or spreading from, formerly Curonian areas. Much less is known about Semigalian and Selonian, though a few putative substratum features are ascribed to these dialects as well. The fourth of the Baltic tribes, the Letgalians, must have played a crucial part in the formation of the Latvian nation and the Latvian language. In the 13th century Chronicon Heinrici the terms Let(t)(h)i ‘Latvians’ and Letthigalli ‘Letgalians’ are used interchangeably (cf. Lethi, qui proprie dicuntur Lethigalli ‘the Latvians, properly called Letgalians’), i.e. the Letgalians were the Latvians par excellence. They did not resist the Livonian Order as fiercely as other tribes (such as the Semigalians) did, and even collaborated with it. In recognition of this they were given preferential treatment by the Order, probably also in the form of allotments of land in areas depopulated by war. It is thus conceivable that the Latvian language as we know it
7
8
Laimute Balode and Axel Holvoet
is the result of Letgalian expansion into areas originally inhabited by Curonians, Semigalians, Selonians and, of course, (non-Baltic) Livonians. This theory, advanced by historians and once commonly held by linguists as well, has been disputed in more recent times by linguists advocating the view that the present-day Letgalian dialect is a separate Baltic language, separate from Latvian. These linguists tend to project the distinctive features of Letgalian into a distant past, and to regard this dialect as the only direct continuation of the ancient language of the Letgalian tribe. This Letgalian separatist view raises serious doubts, however, as none of the distinctive features of Letgalian can be considered ancient. In the domain of phonetics, Letgalian stands clearly apart from the rest of Latvian dialects, but none of its phonetic features can be projected into a distant past. In morphology, Letgalian has a number of archaisms, but these are typical peripheral archaisms also found in parts of Courland. Otherwise Latvian is relatively monolithic: there are no isoglosses going back to a dialectal differentiation within Common Baltic (Proto-Latvian). If we do find forms displaying, in their phonetic shape, a treatment deviating from the common Latvian sound laws, these are substratum forms, which is not surprising if we take into account the theory of Letgalian expansion expounded above. In addition, some of the differentiating features used in classifying Latvian dialects are connected with external influence, i.e. with a non-Indo-European substratum. Apart from the theory of the Letgalian origin of Latvian, other expansion theories have been advanced as well. Some scholars are inclined to assign a certain role to a putative Semigalian expansion as a unifying factor contributing to the relative uniformity of the Latvian dialects. Latvian is attested in writing since the end of the 16th century. From the 16th to the early 19th century Latvian was written mainly by German Lutheran pastors, who used the central dialect as the vehicle of their religious and didactic literature. With regard to its dialectal basis, their language was remarkably uniform from the start. Under the influence of the Latvian national revival, the Latvian written language was reformed from the 1860s onwards: German influences in syntax and lexicon were banished (many German borrowings being replaced by neologisms). From the 18th century onwards, there is also a tradition of writing in the High Latvian (Letgalian) vernacular. Its rise was conditioned by the political and religious separation of (Roman Catholic) Polish Livonia from (Lutheran) Swedish Livonia and Courland. The functional scope of the High Latvian written standard is restricted: it is connected mainly with the pastoral activities of the Roman Catholic Church, though, to a limited extent, it has been used for literary purposes as well, and it enjoys some degree of official recognition, e.g., in education.
The Latvian language and its dialects
1.3 Substrata and adstrata It follows from the above survey of the original population structure of the territory of present-day Latvia that we must reckon with a Finnic substratum in Latvia. In fact, several layers of substratum features may be distinguished. Though archaeological and toponymical data are not unambiguous, it seems probable that the whole of present-day Latvia was once inhabited by Fenno-Ugric tribes, who entered the Baltic area in the 3rd millennium B.C. They were probably pushed back to the north about the 12th century B.C. by Baltic tribes coming from the south-east (the bearers of the so-called Hatched Pottery Culture). A few features distinguishing Latvian as a whole from Lithuanian and Old-Prussian can be ascribed to Finnic influence. The initial stress of Latvian, as opposed to the mobile stress of Lithuanian and Old Prussian, is most often cited as an instance of this (which does not mean that no other explanations for this feature are available). Of course, a stronger admixture of Finnic features may be expected in those areas where the presence of a Finnic population is attested in historic times, i.e. from the 12th century onwards. The so-called ‘Livonian’ (Tamian) dialects of Latvian are defined mainly on the basis of features which can be traced back to a Finnic substratum. On the other hand, there are substratum features in those Baltic dialects which were assimilated by Latvian as a result of the Letgalian (and perhaps also, to a certain extent, the Semigalian) expansion. These are easiest to identify for Curonian, which has had a few distinctive features opposing it to what is now known as Latvian, and partly also to Lithuanian from Common Baltic times onward. Like Lithuanian, Curonian preserves the nasal diphthongs /aN/, /eN/, /iN/, /uN/. These combinations are called diphthongs (or diphthongoid sequences) because, even though the sonorants can no longer be syllabic in Baltic, in combination with vowels, they function as long vocalic segments, i.e. they distinguish tones (see 1.5.5.3 below). In the prehistory of Latvian they changed into nasal vowels which were subsequently denasalized, cf. 1.1.1. Forms which retain the structure /V + N/ occur everywhere in Latvian, perhaps as a result of interdialectal borrowing, but they are particularly frequent in Courlandish dialects, where they are considered Curonianisms. A few other features, recognized as characteristic of Curonian, have left traces only in toponymy. Considerably less is known about Semigalian and Selonian, but the view has been advanced that they had retained the nasal diphthongs in the same way as Curonian had. Some Slavonic languages (Polish, Russian, Belarusian) have evidently played the part of adstratum languages with regard to the High Latvian dialects of Letgalia (Polish Livonia), which show a number of phonetic features strongly reminiscent of Slavonic (apart from strong lexical influences).
9
10
Laimute Balode and Axel Holvoet
1.4 External influences In prehistoric times (i.e. before the 13th century), at least some of the Baltic tribes inhabiting Latvia, mainly the Letgalians, were tributary to the North-Russian principalities of Polock and Pskov, and must have undergone a strong cultural influence from them. The oldest layer of religious terminology connected with Christianity (prior to the activities of German missionaries) is of Slavonic (Old Russian) origin, e.g. baznı¯ca ‘church’ (OR božı˘nica), gre¯ks ‘sin’ (OR greˇchu˘), etc. Slavonic influence was discontinued as a result of the conquest of Livonia by the Swordbearers (the Livonian Order). It was renewed after the incorporation of part of the Livonian Order State (the Duchy of Livonia) into the Polish-Lithuanian Commonwealth, though it had lasting effects only in that part of it where PolishLithuanian ascendancy was maintained after 1660 (so-called Polish Livonia). In this area, Latvian (Letgalian) was exposed to the influence of Polish and East Slavonic (Russian and Belarusian). In Courland and Swedish Livonia, German influence was strong; this is reflected in a large number of loans, the oldest layer of which (up to the 16th century) is Low German. German-Latvian bilingualism was, however, very restricted, as the German gentry tended to isolate itself from the native population rather than to assimilate to it. Structural (e.g., syntactic) influence of German on Latvian should therefore not be overestimated as far as the popular dialects are concerned. In the written language, German syntactic influence was, of course, pervasive, as Latvian was written mainly by German pastors up to the 19th century. After the incorporation of the whole of Latvia into the Russian Empire (Swedish Livonia — 1721, Polish Livonia — 1772, Courland — 1795) Russian influence was initially slight (with the exception of Polish Livonia, incorporated into the Vitebsk Province) owing to the autonomy which the Baltic provinces enjoyed within the Russian Empire. A policy of Russification was pursued more forcibly from the 1880s onward, and it was renewed after the incorporation of Latvia into the Soviet Union in 1940. 1.5 Phonemic inventory A survey of the phonemic inventory is here given for the standard language; divergences from this system will be noted in the survey of the individual dialects. 1.5.1Vowel system Vowels i(˜)
u(˜) e(˜)
o(˜) æ(˜)
a(˜)
The Latvian language and its dialects
In the standard orthography, length is marked by a macron (·a¯Ò, ·e¯Ò etc.). The phonemes /e/ and /æ/, /e˜/ and /æ˜/ are not distinguished in orthography, although the opposition is distinctive, as shown by the minimal pairs cep / tësep/ ‘you bake’ vs. cep /tësæp/ ‘s/he bakes’, /sle˜dëz/ ‘you shut’ vs. /slæ˜dëz/ ‘s/he shuts, they shut’ (both written sle¯dz). In textbooks and dictionaries, the phonemes /æ/ and /æ˜/ are sometimes marked ·eÒ í and ·e¯í Ò. The vowels /o/ and /o˜/ occur in recent loans only, and quantity is not marked here (i.e. the macron is not put over ·oÒ). Rising diphthongs (i.e. with rising degree of sonority) ie
uo
The second component is actually a centralized vowel, which is more fronted in the case of /ie/ than of /uo/. The transcriptions /i6/, /u%/ would describe them best for practical purposes. In standard orthography these diphthongs are written ·ieÒ and ·oÒ respectively; the letter ·oÒ can thus denote either a diphthong (in native words and older borrowings) or the vowel /o/ (in recent borrowings). In historical and comparative literature ·uoÒ is written instead of ·oÒ when the diphthong is meant. Falling diphthongs (i.e. with diminishing degree of sonority) ui
iu ei
(oi) ai
au
/oi/ is restricted to recent borrowings (boikots ‘boycot’). The inclusion of /ui/ and /iu/ is a matter of convention, as both components have a similar degree of sonority here. To the diphthongs occurring in this table we can add the positional diphthongs /eu/ and /ou/, which are variants of the sequences /ev/ and /ov/ before consonants and word-finally, e.g. tev /teu/ ‘you’ (dat) as opposed to tevi /tevi/ ‘you’ (acc). The diphthongs /au/ and /iu/ can also be positional diphthongs, e.g. tavs /taus/ ‘your’ (nom.sg.masc) compared to tava /tava/ ‘your’ (nom.sg.fem) and zivs /zius/ ‘fish’ (nom/gen.sg) as compared to zivis /zivis/ ‘fish’ (nom/acc.pl). 1.5.2Origins of the vowel system Most Latvian vowels come from similar IE vowels, but IE *o has merged with *a into /a/, and its long counterpart *o¯ is represented by the diphthong /uo/. In the 19th century, /o/ and /o˜/ were reintroduced through borrowings. The vowels /e/ and /æ/, and their long counterparts /e˜/ and /æ˜/, are from IE *e and *e¯ respectively; the split was conditioned by the phonetic environment, the closed /e/ occurring when the following syllable contained a front vowel or /j/. Hence metu /mæt˜u/ ‘throw’ (pres.1sg) and met /mæt/ ‘id.’ (pres.3) from an original form *meta, in contrast to met /met/ ‘id.’ (pres.2sg) from an older form meti. The falling diphthongs basically
11
12
Laimute Balode and Axel Holvoet
also continue similar IE diphthongs, with the exception of /ui/ (occurring mainly in borrowings) and /iu/ (in onomatopoeia only). IE *oi and *ai have, of course, coalesced into /ai/; this diphthong, as well as *ei, is sometimes retained as such, but under certain conditions (the precise nature of which is still unclear) both are changed into the rising diphthong /ie/. 1.5.3Consonant system
The characters used in standard orthography are added in angle brackets. The phonemes /f/ and /x/ were originally absent from the Latvian phonemic system; they occur only in recent loans. There is a distinct series of palatal consonants, but palatalized consonants (i.e. with palatalization as a secondary articulation) are now lacking. The only exception used to be palatalized /rj/ (written ·r¸Ò), but this sound is becoming extinct in both dialects and standard language, and in the standard orthography it was abolished in 1946. 1.5.4Origins of the consonant system The voiceless plosives come from similar IE sounds; the voiced plosives may represent IE unaspirated or aspirated voiced plosives (i.e. /d/ may represent IE *d or *dh). The IE palato-velars *k’, *g’ are represented by /s/, /z/ (as opposed to /w/ ·šÒ, /Š/ ·žÒ in Lithuanian), so that Latvian /s/ can be of twofold origin (IE *s or *k’). The Latvian palato-alveolar fricatives /w/ ·šÒ, /Š/ ·žÒ arise from combinations of /s/ and /z/ with /j/ (e.g. šu¯t ‘sew’ as opposed to Lith. siu¯ti). The dental affricates have arisen from original /k/ and /g/ before front vowels, or from the combinations of /k/ and /g/ with /j/ (cf. celt ‘lift, build’ as against Lith. kelti, dzert ‘drink’ as against Lith. gerti). The palato-alveolar affricates /tëw/ ·cˇÒ and /dëŠ/ ·džÒ appear (apart from in loans) as a result of analogical development, having been substituted for original /tës/ and /dëz/ by analogy with the alternation /s/ : /w/ and /z/ : /Š/ appearing in inflectional paradigms, e.g. la¯cis ‘bear’, gen.sg la¯ˇca (instead of *la¯ca) by analogy with the type lu¯sis ‘lynx’, gen.sg lu¯ša. The palatal sonorants /ˆ/ ·l¸Ò and /\/ ·n¸Ò arose from
The Latvian language and its dialects
combinations of /l/ and /n/ with /j/. In native Latvian words, the palatal plosives /c/ ·k¸Ò and /K/ ·g’Ò occur only in the clusters /wc/ ·šk¸Ò and /ŠK/ ·žgˇÒ, arising from /sk/ and /zg/ before front vowels (and from the combinations of these clusters with /j/), e.g. šk¸irt ‘separate, distinguish’ as opposed to Lith. skirti ‘id.’. In isolation, they appear mainly in borrowings, where they represent German /k/ and /g/ before front vowels (e.g. g’ek¸is ‘fool’ from Germ. Geck) and Russian palatalized /tj/, /dj/ (e.g. sog’is ‘judge’ from Old Russ. sudija). 1.5.5Prosodic features 1.5.5.1 Stress. Modern Latvian has a basically non-distinctive, immobile stress on the initial syllable. Exceptions are few; they comprise negative pronouns (ne’kas ‘nothing’), and compounds developed from various types of phrases. The latter group comprises mainly adverbs, such as uzÁreiz ‘at once’ (from the prepositional phrase uz reizi ‘at one time’), and superlatives such as visÁlaba¯kais ‘the best’, originally visu laba¯kais ‘the best of all’. Minimal pairs are extremely rare, e.g. Áa¯re ‘arable land’ vs. a¯Áre ‘look here’. 1.5.5.2 Quantity. Vowels can be short or long. Quantity is independent of stress, but is ultimately connected with tone (see below 1.5.5.3). Quantity is distinctive in any position: saka [Ásak˜a] ‘say’ (pres.3) : sa¯ka [^sa˜ka] ‘begin’ (pret.3) laba [Álaba] ‘good’ (nom.sg.fem.indef) : laba¯ [Álaba˜] ‘id.’ (nom.sg.fem.def)
1.5.5.3 Syllabic tones (syllabic accents). The syllabic tones or accents of Latvian are different modes of realization of word stress. Just as word stress commonly involves (with varying degrees of prominences) three features: expiratory strength, tone height, and quantity (length), so does syllabic accent. The difference consists in that, in the case of long vocalic segments, these features are unevenly spread, and may either be concentrated on its initial part (in the first mora), or be spread over the initial and final parts. An additional feature occurring in combination with tonal features proper is glottalization. As tonal differences (pitch) are not the only feature involved here, the term ‘tones’ is less appropriate than ‘syllabic accent’; still less felicitous is the term ‘intonation’ (still used in Latvian terminology: intona¯cija), which could be confused with sentence intonation. As mentioned, tone is exclusively a feature of long vocalic segments. As such are regarded: 1. long vowels, e.g. ma¯cı¯t [ma˜tësi˜t] ‘teach’ (inf) 2. falling and rising diphthongs with /i/ and /u/ as their second resp. first component, e.g.
13
14
Laimute Balode and Axel Holvoet
teikt ‘say’ (inf) augt ‘grow’ (inf) liepa ‘lime, linden’ (nom.sg) roka ‘hand’ (nom.sg) 3. diphthongoid sequences consisting of a vowel and a sonorant. Though, in principle, sonorants cannot be syllabic in Latvian (they could in Indo-European), their combinations with short vowels yield long (i.e. able to carry tone), vocalic segments, e.g. manta ‘thing, possession’ (nom.sg) varde ‘frog’ (nom.sg) Some of the central dialects (and, theoretically, the standard language as well, see below), have a system of three tones. a. The falling tone (krı¯toša¯ intona¯cija, German fallender Ton). The initial part of the vocalic segment is pronounced with greater intensity and a higher pitch; both intensity and pitch decrease toward the end of the vocalic segment. b. The drawn tone (stiepta¯ intona¯cija, German Dehnton) shows no decrease in either intensity (expiratory strength) or pitch toward the end of the articulation of the vocalic segment: both are maintained at the same level. The total average length of a vocalic segment with drawn tone exceeds that of a segment with falling tone by about 1/3. In the case of diphthongs and diphthongoid sequences, the length of the second component (the half-vowel or sonorant) distinctly exceeds that observed in falling diphthongs. c. The broken tone (lauzta¯ intona¯cija, German Stoßton) is similar to the falling tone in that the initial part is more prominent in both intensity and pitch, but after this initial segment there is a glottal stop, and the remaining part is pronounced voicelessly (whispered). The distinctive nature of the tones is shown by the following near-homonyms, distinguished only by tone (for details on the use of the tone marks cf. 1.5.5.4): lùoks luõks luôgs
‘bow, arch’ ‘onion leaf ’ ‘window’
(falling tone) (drawn tone) (broken tone)
In unstressed position (if the syllable bears neither main nor secondary stress), the opposition between drawn and falling tone is neutralized; only the glottal stop remains: labãs laba¯s
The non-glottalized realization of unstressed long vowels and diphthongs is usually marked with the sign of the drawn tone, though it is more similar to the falling one.
The Latvian language and its dialects
Of course, syllables bearing secondary stress may also display tone oppositions, cf. ap-sèja ‘sow’ (pret.3) ap-s«e ja ‘bind up, bandage’ (pret.3)
However, many speakers of Latvian neutralize the opposition here as well. This means that in unstressed syllables (with the possible exception of syllables bearing secondary stress) tonal oppositions proper are neutralized. There may be an opposition between ‘glottalized’ and ‘non-glottalized’, but this is not, strictly speaking, a tonal feature, though under stress it occurs as a concomitant of tone. In Latvian linguistics, glottalization is considered to be a tonal feature, which means that tone is regarded as independent of word stress. This is reflected in the fact that tone marks are put over all syllables containing long vocalic segments. The system with three tones is variously reduced in most Latvian dialects. In most of the Central dialect, the falling and broken tones have merged into a single, broken tone. In principle, the three-tone system is recommended for the standard language, but the distinction of three tones is hard to elicit from speakers distinguishing only two in their native dialect. For most speakers of the standard language, the opposition should be thus formulated as ‘drawn vs. non-drawn’, further differentiation (with an opposition between ‘falling’ and ‘broken’ instead of one single ‘non-drawn’ tone) being observed only in the pronunciation of those speakers who distinguish three tones in their native dialect. In High Latvian (Letgalian), the falling and drawn tones have merged, which results in an opposition of ‘falling vs. broken’. As the difference is one of glottalization, tone has actually become neutralized in this dialect. There is only an opposition between presence or absence of stød (a situation comparable to that of Danish but not the remaining Scandinavian languages). 1.5.5.4 The use of tone marks. The tones are not marked in the standard orthography, only in linguistic publications. The use of tone marks is as follows: –
–
–
The falling tone is marked with ·`Ò; in the case of diphthongs and diphthongoid sequences this mark is put over the letter denoting the first component: kàsis ‘hook’, màlks ‘draught’, bràukt ‘drive’ The drawn tone is marked with ·«Ò; in the case of diphthongs and diphthongoid sequences this mark is put over the letter denoting the second component: mãte ‘mother’, mañta ‘thing, possession’, laı˜ks ‘time’. The broken tone is marked ·ˆÒ; this mark is put over the second component of a diphthong: vâks ‘lid’, da¥rbs ‘work’, laûzt ‘break’.
Under certain conditions, which vary from one dialect group to another, in original diphthongoid sequences of the type /V + r/ (/ar/, /er/, /ir/, /ur/), the first component, which was originally always short, may be lengthened or diphthongized.
15
16
Laimute Balode and Axel Holvoet
Tonal differences are then realized within the vowel (the sonorant ceasing to belong to the vocalic segment which can carry tone), and all three tone marks are put over the first letter, e.g. kãrpa ‘wart’, dârbs2 ‘work’ (in the Curonian dialects, compared to da¥rbs in the standard language. The significance of the index ·2Ò is explained in the next paragraph). If, in such a case, the tone is falling, a macron is added to the tone mark: v×ards ‘word’ (unlike kàrte ‘map’, a recent loan where the vowel did not undergo lengthening). The three-tone system is the original one. The systems with two tones result from the reduction of this system, and this reduction has completely obscured the original distribution of tones. The historical equivalence between the Latvian and Lithuanian tones can therefore be established only on the basis of the three-tone system. In historical-comparative works on Baltic languages, Latvian words are, in principle, cited with the tone attested in the dialects distinguishing three tones. In citing words and forms from dialects with reduced tone systems, an index ·2Ò is added to indicate that the tone marked on the word in question is the regular outcome of neutralization, and therefore provides no clue as to the original tone of the word, e.g. daîkts2 ‘thing’. 1.5.6Note on Latvian orthography Though Latvian has only one official orthography, a special variety of this is used in linguistic (especially comparative and historical) literature on Latvian and the Baltic languages in general, in etymological dictionaries etc. It differs from the standard orthography on three points: –
– –
2.
instead of the macron (length mark), three different tone marks are used, according to the tone of the syllable (see 1.5.5.4 above), and tone marks are added on diphthongs where these are absent from the standard orthography, e.g. rãdı¯t vs. ra¯dı¯t, naûda vs. nauda; where the letter ·oÒ of the standard orthography denotes the diphthong /uo/, ·uoÒ is written, e.g. rùoka vs. roka; where the letters ·eÒ, ·e¯Ò of the standard orthography denote [æ], [æ˜], the signs ·eÒ, í ·e¯í Ò are used, e.g. víesíels vs. vesels.
The Latvian dialects
2.1 Classification Since August Bielenstein (Bielenstein 1863/64), the Latvian dialects have been classified into three groups: –
the central dialects (Latv. vidus dialekts, Germ. der mittlere Dialekt) are spoken in most of Vidzeme (except for a transitional zone bordering on Letgalia in the east)
The Latvian language and its dialects
–
–
as well as in Semigalia and Courland proper (with the exception of its northern part); they have formed the basis of the standard language since the earliest writings (in the 16th century). the Tamian dialects (Latv. ta¯mnieku or lı¯biskais dialekts, Germ. tahmisch or livisch, also nordwestkurisch) are spoken on both sides of the Bay of Riga, in the north of Courland as well as in the north-west of Vidzeme. The term Tamian is based on the nickname ta¯mnieki, by which the speakers of this dialect were referred to, and which seems to be derived from the adverb ta¯m ‘until now’ (standard Latvian lı¯dz šim). This adverb, now fallen into disuse, seems to have been characteristic of the dialects of Courland as a whole. The term lı¯biskais dialekts (‘Livonian dialects’), used in Latvian terminology, refers to the Livonians, the Finnic tribe originally inhabiting the Tamian area. Strictly, the term Tamian refers only to the dialects spoken in the north of Courland, but it is also used to denote the whole dialect group, and this usage will be followed here. the High Latvian Dialect (Latv. augšzemnieku dialekts, Germ. hochlettisch) is spoken in Letgalia, the easternmost part of Vidzeme bordering on Letgalia, and in Selonia, to the south of the Daugava. This dialect is the basis of a regional written standard, sometimes referred to as the ‘Letgalian language’.
This classification is still used without any modifications. However, the Central Dialect and Tamian are sometimes grouped together and referred to by the term Low Latvian, as opposed to High Latvian. Some linguists of Letgalian extraction regard this as the only major line of division among Latvian dialects, arguing that the features commonly accepted as defining characteristics of Tamian are, in fact, but a more consistent realization of features shared with the central dialect. Though this may be conceded, it must also be added that the defining features of High Latvian, though clearly opposing it to Low Latvian, are probably of relatively recent date (caused presumably by a Slavonic adstratum), so that no strong case can be made for the dichotomous classification. 2.2 The main differentiating features of the Latvian dialects The main criteria used in classifying the Latvian dialects are: – – –
differences with regard to prosodic features (tones); qualitative and quantitative changes in the vowel system; morphological innovations.
The distinguishing features of the three main dialect groups can best be shown by contrasting the Tamian and High Latvian dialects with the central dialects, on which the standard language is based. We should mention that the central dialect group is generally phonetically more conservative than the other dialects (a fact which may be connected with the
17
18
Laimute Balode and Axel Holvoet
less pervasive influence of substratum or adstratum languages). In morphology, however, at least part of this dialect has carried through a number of innovations (see 3.1.2.1–3, below, on the formation of the preterite, the future and reflexive verbs). In the domain of prosody, the conservative character of the central dialect manifests itself in the retention, in part of its area, of the original system of three tones. This can, admittedly, only be observed in a few places (in a compact area around Valmiera and Ce¯sis in Vidzeme, and in the civil parishes of Blı¯diene and Jaunpils in Semigalia), but Tamian and High Latvian have only reduced tone systems. The vowel system of the central dialect is both qualitatively and quantitatively much closer to that of Common Baltic than those of the other dialects. The vowels have undergone relatively few qualitative changes, which distinguishes the central dialect from High Latvian (with its ‘velar vowel shift’); and, compared with the Proto-Latvian state of affairs, there has been no further loss of final short vowels, and shortening (loss) of final long vowels, as observed in Tamian. The following table gives typical representations of five words in the three dialects as well as in the literary language: SL
As can be seen, High Latvian displays mainly qualitative changes: diphthongization of /i˜/ and /u˜/, vowel shift /a/ > /o/. The Tamian dialect shows mainly quantitative changes: reduction of long unstressed vowels to short vowels, and of unstressed short vowels to zero. The peculiarities of the three main dialects are once more illustrated in the following sentence (from a folk song) in its three varieties: Central dialect: ganuôs gãju, kre˛klu šuvu, pìe uôzuola m˜eí ru n¸êmu [Áganu%s «ga˜ju Ð Ákræklu Áwuvu Ð Ápi6 ^u%zu%la «mæ˜ru ^\e˜mu]
The Latvian language and its dialects
Tamian dialect: gane˛s gãj, kre˛kal šûj, pe uôzal2 m˜eí r j˜eí m [Áganæs «ga˜j Ð ‘krækal wu˜j Ð pë ^u%zal «mæ˜r «je˜m] High Latvian dialect: gonûs gùoju2, kraklu šyvu, p’i ûzula màru2 jiêmu [Ágfnu˜s Águ%ju Ð Ákraklu ÁwGvu Ð pji ^ulzula Áma˜ru ^ji7mu] In the standard orthography: ganos ga¯ju, kreklu šuvu, pie ozola herdsman:loc.pl go:pret.1 shirt:acc.sg sew:pret.1 by oak:gen.sg me¯ru n¸¯emu measure:acc.sg take:pret.1 ‘I went a-pasturing, I sewed a shirt, I took the measure by an oak’
Commentaries: The Central dialect is here represented by a variety distinguishing three tones. The phonetic shape of this fragment corresponds exactly to the standard language. The prothetic /T/ in n¸emt ‘take’ is idiosyncratic. The Tamian equivalent shows shortening and concomitant monophthongization of unstressed diphthongs (/uo/ > /a/ or /æ/, /ie/ > /e/ etc.) as well as the consistent loss of final short vowels, with epenthesis of /a/ where this would lead to a word-final sequence ‘obstruent + sonorant’ (/krækal/ < /krækÏ/ < /kræklu/), and with compensatory lengthening (šu¯j). High Latvian shows the High Latvian vowel shift (/a/ > /o/, /a˜/ > /uo/, /æ/ > /a/ and /æ˜/ > /a˜/) and monophthongization of diphthongs (/uo/ > /u˜/). The High Latvian palatalization is marked on p’i. In the Tamian version, the index ·2Ò warns that the broken tone may represent an original broken or falling tone: in the High Latvian version it warns that the falling tone may represent an original falling or drawn tone. Until now only vocalic features have been mentioned. If we take into account the consonant system as well, we will see that High Latvian is more distinctive here. 1. High Latvian has a full-fledged correlation of palatalization, as a result of assimilatory palatalization before front vowels (which subsequently became distinctive as a result of vowel loss, e.g., word-finally). This feature is shared by Lithuanian, but distinguishes High Latvian from the central dialects, which have no palatalization as a secondary articulation. Tamian also has a certain degree of assimilatory palatalization, but only stem-finally, probably a result of Finnic influence, as will be shown below. The central dialects have no palatalization as a secondary articulation, but they have a separate palatal series comprising the nasal /\/, the lateral /ˆ/ and the stops /c/ and /K/, as seen in the table of consonants. This is in accordance with Trubetzkoy’s observation: if a language has a correlation of palatalization, it does not have a palatal series, because palatals are identified as palatalized dentals or alveolars. Where Low Latvian has /c/, /K/, High Latvian
19
20
Laimute Balode and Axel Holvoet
substitutes /tëw/, /dëŠ/ (see 5.1.8 below) and Letgalians speaking standard Latvian substitute palatalized velars /kj /, /gj/. Widespread assimilatory palatalization seems to be associated with a Slavonic adstratum. Although most Balticists ascribe this feature to Common Baltic (or Balto-Slavonic) because it occurs in Lithuanian, High Latvian palatalization cannot have been inherited from Common Baltic, because: a.
if Latvian had once been characterized by generalized assimilatory softening and subsequently lost it, it would have lost its palatal series as well, according to Trubetzkoy’s observation, and b. wherever Letgalian has the vowel shift /æ/ > /a/ (cf. HL vacs ‘old’, standard Latvian vecs [vætës.], Lithuanian (obs., reg.) v˜etušas [«vjæ˜tuwas]), the preceding consonant is not palatalized, which means that palatalization in High Latvian cannot be older than the High Latvian vowel shift. 2. High Latvian devoices obstruents word-finally, like Lithuanian and most of Slavonic, but unlike Low Latvian, e.g. vajag ‘it is necessary’ is pronounced /vajak/ in High Latvian, while final voicing is retained in Low Latvian and the standard language. Slavonic influence seems to be involved here as well. 3. Low Latvian, unlike High Latvian, is characterized by non-distinctive lengthening of unvoiced obstruents. This occurs between short vowels, especially in disyllables, where the final vowel can be unvoiced (whispered): lapa visi
[lap˜a] or [lap˜a] [vis˜i] or [vis˜i]
It does not occur when either of the vowels is long: saka sa¯ka saka¯m sa¯ka¯m
[sak˜a] [sa˜ka] [saka˜m] [sa˜ka˜m]
‘say (pres.3)’, also [sak˜a] ‘begin (pret.3)’, also [sa˜ka] ‘say (pres.1pl)’ ‘begin (pret.1pl)’
There is a potential opposition between geminated and ungeminated obstruents on the morpheme border, as in ata¯ls ‘aftermath’ and atta¯ls ‘distant’, but usually a short vowel is heard in both cases unless a pause is inserted between the final [t] of at- and the initial [t] of -ta¯ls. Though non-distinctive, this lengthening is part of the standard pronunciation and causes notorious difficulties to speakers of High Latvian, who have only geminates in their dialects. In this respect, High Latvian resembles Lithuanian (which, however, goes further and usually even shortens geminates) and such Slavonic languages as Polish and Russian, which have only geminates, but no phonetic lengthening of obstruents. The difference between Low and High Latvian may be connected with differences in syllable structure, but this still has to be investigated in detail. Whereas the state of affairs observed in High Latvian is
The Latvian language and its dialects
reminiscent of Slavonic, the Low Latvian situation reminds us of Finnic, where a correlation of quantity in obstruents is a common phenomenon. More specifically, we are reminded of the lengthening of voiceless obstruents after short vowels in Livonian, as described by Kettunen (1938). In Livonian, this lengthening is not dependent on grade alternation conditioned by the opposition of original open and closed syllables, but by the reduction (shortening and contraction) of final syllables.
3.
The central dialects (vidus dialekts)
3.1 Common features 3.1.1 As mentioned above, the central dialects are the phonetically most conservative, so that their distinguishing features, as compared to Tamian and High Latvian, can be formulated negatively (as the lack of some phonetic change or other). The basic features of the phonemic and prosodic system coincide with those described above for the standard language. More specifically, we may note the retention of short vowels in final syllables. True, these are often unvoiced (whispered) or reduced to zero, but they may always reappear in more careful speech. In some types of polysyllabic words, however, complete vowel loss is observed in the central dialects as well, e.g. in nouns ending in -ı¯tis, -ietis: SL
CD
meaning
bra¯lı¯tis latvietis
bra¯lı¯ts latviets
‘brother’ (dim.nom.sg) ‘Latvian’ (nom.sg)
3.1.2 In morphology, on the other hand, at least some of the central dialects have carried through a number of innovations, which are also characteristic of the standard language. They distinguish most of the middle dialect, and the standard language, from High Latvian on the one hand, but also from the Curonian dialects of the central group. Of course, this does not point to a closer connection between Curonian and High Latvian; it is simply an instance of a peripheral archaism. The innovations are the following. 3.1.2.1 Common Baltic had two kinds of past tense stems, in -a¯- and in -e¯-. Lithuanian has retained both, but the central dialects of Latvian, and the standard language, have generalized the preterite in -a¯-. The changes undergone by stemfinal velar consonants in the preterite show that preterites in -e¯- must once have existed everywhere in Latvian. If standard Latvian has a preterite (3rd person) na¯ca
21
22
Laimute Balode and Axel Holvoet
from the verb na¯kt ‘come’, then this points to an original preterite na¯ce, because it is only before a front vowel that /k/ is changed into /tës/. High Latvian has retained the -e¯-preterites, and so have some of the Curonian dialects; e.g. CD (Livonian) and SL
3.1.2.2 The Baltic future is formed with the suffix -si-, added to the infinitive stem. If the infinitive stem ends in a dental or alveolar consonant, consonant clusters arise which are normally simplified, rendering the forms morphologically opaque. In some of the Livonian dialects this is avoided by inserting /i˜/ (borrowed from a verb class with an infinitive stem in /i˜/). Alongside daru ‘I do’ there is a future darı¯šu ‘I will do’, based on the infinitive darı¯t, and this gave rise to an analogical relationship metu : metı¯šu. High Latvian and some of the Curonian dialects have retained the original forms without insertion: CD (Livonian) and SL
CD (Curonian)
LITH
meaning
lauzı¯šu vedı¯šu
lauššu veššu
laušiu vesiu
‘break’ (fut.1sg) ‘lead’ (fut.1sg)
3.1.2.3 In the Baltic languages the reflexive pronoun has come to be integrated into the verb form. Basically, it is now affixed to it (as is the case in east Slavonic). If the verb contains a prefix, however, the reflexive marker is inserted between this prefix and the lexical verb stem, which is a relic of the original mobility of the reflexive pronoun. Lithuanian still shows this state of affairs: baigend baigend papref papref
-e˙pret -e˙pret baigend sirefl
-Ø 3 -Ø 3 -e˙pret baigend
-si refl -Ø 3 -e˙pret
-Ø 3
The majority of the central dialects of Latvian (and the standard language) have generalized a single pattern, with the reflexive marker affixed at the end, after the personal endings; but the pattern with infixed reflexive marker can be shown to
The Latvian language and its dialects
have been general in Latvian as well. The Curonian dialects have, to a certain extent, retained the original state of affairs (as have the High Latvian dialects). Often, however, the resulting verb forms have double reflexive markers, one inserted, the other affixed word-finally. The verb form in the example means ‘ends’ (intransitive, 3rd person present): CD (Liv.)/SL nopref
beidzend
-aTHEM
-s refl
CD (Cur)
-sarefl
beidzend
-aTHEM
nopref
-s refl
3.2 Subdivision of the central dialect group The central dialect group is subdivided into three subgroups: – – –
Livonian dialects (Latv. Vidzemes vidus izloksnes) Semigalian dialects (Latv. Zemgales izloksnes) Curonian dialects (Latv. kursiska¯s izloksnes)
Among these, the Curonian dialects are most specific, whereas the differences between the Livonian and Semigalian dialects are relatively slight. 3.2.1The Curonian dialects (kursiska¯s izloksnes) 3.2.1.1 The most characteristic feature of the Curonian dialects is the lengthening and/or diphthongization of the vowel in original diphthongoid sequences of the type /V + r/. Under falling and drawn tone all Latvian dialects lengthened the vowel if it was a non-high vowel (a process mentioned above, 1.5.5.4). In the Curonian dialects, this process extends to the high vowels /i/ and /u/, which, according to the dialect, are either lengthened into /i˜/ and /u˜/ or diphthongized into /ie/ and /uo/. Moreover, this lengthening occurs under all three (original) tones: SL
When the first component of the original diphthongoid sequence was a non-high vowel, lengthening is general in the Curonian dialects, whereas elsewhere it did not operate under broken tone:
23
24
Laimute Balode and Axel Holvoet
SL
CD (Livonian)
CD (Curonian)
meaning
darbs dzert
darps dzert
dârps2 dzêí rt2
‘work’ (nom.sg) ‘drink’ (inf)
3.2.1.2 An archaism of the Curonian dialects is the retention of /u/ before /v/ and /b/; in the other dialects this vowel has undergone loss of lip rounding as a result of dissimilation, yielding /i/: SL
CD (Livonian)
CD (Curonian)
meaning
zivs dibens
zivs /zius/ dibíens
zuve dubíenc
‘fish’ (nom.sg) ‘bottom’ (nom.sg)
The archaic character of the forms with /u/ is shown by Lith. žuvìs, dùgnas (< *dubnas). 3.2.1.3 Another archaism, not, in this case, a regular phonetic peculiarity, but a substratum feature, is the relatively high incidence of forms with retention of the Common Baltic diphthongoid sequences of the type /V + N/. In Latvian, these were, in principle, originally changed into nasal vowels which subsequently lost their nasal resonance to yield long oral vowels or diphthongs. The now extinct Curonian language retained these sequences in their original shape. SL
CD (Livonian)
CD (Curonian)
meaning
bezdelı¯ga
bezdelı¯ga
bezdelinga
‘swallow’ (nom.sg)
A number of forms with retention of the sequence /V + N/, where loss of the nasal would be expected, are found everywhere in Latvian; these probably result from dialect borrowing (expansion of Curonian forms), e.g. menca, mence ‘cod’, corresponding to Lith. ménke˙; the regular treatment in Latvian would have been *miece. 3.2.1.4 In the Curonian dialects, /v/ is lost after /l/, with possible compensatory lengthening of the preceding vowel:
3.2.1.5 In morphology, the Curonian dialects are characterized by a number of archaisms as compared to the innovating Livonian dialects. They have already been mentioned above, 3.1.2.1–3. 3.2.2The Livonian dialects (Vidzemes vidus izloksnes) The most characteristic feature of these dialects, occurring nowhere else, is the extension of /i˜/-insertion in the future tense (cf. 3.1.2.2) beyond the group of verbs with infinitive stems in dental consonants. Here we find /i˜/ generalized for all 1st conjugation verbs, i.e. verbs with monosyllabic stems in infinitive, present and preterite. Hence we have: SL
3.2.3The Semigalian dialects (Zemgales izloksnes) Semigalian is characterized by a tendency toward vowel insertion between /r/ (and, less frequently, /l/) and a following consonant, a feature found nowhere in the other dialect groups. The vowel preceding /r/ is short, whereas in other dialects it is lengthened at least with some tones. It is evidently the absence of lengthening that has conditioned the vowel insertion lacking in other dialects. This feature, which is not found outside Semigalian, is rapidly receding and is now observed in aged dialect speakers only. SL
4. The Tamian dialect group (lı¯biskais/ta¯mnieku dialekts) 4.1 Common features Tamian is divided into two groups: the dialects of northern Courland (Kurzemes lı¯biskais dialekts), which, as mentioned above, are the Tamian dialects proper, and those of north-western Livonia (Vidzemes lı¯biskais dialekts). Within the western
25
26
Laimute Balode and Axel Holvoet
(Courlandish) group, those spoken farthest to the north display the Tamian features in the most pronounced way; they are sometimes referred to in German terminology as strengtahmisch. 4.1.1 Tamian as a whole agrees with the majority of central Latvian dialects in having merged the falling and broken tones; both are represented by the broken tone, so that (2) and (3) are homonymous: (1) luõks
‘onion leaf ’ (nom.sg) 2
‘bow, arch’ (nom.sg)
2
‘window’ (nom.sg)
(2) luôks (3) luôgs
4.1.2 The most characteristic feature of Tamian is the reduction (loss or shortening) of unstressed vowels. Short vowels in word-final position are consistently lost. If the preceding syllable originally contained a short vowel followed by a voiced consonant, this vowel undergoes compensatory lengthening: SL
Tamian
meaning
labi kaza
la¯b ka¯z
‘well’ (adv) ‘goat’ (nom.sg)
If the consonant preceding the apocopated final vowel is a voiceless obstruent, a short vowel in the preceding syllable is not lengthened. In such cases, the obstruent is lengthened, as is generally the case between short vowels. The retention of long voiceless obstruents in word-final position is a peculiarity of Tamian and obviously accounts for the lack of compensatory lengthening of the vowel. SL saka lasa upe
A tendency towards loss of final short vowels can be observed in the central dialect as well, and the extent of this phenomenon is gradually increasing. In Tamian this process has been carried through consistently. Instead of occasional reduction of final vowels (which may either be dropped or be pronounced voicelessly), we observe their complete loss in Tamian, and this dialect has also developed compensatory mechanisms restricting the impact of vowel loss on the prosodic structure of the word.
The Latvian language and its dialects
The widespread loss of unstressed vowels seems to be connected with a similar development in Livonian, so that it could be considered a substratum feature. Livonian shows widespread loss of internal and final short vowels when compared to Estonian and other Baltic Finnic dialects, cf. Livonian
Estonian
Proto-Finnic
meaning
lapst tu’lb6D
lapsid tulevad
*lapset *tulepat
‘children’ ‘they come’
4.1.3 Unstressed long vowels and diphthongs, i.e. those occurring in non-initial syllables, are consistently shortened: SL
Further reduction may lead to complete loss of unstressed long vowels and diphthongs: runa¯t uozuolin¸š
ru¯nt uozliš
‘speak’ (inf) ‘oak’ (dim.nom.sg)
The fact that even long unstressed vowels undergo shortening and may be completely lost in Tamian (a feature unknown to other dialects) may be connected with the Livonian substratum as well. Endzelin connected it with a feature of the Baltic Finnic prosodic system, which originally distinguished quantity in root syllables only, long vowels in non-root syllables being of recent date. 4.1.4 The diphthong /au/ appears as /åu/ or /ou/: saukt kauns
såûkt2/soûkt2 kåûns2
The vowel /a/ appears as /o/ before /v/:
‘call’ (inf) ‘shame’ (nom.sg)
27
28
Laimute Balode and Axel Holvoet
sava¯da¯ks
sova¯daks
‘different’ (nom.sg)
Tamian /ou/ < /au/ and /ov/ < /av/ might be of Livonian origin, cf. Livonian
Estonian
meaning
loùl ovàr
laul avar
‘song’ ‘open, spacious’
However, Livonian influence is not certain here, because this feature has been observed (less consistently) in other dialects as well. 4.1.5 In some Tamian dialects, /e˜/ develops into /ei/, especially before obstruents and with broken tone, e.g. SL
TD
meaning
¯ešana se¯d
eîšan2 seîd2
‘eating’ (nom.sg) ‘sit’ (pres.3)
4.1.6 In some Tamian dialects, dental and alveolar consonants are palatalized before front vowels stem-finally, i.e. between stem and affix, and this feature also occurs where an original front vowel has been lost. SL
TD
meaning
atna¯cis sacı¯t
atnãc’s [Áatna˜ tësj˜] sac’t [satësj˜t]
‘come’ (part.pret.act.masc.sg.nom) ‘say’ (inf)
This type of palatalization is also attested in Livonian and Estonian, where it is associated with the (original) presence of /i/. The fact that it does not occur steminitially and is restricted to dentals and alveolars distinguishes it from the High Latvian palatalization, which extends to all consonants and is independent of their position in the word. This points to the conclusion that Tamian palatalization is not historically connected with High Latvian palatalization, but rather reflects the influence of the Livonian substratum. 4.1.7 In the domain of morphology, the most striking feature of Tamian is the tendency towards reduction of the gender system. It is generally agreed upon that this must
The Latvian language and its dialects
be due to the influence of the Livonian substratum, as the category of grammatical gender is unknown to Finnic. The extent of the substitution of masculine forms for feminine ones varies from one place to another. The most common manifestation of this tendency is the use of masculine forms of the personal pronoun (vin¸ˇc ‘he’) to refer to females. The next stage of this substitution (attested in a smaller area) is the use of masculine forms of adjectives and pronouns to agree with feminine nouns, which themselves retain their characteristic feminine endings; in the case of adjectives, this occurs both in attributive and in predicative use. SL ta¯s se¯n -es these:nom.pl.fem mushroom nom.pl TD tie se¯n -s these:nom.pl.masc mushroom nom.pl ‘these mushrooms’ SL zeme ir slapj- -a earth (fem) is wet nom.sg.fem TD ze¯m i(r) slap- -š earth is wet nom.sg.masc
Finally, feminine nouns themselves may adopt endings characteristic of masculine nouns. This amounts to the virtual elimination not only of gender as a morphosyntactic category, but also of the very forms historically connected with the feminine gender. This third stage has been reached in a still smaller area than the second. SL maz-a siev- -in¸- -a little-fem wife dim nom.sg.fem TD mas (< maz-s) siev- -i- (< -in¸-) -š little-masc wife dim nom.sg.masc
4.1.8 Apart from morphosyntactically determined changes, phonetic processes have also had their impact on the morphology of the Tamian dialects. In verbal inflection, the loss of conjugational endings as a result of the loss of final vowels has led to widespread homonymy of personal forms, at least in the singular. This pattern has been extended to the plural, which results in the generalized use of the original 3rd person form for all persons. In the following table, the conjugation of iet ‘go’ is shown for the standard language and for Tamian. Note that the homonymy of the 3rd person singular and plural in all finite verb forms (the common form continues the original singular form) is an (as yet unexplained) common feature of the Baltic languages. This homonymy may have facilitated the spread of the endingless singular form into the plural.
29
30
Laimute Balode and Axel Holvoet
es tu vin¸š me¯s ju¯s vin¸i
SL
TD
ej-u ej iet ej-am ej-at iet
es tu vin¸ˇc (viš) me¯s (míes) ju¯s vîn¸2
¸ Ô Ô ˝ iet Ô Ô ˛
4.1.9 Tamian uses the accusative singular with prepositions governing the genitive singular in other dialects. TD Es bi pe bra¯l (< bra¯li) I be:pret1 at brother:acc SL Es biju pie bra¯¸la I be:pret1 at brother:gen ‘I was at my brother’s place’
As mentioned in 1.1.3, all dialects have generalized the dative plural with all or nearly all prepositions. This pattern of government is based on that of the prepositions governing a dative plural (originally an instrumental plural) as opposed to an accusative singular (which could be an original accusative or an original instrumental). The process was triggered by the coalescence of the instrumental with the accusative in the singular and the dative in the plural. Tamian seems to have carried this process to its logical conclusion by generalizing the accusative in the singular. 4.1.10 Among the syntactic properties of Tamian, one should mention the widespread use of the dative, which has partly ousted the genitive even in adnominal position, as in Dêls í valkeí te˜í vam cepur í son:nom.sg wear:pres.3 father:dat.sg cap:acc.sg ‘The son wears his father’s cap’
as compared to standard Latvian De¯ls valka¯ te¯va cepuri. son:nom.sg wear:pres.3 father:gen.sg cap:acc.sg
This is usually described as a consequence of the loss of final vowels: tí˜ev’ can be gen.sg, acc.sg or nom.pl, whereas dat.sg tí˜evam has a clearly identifiable ending.
The Latvian language and its dialects
4.1.11 The partitive genitive with quantifiers is also virtually extinct in Tamian. Thus, with a quantifier like daudz ‘much, many’, one finds the nominative (in subject function) or the accusative (in object function), e.g. SL ju¯ra¯ daudz laivu sea:loc.sg many boat:gen.pl TD ju¯reí doudz laivs sea:loc.sg many boat:nom.pl ‘There are many boats at sea.’
The genitive of negation, used as an object case with transitive verbs and as a subject case with intransitive verbs, is rare as well. In negated existential constructions, for instance, the nominative is normally used, not the genitive as would be the case in the standard language and in the central and High Latvian dialects: ju¯rmaleí nau ve¯š beach:loc be:pres.3sg.neg wind:nom.sg ‘There is no wind on the beach.’
The partitive genitive with quantifiers and the genitive of negation are much better retained in the central dialects, and their use is most consistent in High Latvian, cf. 5.1.12. 4.2 Subdivision of the Tamian dialects This dialect group can be subdivided into Tamian proper, i.e. the dialects of northern Courland, and the related dialects in the north-west of Vidzeme. It goes without saying that the former are more closely related to the Curonian subgroup of the central dialects, and the latter to the Livonian subgroup. The influence of the Finnic (Livonian) substratum language is common to both, and may be viewed as the defining feature of the Tamian dialect group as such. The following features are peculiar to Tamian proper: 4.2.1 Vowel reduction is carried further, affecting to a larger extent not only word-final, but also word-internal syllables. Short vowels may be completely lost not only in inflectional endings, but also in derivational suffixes (if they bear no secondary stress). This rule extends to short internal vowels resulting from the shortening of original long vowels and diphthongs (and still appearing as such in the Tamian dialects of Vidzeme):
31
32
Laimute Balode and Axel Holvoet
SL
TD (Vidzeme)
TD (Courland)
meaning
sacı¯t kumelin¸š
sacit [satësit] kumelin¸ˇc
sac’t [satësjt] kûmliš
‘say’ (inf) ‘horse (poet.)’ (dim.nom.sg)
4.2.2 Some Tamian dialects replace word-final unvoiced stops (fortes) with unvoiced lenes: SL
TD
meaning
iet
ied®
‘go’ (pres.3)
This seems to be due to Livonian influence, as the Finnic languages lack voiced stops (the Fenno-Ugricists’ transcription would be ieD). 4.2.3 A few Tamian dialects of Courland have retained the archaic dative plural endings of the type -Vms, replaced by -Vm in all other dialects (in the examples below it should be noted that, phonetically, the standard Latvian forms are more archaic in that they retain the original long vowels in the endings): SL
TD
meaning
sieva¯m ma¯te¯m
sievams ma¯tíems
‘wife’ (dat.pl) ‘mother’ (dat.pl)
4.2.4 The Tamian dialects of Courland display a number of features that are not peculiar to Tamian as such, but are shared with the Curonian dialects of the central dialect group to which it is contiguous, e.g., lengthening or diphthongization of /ir/, /ur/ into /i˜r/, /ier/, /u˜r/, /uor/ etc., cf. 3.2.1.1.
The Latvian language and its dialects
5.
The High Latvian dialects (augšzemnieku dialekts)
5.1 Common features 5.1.1 The opposition between falling and drawn tone is neutralized, cf. 1.5.5.3. In the examples, (1) and (2) are homonymous: (1) lùoks2/lùks2
‘onion leaf ’ (nom.sg)
2
2
‘bow, arch’ (nom.sg)
2
2
‘window’ (nom.sg)
(2) lùoks /lùks (3) luôgs /lûgs
5.1.2 Long /a˜/ acquires a more closed articulation; according to the dialect, it develops into /o˜/, /uo/, /oa/, /a¯o/, /oa¯/ etc.: SL
5.1.4 The whole of High Latvian is characterized by a series of vowel changes conditional upon the ‘velar’ nature of the following syllable. These changes occur when the following syllable contains, or originally contained, a back vowel, as long as no /j/ nor any palatalized consonant arising from a combination with /j/ intervened between the two vowels. The scope of these changes may vary (they may, for instance, be restricted to the ‘deep’ High Latvian dialects), but, considered together, they reflect a common tendency: vowels are shifted to the back if there is no palatalizing (fronting) factor in the following syllable. These changes are sometimes referred to as the High Latvian velar vowel shift. Synchronically, it conditions a series of alternations in inflection and derivation. It should be noted that Latvian,
33
34
Laimute Balode and Axel Holvoet
as a whole, had one process dependent on similar conditions: the split of IE *e and *e¯ into /e(˜)/ and /æ(˜)/ was also connected with the presence or lack of a palatalizing factor in the next syllable. In the High Latvian vowel shift this process has been carried much farther. 5.1.4.1 Where Low Latvian has /æ/ and /æ˜/ (which is the case before original ‘back’ syllables, see 1.5.2), High Latvian has /a/ and /a˜/:
5.1.4.2 Original short /a/ shifts to /o/ under the same conditions as mentioned above for the occurrence of /æ/, i.e. when the following syllable contains a back vowel and no /j/ or palatalized consonant intervenes. In final syllables (whether open or closed) /a/ remains unchanged: SL
HLD
meaning
lapa maza
lopa moza
‘leaf ’ (nom.sg) ‘small’ (nom.sg.fem)
In some dialects the shift /a/ > /o/ is restricted to initial syllables: SL
HLD
meaning
vasara
vosora/vosara
‘summer’ (nom.sg)
5.1.4.3 In some of the High Latvian dialects (the so-called ‘deep’ High Latvian dialects) a short /i/ also undergoes changes under the same conditions as described above: it is replaced by the vowel /G/, usually transcribed ·yÒ in dialect recordings. This vowel does not exist elsewhere in Latvian; it is a high centralized vowel similar to Russian or Polish /y/. SL
HLD
meaning
miza tikums
myza tykums
‘bark’ (nom.sg) ‘virtue’ (nom.sg)
The Latvian language and its dialects
The same vowel is also found in some inflectional endings, where it has either taken the place of original /a/ (probably as a reduced vowel) or has been inserted between a stem-final cluster ‘obstruent + sonorant’ and the nom.sg ending -s (as /i/ is inserted in similar circumstances in the central dialects): tas kakls (< *kaklas)
tys kaklys
‘that’ (nom.sg) ‘neck’ (nom.sg)
5.1.5 There is a strong tendency to diphthongize the long closed vowels /i˜/ and /u˜/; /i˜/ becomes /ei/ (the first component may also be more closed), whereas /u˜/ yields either /ou/ or (in deep High Latvian) /yu/, /iu/ or /eu/: SL
5.1.6 In the deep High Latvian dialects the rising diphthongs /ie/ and /uo/ are monophthongized into /i˜/, /u˜/: SL ieva roka
[ru%ka]
HLD
meaning
¯ıva ru¯ka
‘bird-cherry’ (nom.sg) ‘hand’ (nom.sg)
A few general remarks may be made here on the changes in the High Latvian vowel system. These changes may be divided into two groups: (1) those dependent on the quality of the next syllable, and (2) those independent of phonetic environment. If we assume an interconnection to have existed between both groups, we could ascribe the changes in group (2) to shifts within the vowel system caused by the conditioned vowel changes in group (1). The whole series of changes might have been triggered by the velar vowel shift, the effect of which amounted to a rearrangement of the system of short vowels. The initial and final stage of this process are shown here as (A) and (B):
35
36
Laimute Balode and Axel Holvoet
(A)
(B)
i
u
i
e
(y) e
æ
a
u o
a
It is interesting to note that by adding /y/, the deep High Latvian dialects have achieved a short vowel system identical with the vowel system of the neighbouring Slavonic languages (Polish, Russian and Belarusian). A new /æ/ (from original /e/) occurs in High Latvian, but it is a variant of /e/ rather than an independent phoneme. The changes among long vowels and diphthongs might have been (at least partly) triggered by the change of /æ˜/, which was ‘swept along’ by its short counterpart /æ/ and passed to /a˜/. In some dialects its place was taken by a new /æ˜/ from /e˜/. Original /a˜/ had to shift towards /o˜/ or /uo/ in order to be kept distinct from the new /a˜/. Its passage to /uo/ must be related to the monophthongization of original /uo/ in some of the High Latvian dialects; another factor contributing to the monophthongization of both /ie/ and /uo/ must have been the loss of original /i˜/ and /u˜/ as a result of diphthongization. 5.1.7 In the consonant system the most striking feature of High Latvian is the (relatively slight) palatalization of consonants before front vowels: SL
HLD
ma¯te sieva ve¯jš lakstı¯gala
mùot’íe2 s’ı¯va v’iejs´ laks’t’eigola
meaning [Ámu%tjæ] [sji˜va] [vjiejsj] [laksjtjeigola]
5.1.8 As mentioned above, most High Latvian dialects do not have the palatal plosives /c/ and /L/; where these appear (normally as a consequence of interdialectal borrowing), the palato-alveolar affricates /tëw/ and /dëŠ/ are substituted: SL k¸¯eve kak¸a
HLD [ce˜ve] [kaca]
2
ˇcève kacˇa
meaning [tëwe:vje] [katëwa]
‘mare’ (nom.sg) ‘cat’ (gen.sg)
The Latvian language and its dialects
Where Low Latvian has the clusters /wc/ ·šk¸Ò, /ŠL/ ·žg’Ò, High Latvian has /wk/, /Šg/; if a front vowel follows, the velar consonant is, of course, palatalized, but it is still a palatalized velar, not a palatal. SL
5.1.9 Syllable-final /l/ is often velarized into /ł/ unless followed by a palatalized consonant; sometimes this velarized /ł/ is vocalized into /u/: SL
HLD
malka celms
[tësælms]
mòłka2 caÛûms
meaning [tësaums]
firewood (nom.sg) stump (nom.sg)
5.1.10 In the domain of morphology we find the same archaisms already mentioned for the Curonian dialects: – – –
retention of -e¯-stems in the preterite (cf. 3.1.2.1); lack of /i˜/- insertion in the future of verbs with stems in dental consonants (cf. 3.1.2.2); insertion of the reflexive marker (cf. 3.1.2.3).
Furthermore, some High Latvian dialects have preserved ancient locatives in -ie (for -i-stems) and -uo (for -u-stems), as opposed to the innovative forms in -ı¯, -û in the other dialects: SL
HLD
meaning
ausı¯ medu¯
ausie maduo
‘ear’ (loc.sg) ‘honey’ (loc.sg)
5.1.11 High Latvian uses another personal pronoun for the 3rd person than Low Latvian: jis//jys ‘he’ and jì2 // jèi2 ‘she’, cf. Lith. jìs, jì as well as Slavonic (where it is now used in the oblique cases only) Russ. jego, jemu…. Low Latvian has vin¸š, vin¸a instead, which no doubt contains the same pronoun *jis, the initial element *vin- being, however, unclear.
37
38
Laimute Balode and Axel Holvoet
5.1.12 In the domain of morphosyntax we may note the frequent use of the genitive of negation, i.e. the object of a negated verb is in the genitive instead of the accusative. This feature is also characteristic of Slavonic (especially Polish) and Lithuanian, and it is often viewed as a common feature of Baltic and Slavonic; in High Latvian, Slavonic influence has certainly contributed to its retention, whereas it is rare in Low Latvian. SL, CD ne-mazga¯ muti neg-wash:pres.3sg face:acc.sg HLD namåzgà2 muitis neg-wash:pres.3sg face:gen.sg ‘He doesn’t wash his face.’
Apart from the genitive of negation, the partitive genitive is also regularly used in High Latvian, more consistently than in the central dialects and especially in Tamian, where it is virtually extinct, cf. 4.1.11. High Latvian is, generally speaking, more archaic in morphosyntax than the remaining dialects. This can be seen in the government of prepositions, where High Latvian has not consistently carried through the generalization of the dative in the plural (cf. 1.1.3). The genitive plural is often retained with prepositions governing the genitive in the singular, and the deep High Latvian dialects may also retain the accusative plural (on this notion cf. 5.2). Tamian is most innovative in this respect, cf. 4.1.9. 5.1.13 In the lexicon, an interesting feature of High Latvian is the lack of the verb ‘to come’, which Low Latvian distinguishes from ‘to go’, like Germanic and Baltic Finnic (cf. Finnish mennä : tulla, German gehen : kommen), but unlike Slavonic and Lithuanian. Thus High Latvian has atı¯t’ = Lith. ateiti ‘to come’, cf. Russ. prijti, Polish przyjs´´c etc. Low Latvian has iet ‘to go’ and na¯kt ‘to come’, also in compounds. Perhaps Finnic influence may be suspected in Latvian. 5.2 Subdivision of High Latvian In the internal classification of High Latvian dialects, several lines of division are drawn. The western dialects form a kind of transition between the central dialect and eastern High Latvian, which has the most pronounced High Latvian features and is also referred to as ‘deep’ High Latvian (Latv. dzil¸a¯s augšzemnieku izloksnes, Germ. tiefhochlettisch). On the other hand, within both western and eastern High Latvian a distinction is drawn between Selonian dialects (Latv. se¯liska¯s izloksnes, in Upper Courland and south-eastern Vidzeme) and non-Selonian dialects (Latv.
The Latvian language and its dialects
nese¯liska¯s izloksnes, in Latgale and adjacent north-eastern Vidzeme). The intersection of these two lines of division yields four subgroups within High Latvian. 5.2.1 The main distinguishing feature of the Selonian dialects is that instead of the broken tone they have a rising tone which is not known in any other dialect group. It is marked with an acute over the second letter, e.g. SL
HLD (non-Selonian)
HLD (Selonian) meaning
gailis
gail’s’
gaíl’s’
[Ágailjsj]
‘cock’ (nom.sg)
5.2.2 The ‘deep’ High Latvian dialects are distinguished from the remaining ones by the following features, some of which were mentioned above: 5.2.2.1 monophthongization of /ie/, /uo/, cf. 5.1.6. 5.2.2.2 The rise of the vowel /G/, cf. 5.1.4.3. 5.2.2.3 occurrence of /G/ and /i/ instead of /a/ and /e/ in the inflectional endings -as and -es (gen.sg and nom/acc.pl): SL ma¯sas ma¯tes
5.2.2.4 in morphology, the diminutive suffix -en¸- instead of -in¸-: SL
deep HLD
rocin¸a
rùc’en’a2
meaning [Áru˜tësjenja]
‘hand’ (dim.nom.sg)
Note * The authors express their gatitute to Oja¯rs Busš (Riga), who kindly commented on the draft version of this paper. Of course, he bears no responsibility for its shortcommings.
39
40
Laimute Balode and Axel Holvoet
References Bielenstein, August. 1863–64. Die lettische Sprache nach ihren Lauten und Formen vergleichend und erklärend dargestellt. Bd. 1–2. Berlin: Ferdinand Dümmler. Endzelin, Janis. 1922. Lettische Grammatik. Riga: Gulbis. Endzelı¯ns, Ja¯nis. 1951. Latviešu valodas gramatika. Rı¯ga: Latvijas Valsts izdevniecı¯ba. (Revised Latvian version of the preceding.) Ga¯ters, Alfreds. 1977. Die lettische Sprache und ihre Dialekte. The Hague-Paris-New York: Mouton. Kettunen, Lauri. 1938. Livisches Wörterbuch mit grammatischer Einleitung. Helsinki: SuomalaisUgrilainen Seura. (Lexica Societatis Fenno-Ugricae, V.) Latviešu valodas dialektu atlants. Leksika. Rı¯ga, 1999. Rudzı¯te, Marta. 1964. Latviešu dialektolog’ija. Rı¯ga. Rudzı¯te, Marta. 1993. Latviešu valodas v˜esturiska¯ fone¯tika. Rı¯ga. Some of the examples from Latvian dialects have been taken from the card files of the Latvian Language Institute, Latvian Academy of Sciences, Riga.
AUTHOR "Laimute Balode and Axel Holvoet"
TITLE "The Lithuanian language and its dialects"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
The Lithuanian language and its dialects* Laimute Balode and Axel Holvoet
General data 0.1 Number of speakers Lithuanian is now spoken by about 3,500,000 people in Lithuania and about 600,000 people spread over a number of other countries. Outside Lithuania, autochthonous Lithuanian populations can be found in north-eastern Poland near the Lithuanian border and in a few villages in Belarus. No figures are available for the number of dialect speakers compared to speakers of the standard language, or for the number of speakers of the particular dialects. Before World War II, 69.2% of the population consisted of Lithuanians; the most important minorities were: Poles (15%), Jews (8%), Germans (3.4%), Russians (2.5%). In 1989, Lithuanians constituted 81% of the population, Russians 98.3%, Poles 6.9% and Belarusians 1.7%. 0.2 Territorial division The territory inhabited by Lithuanians is often referred to as ethnic Lithuania, in order to distinguish it from the historical Grand Duchy of Lithuania, which was mainly made up of territories inhabited by Eastern Slavs (13th–16th centuries, subsequently as a part of the Polish-Lithuanian Commonwealth). Outside the Grand Duchy, Lithuanian was also spoken (up to World War II) in some areas of former East Prussia, where Lithuanians settled from the 14th century onwards (the autochthonous Prussian population, never very dense, had been decimated by war). These areas, formerly known as Prussian Lithuania (Germ. Preussisch-Litauen, now largely part of the Kaliningrad District in the Russian Federation), are often referred to as Lithuania Minor, as opposed to Lithuania Maior or Lithuania propria, i.e. the ethnically Lithuanian part of the Grand Duchy. Their inhabitants were evacuated to Germany after World War II, and this was also the case with those of the Klaipe˙da region (Germ. Memelland), an originally Lithuanian area also formerly belonging to Prussia. The Klaipe˙da region is once more inhabited by Lithuanians, of course, but for the most part they are not autochthones.
42
Laimute Balode and Axel Holvoet
Telšiai gr.
Šiauliai gr.
North
Kretinga gr.
West
Zemaitian
Kupiškis gr.
Paneve· zys gr.
Varniai gr.
Anykšciai ¦ gr.
South
Utena gr.
Raseiniai gr.
East Aukštaitian West Aukštaitian Kaunas group
Širvintos gr.
Vilnius gr.
South Aukštaitian
Lithuania Maior is divided into two regions of unequal size: –
–
Lower Lithuania or the žemaicˇiai region, covering the north-western part of the Lithuanian state, amounting to about 1/4 of its territory. The original žemaı˜ˇciai ‘Lowlanders’ were one of the tribes from which the Lithuanian nation was formed. The name of this region appears in Old Russian sources as Žomoitı˘, whence Polish Z˙mudz´ and Middle High German Samaythen, the form underlying the Latinized name Samogitia. Upper Lithuania or the aukštaicˇiai region, covering the remaining part of the Lithuanian territory. The aukštaicˇiai ‘Highlanders’ (whence Middle High German Eustoythen, Auxtote etc. and Polish Auksztota), were a tribe originally inhabiting the uplands of Ošmjany, to the south-east of Vilnius, a territory now predominantly Belarusian. The name acquired its modern meaning by becoming opposed to Samogitia.
Samogitia was the last ethnically Lithuanian region to be integrated into the Lithuanian state, and even after that, it retained a certain administrative autonomy and cultural identity. Until recent times the terms ‘Samogitia’ and ‘Samogitian’ were often opposed to ‘Lithuania’ and ‘Lithuanian’, which were taken to mean only Higher Lithuania. Quite often a Žemaitian language was opposed to Lithuanian, though this did not correspond to the linguistic reality; linguists distinguish
The Lithuanian language and its dialects
Žemaitian dialects, but they extend over a smaller area than former Samogitia as a territorial unit.
1.
The Lithuanian language
1.1 Lithuanian among the Baltic languages Together with Latvian, Lithuanian is said to constitute the eastern branch of the Baltic languages, as opposed to extinct Old Prussian. Lithuanian is extremely conservative in both phonetics and morphology, and, in most cases, Lithuanian forms can be taken to represent the nearest approximation to Common East Baltic and even Common Baltic forms. As compared to Latvian, Lithuanian is characterized, among other things, by the retention of free and mobile word stress and the original two-tone system, by less pervasive loss and shortening of final vowels, and by a more conservative morphology (with better retention of the inherited IE declension classes). It cannot be established with certainty whether the important part played by palatalization of consonants is an innovation of Lithuanian (influenced by a Slavic adstratum), or a feature inherited from Common Baltic. In morphology, Lithuanian has a few innovations as well: it has developed a special habitual past tense (which is, however, lacking in some dialects, cf. 4.1.10), as well as a new type of imperative with the affix -k(i)- (both lacking in Latvian). In syntax, Lithuanian is often more archaic than Latvian, using, for instance, prepositionless case forms where Latvian uses prepositional phrases. It has preserved the genitive of negation and the predicative instrumental (generally considered to be common Balto-Slavic features), whereas Latvian has lost both. Passive constructions differ in both languages: while Latvian has only an agentless passive, Lithuanian has an agented passive, with the agent phrase in the genitive (a feature regarded as an archaism of Indo-European date by some, but as a Lithuanian innovation by others). The widespread use of participles instead of finite verb forms in subordinate clauses is probably an archaic feature of Lithuanian (better preserved than in Latvian). On the other hand, the use of participles instead of finite verb forms as a special evidential form (the so-called oblique mood) is an innovation, shared with Latvian, and with parallels in Baltic Finnic. 1.2 The formation of the Lithuanian language It is known from historical sources that the Lithuanian nation emerged from a conglomerate of Baltic tribes, and that the Lithuanians were originally but one among these. This tribe apparently took the lead in organizing a unified Lithuanian
43
44
Laimute Balode and Axel Holvoet
state in the 13th century. However, no link can be established between the original tribal structure of Lithuania and its subsequent dialectal differentiation. Lithuanian is attested in writing from the beginning of the 16th century onwards. As it was never used before as an official language, and rarely in writing (the main exception being religious literature), the formation of a standard language was a slow process. The foundations of the present-day standard language were laid in the 1880s by publicists connected with the Lithuanian national revival, whose native dialect was in most cases the South-West Aukštaitian dialect of the socalled suvalkiecˇiai region (this term refers to the Lithuanian areas belonging, in Czarist times, to the province or ‘gubernya’ of Suvalki, which derived its name from the town Suwałki in present-day Poland). Another factor contributing to the role of South-West Aukštaitian in the formation of the standard language was the prestige enjoyed by this dialect as the vehicle of the tradition of Lithuanian writing in Prussian Lithuania (where the first grammars and dictionaries were also compiled). The final codification of standard Lithuanian was only possible after the establishment of an independent state in 1918. 1.3 Substrata and adstrata The oldest population whose presence on the territory of present-day Lithuania can be established were representatives of the Narew culture, which cannot be identified with any language group. A new migration wave, identified as Fenno-Ugric, entered the Baltic area in the 3rd millennium B.C. The Finnic population did not extend over the whole of present-day Lithuania (water names of Finnic origin are found predominantly in its northern and eastern parts) and was probably not very dense. It was ousted towards the close of the 3rd millennium B.C. by the bearers of the Corded Ware and Stone Battle-Axe Culture, who entered the Baltic area from the East. These were probably Indo-Europeans, but it is not clear whether they were Balts. The next migration wave, associated with the Hatched Pottery Culture, was certainly Baltic. It extended over a large area, stretching from the river Daugava in the North to the river Pripet in the South, and the upper reaches of the Volga and Oka in the East. The eastern part of this area is now inhabited by Eastern Slavs, but numerous water names testify to the former presence of Balts. Finnic substratum features cannot be pointed out as clearly in Lithuanian as they can in Latvian, though a few phenomena are reminiscent of Finnic. So, for instance, the series of local cases (illative, adessive, allative, cf. 3.2.4) attested in Old Lithuanian and the Lithuanian dialects (only fossilized remnants are found in Latvian) seems to echo the local cases of Baltic Finnic. The Lithuanians were originally one of several Baltic tribes inhabiting what would later become the Lithuanian language area. The north-western part, more or less coinciding with the territory of Samogitia, was once inhabited by Curonians, a
The Lithuanian language and its dialects
tribe whose territory extended northwards into present-day Latvia (the Samogitians themselves must originally have lived on the plains by the river Neve˙žis). Semigalians, whose territory also extended into present-day Latvia inhabited the area further to the east (north of Paneve˙žys). The Curonian and Semigalian substrata might be responsible for some features of the dialects of these areas. So, for instance, the character of the Žemaitian tones, and their similarity to those of Latvian, might somehow be connected with the Curonian substratum. Some would ascribe stress retraction in both Žemaitian and northern East Aukštaitian dialects to a Curonian or Semigalian substratum. Needless to say, it will never be possible to substantiate this. Among features which might be connected with a Slavic adstratum, so-called dzu¯kavimas, i.e., the affrication of dental plosives before high front vowels, must be mentioned. It occurs in the South Lithuanian area adjacent to Belarusian, where a similar affrication is observed (Polish affrication is of a different type). However, most scholars now explain dzu¯kavimas as an independent development in Lithuanian. 1.4 External influences From the 10th century onwards, the Lithuanian tribes, though successfully opposing the expansionism of Kievian Russia and the northern Russian principalities, were exposed to East Slavic influence. This is reflected in a number of loan words whose early date is attested by their archaic sound shape. The Lithuanian state which was formed in the 13th century expanded mainly into Ruthenian (Western and North-Western Russian) territories, and the Ruthenian element became prevalent in the culture, language and religion. The only language to be used in writing was the northwestern variety of Old Russian Church Slavic (the Ducal Chancery style). The separate development of the East Slavic dialects within the Grand Duchy of Lithuania eventually led to the rise of the Belarusian (White Russian) language. As a result of the Polish-Lithuanian Union (1569–1795), Polish language and culture became predominant among the upper classes from the 16th century onward. Lithuanian was hardly ever used in public life or for literary purposes, and even the Church did not encourage its use in religious life. Lithuanian survived only among the peasantry. In writing, Lithuanian was cultivated mainly in Prussia from the 16th century onward, under the influence of the Reformation. The symbiosis of languages and cultures within the Grand Duchy of Lithuania (Lithuanian, Belarusian, Polish) gave rise to a situation of widespread bilingualism even among the rural population, and Lithuanian was thus exposed to a strong Slavic (Belarusian and Polish) influence, which made itself felt in vocabulary, phraseology and syntax. The language developed a system of regular sound substitutions for Slavic loan words. The impact of (Great) Russian began to be felt after the incorporation of Lithuania into the Russian Empire (1795), especially after 1865, when the Czarist authorities began to enforce Russification and still further
45
46
Laimute Balode and Axel Holvoet
restricted the use of Lithuanian through administrative measures. The pervasive Slavic influence in the lexicon as well as in syntax was resisted by Lithuanian writers, philologists etc. from the 1880s onward (the Lithuanian national revival), and superfluous Slavic loan words were ousted from the standard language. A policy of Russification was again pursued after the incorporation of Lithuania into the Soviet Union in 1940. 1.5 Phonemic inventory 1.5.1The vowel system Vowels Short i ·iÒ
u ·uÒ o ·oÒ æ ·eÒ
a ·aÒ
Long /i˜/ ·y, ˛iÒ
/u˜/ ·u¯, u˛Ò /e˜/ ·e˙Ò
/o˜/ ·oÒ /æ˜/ ·e, e˛Ò
/a˜/ ·a, a˛Ò
Of the short vowels, /o/ occurs only in relatively recent loan words, such as kodas ‘code’. Some speakers of Lithuanian have a short /e/ ·eÒ, distinct from /æ/ ·eÒ, in such loan words as metras ‘metre’, but the autonomous existence of this vowel is not well enough established to justify its inclusion into the phonemic inventory. Diphthongs Rising diphthongs /ie/
/uo/
The first components are /i/ and /u/ respectively, but the second part is hard to identify (it is a centralized vowel /6/ rather than /e/ resp. /o/ in the dialects, these sounds are often realized as monophthongs. They also tend to lose tone differences in those dialects where long monophthongs lose them. For these reasons /ie/ and /uo/ are often described as non-homogeneous monophthongs rather than as diphthongs.
The Lithuanian language and its dialects
Falling diphthongs ui (oi) ei
ai
eu
au
/oi/ exists only in recent loan words such as boikotas ‘boycot’; /ui/ is exceedingly rare. Apart from loan words like Europa ‘Europe’, /eu/ exists only as a front counterpart of /au/ after palatalized consonants and is then written /au/, e.g. macˇiau [matëwjeu] ‘see’ (pret.1sg) 1.5.2Origins of the vowel system Lith. /a/ is the result of the merger of IE *a and *o. Short /o/ was reintroduced into the standard language quite recently, along with loan words (until then foreign /o/ was rendered as /a/, cf. Lith. (obs., reg.) gaspadorius ‘farmer, landlord’ from Polish gospodarz). /æ/ arises not only from IE *e, but also from original /a/ when preceded by a palatalized consonant, e.g. ve«rˇcia ‘turn’ (pres.3) is pronounced [«vjærjtëwjæ ] (/a/ and /e/ are thus in complementary distribution everywhere except word-initially). The long high vowels /i˜/ and /u˜/ are from IE *ı¯, *u¯, but also nasal /ı˜/ and /u˜/, from the sequences /iN/ and /uN/ (in the latter case, they are written ·i˛Ò, ·u˛Ò). The long mid vowels /e˜/ and /o˜/ are from IE long *e¯, *a¯. IE *e¯ was probably represented in Common Baltic by a broad /æ/, *a¯ by Common Baltic /a˜/; the narrowing occurred in the separate history of Lithuanian. The long low vowels /æ˜/ and /a˜/ have arisen from former nasal /ã/ and /«e/ (in this case, they are written ·a˛Ò, ·e˛Ò), but under stress they may also come from original short /æ/ and /a/, which were lengthened in this position while remaining short elsewhere. In the latter case, they are not distinguished from short /æ/ and /a/ in normal orthography; they automatically acquire circumflex tone, e.g. nãmas ‘house’ (compared to nami«e ‘at home’, where the unstressed /a/ has remained short).
In this table the representation of the sounds in standard orthography is given in angle brackets (unless trivial). All palatalized consonants are rendered by the addition of the letter ·iÒ if followed by a back vowel, e.g. liu¯tas [lju˜tas] ‘lion (nom.sg)’; when they are followed by a front vowel, palatalization it is not marked at all, as all consonants are palatalized in this position, e.g. dešine˙ [djæwji«nje˜] ‘right(-hand)’ (nom.sg.fem). Word-finally, palatalized consonants are replaced by their non-palatalized correlates (the opposition is neutralized in the standard language, though not in all dialects), e.g. eit [eit] ‘go’ (inf), shortened form of eiti [eitji] ‘id.’. Before another consonant, palatalization is not marked because the last consonant of a cluster always imposes its palatalized or non-palatalized articulation on the preceding ones, e.g. kélsiu [´kjæ.ljsju] ‘lift’ (fut.1sg) compared with ke˜ls [«kjæl.s] ‘id.’ (fut.3). 1.5.4Origins of the Lithuanian consonants The voiceless plosives come from similar IE consonants; voiced plosives may correspond to IE aspirated or unaspirated voiced plosives, e.g. /d/ may be from *d or *dh. The dental fricative /s/ is inherited, /z/ occurs mainly in loan words and onomatopoeia (or as an allophone of /s/). The palato-alveolar fricatives originate from IE palato-velar plosives, e.g. šimtas « ‘hundred’ (Lat. centum), žinóti ‘know’
The Lithuanian language and its dialects
(inf) (cf. Old Lat. gnosco). The soft palato-alveolar affricates arise from the clusters *-tj-, *-dj-, e.g. gaı˜džio < *gaid-j-a¯ ‘cock’ (gen.sg) alongside gaid«ys < *gaidij(a)s ‘id.’ (nom.sg). The corresponding non-palatalized affricates are rare apart from loan words. The affricates /tës/, /dëz/, /tësj/ and /dëzj/ are exceedingly rare in standard Lithuanian (except for loan words and onomatopoeia), but they are frequent in the ‘Dzukian’ dialects, see 3.2.2 below. The phonemes /x/, /xj/, />/, />j/ and /f/, /fj/ were introduced into the standard language at the end of the 19th century, along with foreign loans. All palatalized consonants have arisen either (1) by assimilatory palatalization before front vowels, or (2) from combinations with /j/. The phonological relevance of palatalization is established on the basis of minimal pairs with oppositions before back vowels, e.g. gilus [gjilus]‘deep’ (nom.sg.masc) vs. gilius [gjiljus]‘deep’ (acc.pl.masc). 1.5.5Prosodic features 1.5.5.1 Stress. Lithuanian stress is free and mobile, i.e. not only can it fall on any syllable of the word, but it often varies within inflectional paradigms. The system of word stress is closely bound up with that of tones or syllable accents (see below). This connection is of a twofold nature: (1) when stress falls on a syllable containing a long vocalic segment (on this notion cf. 1.5.5.3), it must obligatorily be realized in one of two ways: as acute or circumflex tone or syllable accent. Within lexical, derivational and grammatical morphemes, tone is, in principle, constant, i.e. whenever a given morpheme bears word stress, it displays a constant tone; change of tone (metatony) may, however, occur as a derivational device. So, for example, in laı˜svas ‘free’ the stem laisv- does not always bear stress, but when it does, it will be realized as circumflex tone; in láisve˙ ‘freedom’ the change of circumflex into acute tone is a marker of derivation (together with the change of declension type). (2) the tone of the stem syllable (or, in the case of polysyllabic stems, the syllable that bears stress within the stem) influences the pattern along which word stress may vary. Stress may be fixed (i.e., remain on the stem in all inflectional forms) only when it bears acute tone on the last stem syllable (as in vý-ras ‘man, husband’), or stress of any kind on a non-final stem syllable (as in vãsara ‘summer’, gárbana ‘curl’, nùgara ‘back’), otherwise it is always mobile within the paradigm. The patterns along which stress may vary are called accent paradigms; four of them are distinguished for nouns. 1.5.5.2 Quantity. Vowel length is distinctive in both stressed and unstressed syllables. In unstressed syllables, however, long vowels are realized as half-long, and
49
50
Laimute Balode and Axel Holvoet
there is a strong tendency for vowel length to be neutralized both in dialects and in the standard language. When a long vowel is stressed, it displays one of the two tones, (see below). Thus, under stress, length cannot be realized independently of tone. 1.5.5.3 Tone (syllable accent). Tones can be defined as alternative ways of realizing word stress in the case of long vocalic segments. These include: – – –
long vowels; diphthongs (both rising and falling); and diphthongoid sequences of the type /V + R/.
Though the sonorants /r/, /l/, /m/, /n/ (and /]/) cannot act as vowels in the sense of syllabic nuclei, their combinations with vowels behave like diphthongs as far as tonal differences are concerned. So, for instance, the sequence /um/ in trumpas « ‘short’ is marked for tone, just like the diphthong /ui/ in muı˜las ‘soap’ (nom.sg); but in lùpti ‘peel’ (inf) the first syllable simply contains a short vowel incapable of displaying tonal differences. When word stress falls on a long vocalic segment, it is realized as either acute or circumflex tone. This means that the constitutive features of stress, intensity (expiratory strength), pitch (tone height) and length, are spread over the vocalic segment in different ways. They can either be concentrated on the initial part (mora) of the vocalic segment, or they can be spread more or less evenly over the vocalic segment. Accordingly, we distinguish: – –
the acute tone, also called sharp (Lith. staigine˙) or, less appropriately, falling tone, and the circumflex tone, also called even or drawn (Lith. te˛stine˙) or, less appropriately, rising tone.
The distinctive nature of the opposition is shown by the following minimal pair: klóste˙ klõste˙
‘spread, pleat’ (pret.3) ‘pleat, crease’ (nom.sg)
Tone differences are most pronounced in the case of diphthongs and diphthongoid sequences, because these consist of articulatorily heterogeneous segments, the relative prominence of which is more clearly marked. The vocalic segment is then, as it were, split into two more or less distinct units (morae), the first or the last of which can be more prominent. Compare: láiškas laı˜kas
[¢la.iwkas] [«l6i.kas]
‘letter’ (nom.sg) ‘time’ (nom.sg)
There is a difference in relative length between both components; furthermore, the first vowel acquires a more closed and centralized articulation with circumflex tone,
The Lithuanian language and its dialects
and tends to merge with the second component to a larger extent than with acute tone. The realization of tones in the case of diphthongs seems to match the traditional notions of ‘falling’ and ‘rising’ tone most accurately. In the case of long vowels, the distinctive features described for diphthongs do not apply, as nothing seems to justify the distinction of two morae within the vocalic segment. But here as well, tone distinctions are accompanied by qualitative differences: vowels with the acute tone have been described as more tense, so that the above-mentioned forms klóste˙ and klõste˙ could almost be transcribed [´klo˜sjtje.] and [«klf˜sjtje.] respectively. Vowels with circumflex tone are, on average, also slightly longer than those with acute tone. The tones of the Žemaitian dialects differ considerably from the Aukštaitian and standard tones described here, cf. 4.1.8. In many Aukštaitian dialects, there is a tendency towards loss of tone differences in long monophthongs and in the rising diphthongs /ie/ and /uo/; at least, the distinction between the tones is perceptually slight. The tones are retained much better in the case of falling diphthongs, where their differentiation is facilitated by the differences in relative prominence of the initial and final components. 1.5.5.4 The use of tone marks. In principle, the circumflex accent (˜) is used to mark circumflex tone, the acute (´) to mark acute tone, and the grave (`) to mark a stressed short vowel, i.e. the lack of tonal differentiation (as only long segments are capable of distinguishing tones). There is, however, one exception from this rule. As mentioned above, in the case of diphthongs and diphthongoid sequences there is a marked difference in relative length between the two components. If the first component is /a/ or /æ/, it becomes half-long with the acute tone, but if the first component is /u/ or /i/, it is not perceptibly lengthened in the literary language (though it is in some dialects). In this case the acute tone is marked by putting a grave accent over the first letter in order to avoid the suggestion that it is lengthened, e.g. tvìrtas [´tvirtas] ‘firm’ (as opposed to tvártas [´tva.rtas] ‘stable, sty’). Here the grave marks acute tone, whereas in kìtas ‘other’ it marks stress and lack of tone. Moreover, the grave is used to mark the acute tone in a number of loan words, where no lengthening of the first component occurs, e.g. fèrma ‘farm’, spòrtas ‘sport’.
2.
The Lithuanian dialects
2.1 Classification The classification of Lithuanian into two main dialects, Aukštaitian and Žemaitian, has been known for centuries; they were often described as separate languages (‘Žemaitian’ being opposed to ‘Lithuanian’). The first attempt at a classification in
51
52
Laimute Balode and Axel Holvoet
modern times was made at the close of the 19th century by Antanas Baranauskas (Baranowski) and Kazimieras Jaunius. During the period of Lithuanian independence, their classification was improved by professional linguists such as Antanas Salys. The classification used nowadays was proposed after World War II by Zigmas Zinkevicˇius and Aleksas Girdenis, but in its basic shape it goes back to Baranauskas. 2.2 The differentiating features of the Lithuanian dialects The criteria used in classifying the Lithuanian dialects are: (1) In the vowel system: – – – – –
the treatment of the original sequences /aN/ and /eN/ (the main criterion in classifying the Aukštaitian dialects); the treatment of the diphthongs /ie/ and /uo/ (the main criterion in classifying the Žemaitian dialects and setting them apart from Aukštaitian); the treatment of the long vowels /e˜/ and /o˜/; the treatment of the short vowels /i/ and /u/; vowel loss in endings.
(2) In the consonant system: – the distribution of the dental plosives /t/ and /d/ and the affricates /tëw/ and /dëŠ/ (the Žemaitian Sound Law, used to set apart Žemaitian from Aukštaitian); – occurrence of palatalized and non-palatalized /l/ before non-high front vowels. (3) Prosodic features: – stress retraction (used in classifying both Žemaitian and Aukštaitian dialects); – the character of tones (sets apart Žemaitian from Aukštaitian).
3.
The Aukštaitian (High Lithuanian) dialects (aukštaicˇiu˛ tarme˙s)
3.1 Criteria of classification The main criterion used in classifying the Aukštaitian dialects is the treatment of the Baltic diphthongoid sequences of the type /aN/ and /eN/. Everywhere in Lithuanian, the nasal sonorant is preserved in such sequences when it occurs before a plosive, e.g. rankà ‘hand’ (nom.sg). However, where such a sequence occurred before a fricative or another sonorant, or word-finally, it was changed into a nasal vowel, which subsequently lost its nasal resonance and become a long oral vowel (still marked as a ‘nasal vowel’ in orthography, see above). If the vowel occurring in an original diphthongoid sequence with a nasal was *a or *e, this sequence may be of three kinds:
The Lithuanian language and its dialects
(1) in West Aukštaitian, the vowel did not undergo any change, and the diphthongoid sequence is reflected as either /aN/ and /æN/ or /a˜/ (·a˛Ò), /æ˜/ (·e˛Ò): Proto-Lith.
SL and WA
*ranka¯ *kampas *menkas *tempja *žansis *trenšja
rankà kampas « me«nkas tempia « ža˛sìs tr«e˛šia
meaning [ra]Áka] [«kam.pas] [«mjæ].kas] [«tjæmj.pjæ] [Ša.Ásjis] [«tjrj˜æwjæ]
(2) in South Aukštaitian, the vowel underwent no change when the sequences /aN/ and /æN/ were retained as such, but when a nasal vowel arose from such a sequence, it was narrowed into nasal /u˜˜/ and /«˜/, eventually giving rise to long oral /u˜/ and /i˜/: SL and WA
(3) in East Aukštaitian, the development of the nasal vowels arising from /aN/ and /æN/ was similar as in South Aukštaitian, but, in addition, the vowels /a˜/ and /æ˜/ were narrowed into /u/ and /i/ when the nasal sonorants were also retained: SL and WA
3.2 West Aukštaitian (vakaru ˛ aukštaicˇiu˛ tarme˙s) West Aukštaitian is in its turn divided into a southern subgroup, the dialects of the Kaunas region, and a northern subgroup, those of the Šauliai region. The latter are
53
54
Laimute Balode and Axel Holvoet
more innovative than the former: they are characterized by stress retraction and by the neutralization of vowel quantity in unstressed syllables.
˛ tarme˙s) 3.2.1The dialects of the Kaunas region (kauniškiu These dialects are sometimes referred to as dialects of the suvalkiecˇiai. The reason is that within Czarist Russia the major part of this region belonged to the province (‘gubernya’) of Suwalki (Polish Suwałki, a town situated in present-day Poland). As mentioned above (in 1.2), the standard language is based on this dialect group. The dialects of the Kaunas dialect happen to be the most conservative Lithuanian dialects, particularly with regard to the prosodic features described above for the literary language: 1. lack of stress retraction; 2. long vowels are distinguished form short ones even in unstressed position. The characteristics of this dialect can therefore, for the most part, be formulated negatively, as the lack of the features enumerated further on for the other dialects. The phonological system can be identified with the one described for the standard language in 1.5 above; as to its place within Aukštaitian cf. 3.0 above. The following features distinguish the Kaunas dialect from the standard language (most of them can also be found in one or other of the remaining dialects): 3.2.1.1 In some dialects of this group the oppositions of /a/ and /æ/, /a:/ and /æ˜/, /ai/ and /ei/ are neutralized word-initially in favour of /a/, /a˜/ and /ai/: SL
WA (Kaunas)
meaning
«ežeras eı˜ti
ãžeras a«Ht
‘lake’ (nom.sg) ‘go’ (inf)
As mentioned above (cf. 1.5.2), /a/ and /æ/ are in complementary distribution in all positions except word-initially, hence a strong tendency to neutralize the opposition in this position as well. It recurs in several dialects (cf. 3.1.2.5, 3.3.0.4 below). 3.2.1.2 In diphthongoid sequences of the type /V + R/ there is a tendency to lengthen the vowels /u/ and /i/ when the tone is acute (falling), whereas in the standard language this process is restricted to the vowels /a/ and /æ/ (see above, 1.5.2). This feature may be observed especially in the south-eastern part of the Kaunas region. In transcriptions, it is marked by using the acute instead of the grave accent (as in the case of the sequences /a + R/ and /æ + R/), e.g.
3.2.1.3 Prothesis of /v/ before a word-initial vowel /o˜/ or diphthong /uo/ is observed in part of the Kaunas dialects: SL
WA (Kaunas)
meaning
uodegà óras
vuodegà vóras
‘tail’ (nom.sg) ‘air, weather’ (nom.sg)
3.2.1.4 The main prosodic feature deviating from the standard language is the circumflex tone of a few suffixes which have acute tone in standard Lithuanian: -ybe˙ ‘-ness, -hood’, -ytis and -aitis (diminutive suffixes). SL
3.2.1.5 In the conjugation of verbs in -inti, -enti a different type of paradigmatic levelling has taken place from what is prescribed by the standard language. Sequences of the type /V + N/ normally undergo different treatments according to the nature of the following consonant (retention before obstruent, rise of a long oral vowel through an intermediary stage with a nasal vowel before a fricative or sonorant, cf. above, 3.0). This should give rise to alternations in the conjugation of the verbs in question. The standard language generalizes the stem variants in /æn/ and /in/: gyvenù gyvénti gyvénsiu
[gji.vjæÁnu] [gji.¢vjæ.njtji] [gji.¢vjæ.njsju]
‘live’ (pres.1sg) ‘live’ (inf) ‘live’ (fut.1sg)
Here the future has retention of /æN/ where /æ˜/ would be expected before a fricative. In the Kaunas dialects described here, levelling has taken the opposite direction, and the long oral vowel is generalized:
55
56
Laimute Balode and Axel Holvoet
SL
WA (Kaunas)
gyvénti gyvénk skõlinti skõlintu˛
gyv¢e˛ti gyv¢e˛k skõlyti skõlytu˛
meaning [gji.¢vjæ˜tji] [gji.¢vjæ˜k] [«sko˜lji.tji] [«sko˜lji.tu.]
3.2.1.6 In the north-western part of the Kaunas dialect group the dual is still used as a living inflectional form, whereas in the remaining Aukštaitian dialects it has virtually gone out of use (in Žemaitian, it is also retained in verbs): SL
3.2.2The dialects of the Šiauliai region (šiauliškiu ˛ tarme˙s) 3.2.2.1 In the prosodic system, the neutralization of the opposition of quantity in unstressed vowels is observed. In the case of unstressed high vowels, this results in shortening without qualitative changes: SL su¯n«elis gyv«enimas
WA (Šiauliai) [su.«njæ˜ljis] sun«elis [gji.«vjæ˜njimas] giv«e nimas
meaning [su«njæ˜ljis] [gji«vjæ˜njimas]
‘little son’ (nom.sg) ‘life’ (nom.sg)
In the case of the unstressed mid vowels /e./ and /o./ this shortening results in short /6/ and /f/, but there may also occur narrowing into /i/ and /u/: SL
3.2.2.2 On the other hand, the short high vowels /i/ and /u/ are somewhat lengthened under stress, yielding half-long vowels /i./ and /u./. As mentioned in 1.5.2, in the standard language this lengthening is characteristic only of the open vowels /æ/
The Lithuanian language and its dialects
and /a/. The half-long vowels resulting from this lengthening are apparently not susceptible to tone distinctions; Lithuanian dialectologists ascribe it an “intermediate” tone marked ·sÒ. Phonologically, this half-long pronunciation is evidently a variety of shortness, and the so-called ‘intermediate’ tone cannot, therefore, properly be considered a tone. SL pùse˙ kìšti
WA (Šiauliai) [Ápusje.] [Ákjiwjtji]
s
pu se˙ kisšti
meaning [pu.sje.] [kji.wjtji]
‘half ’ (nom.sg) ‘cram, thrust’ (inf)
3.2.2.3 Compared with the standard language and the Kaunas dialects, there is frequent stress retraction from endings onto stem syllables. There is some divergence between the dialects of this group, but usually stress is retracted only from a short final vowel onto a long vowel or diphthong in the stem: SL
WA (Šiauliai)
meaning
plytà rankà
pl«yta rañka
‘brick’ (nom.sg) ‘hand’ (nom.sg)
In the north of this region stress retraction is more general, and stress can be retracted from a long circumflected vowel in the ending onto a short vowel in the stem: SL
WA (Šiauliai)
meaning
šak«u˛
šàku˛
‘branch’ (gen.pl)
3.2.2.4 Word-final vowels are not preserved as well as in the Kaunas dialects: in final syllables short /a/ is usually apocopated (though short /a/ resulting from shortening of Proto-Baltic *a¯ is not affected by this process): SL
WA (Šiauliai)
meaning
výras dvãras
výrs dvárs
‘man, husband’ (nom.sg) ‘estate’ (nom.sg)
Note that as a result of apocope the sequence /ar/ acquires the status of a tautosyllabic diphthongoid sequence, which automatically acquires acute tone. 3.2.2.5 Word-initially the opposition between /a/, /a˜/, /ai/ and /æ/, /æ˜/, /ei/ is neutralized in favour of /a/, /a˜/, /ai/, as in some of the dialects of the Kaunas region (cf. 3.1.1.1):
57
58
Laimute Balode and Axel Holvoet
SL «ežeras eı˜ti
WA (Šiaul) a«žeras aı˜t
[«æŠjæras]
meaning [«a˜Šjæras]
‘lake’ (nom.sg) ‘go’ (inf)
3.3 South Aukštaitian (pietu˛ aukštaicˇiu˛ tarme˙s) This dialect group is also referred to as Dzu¯kian (dzu¯ku˛ tarme˙s), a name that is based on one of the most striking features of these dialects, so-called dzu¯kavimas (cf. 3.2.2). This use of the term is not accurate, however, as dzu¯kavimas occurs also in some East Aukštaitian dialects (the main criterion of classification being the treatment of /aN/ and /eN/ rather than dzu¯kavimas). 3.3.1 With regard to the chief criterion of classification, the treatment of the original sequences /aN/ and /æN/, the characteristic features of South Aukštaitian were already mentioned (cf. above, 3.0). These dialects have /aN/ and /æN/ without any change of the vowel, as in West Aukštaitian, but wherever these sequences have developed into nasal vowels, with subsequent loss of nasalization, the characteristic narrowing into /u˜/ and /i˜/ is observed: SL ža˛sìs vaı˜ka˛ tr«e˛šia kãte˛
SA [Ša.Ás’is] [«vai.ka.] [«tjrjæ˜wjæ] [«ka˜t’æ.]
žu¯sìs vaı˜ku˛ tr«yšia kãty
meaning [Šu.Ás’is] [«vai.ku.] [«tjrji˜wjæ] [«ka˜t’i.]
3.3.2 Almost all dialects of this group are characterized by a series of sound changes referred to by the term dzu¯kavimas. This term is derived from the name dzu¯kai (singular dzu¯kas), which, in principle, is used to refer to the speakers of dialects where this dzu¯kavimas occurs, though, in practice, it usually refers to the South Lithuanians. The name dzu¯kas is onomatopoeic and suggestive of the high frequency of the affricates /tës/, /tësj/, /dëz/ and /dëzj/ which are lacking in other dialects. Dzu¯kavimas comprises two processes: (1) The affrication of /tj/ to /tësj/ and /dj/ to /dëzj/ before the high front vowels /i/ and /i˜/ (regardless of whether it developed from Proto-Baltic *ı¯ or *iN), as well as before the diphthong /ie/, the first component of which is also a high vowel.
The Lithuanian language and its dialects
SL tìk pi«r ti˛ tylùs tìek dìdis gaı˜di˛ dienà
SA [tjik] [«pjirj.tji.] [tji.Álus] [´tjiek] [Ádjidjis] [«gai.dji.] [djieÁna]
This affrication is not impeded by /v/ intervening between the dental and the vowel: tviñko dvýlika
[«tjvji].ko.] [´djvji.ljika]
cviñko dzvýlika
[«tësjvji].ko.] [´dëzjvji.ljika]
‘swell’ (pret.3) ‘12’ (nom/acc)
Affrication does not occur before the other front vowels, /e(˜)/ and /æ(˜)/, which is obviously connected with the degree of palatalization: not only is palatalization usually weaker before the mid front vowels than before high front vowels, but it is also conceivable that assimilatory palatalization had not yet affected consonants followed by mid front vowels at the time when the affrication developed. Affrication is also lacking before a long /i˜/ arising from original *eN (as described in 3.2.1), which shows that dzu¯kavimas had ceased to operate when nasal /«e/ was narrowed. (2) The occurrence of /tës/ and /dëz/ (or /tësj/ and /dëzj/) as a result of the combinations of /t/ and /d/ with /j/, corresponding to /tëwj/ and /dëŠj/ of the other Aukštaitian dialects. Proto-Baltic SL
SA
(?) *svet-j-ai
ˇcià svecˇiaı˜
[sjvjæ« tëwjei.]
*med-j-ai
m«edžiai
[«mjæ˜dëzjei]
cià [tësjæ] sveciaı˜ svecaı˜ m«edziai m«edzai
meaning // cà [tsa] [sjvjæ«tësjei. ] [sjvjæ« tësai.] [«mjæ˜dëzjei] [«mjæ˜dëzai]
‘here’ ‘guest’ (nom.pl) ‘tree’ (nom.pl)
This second form of dzu¯kavimas extends over a slightly larger area than the affrication before front vowels (on East Lithuanian territory, its isogloss stretches north beyond that of affrication). It is clear that, given the affrication of the dentals before high front vowels, the occurrence of /tësj/ and /dëzj/ instead of /tëwj/ and /dëŠj/ implies a simplification of the morphophonological system: a binary alternation of the type /tj/ : /tësj/ is substituted for the ternary alternation /tj/ : /tësj/ : /tëwj/ which would otherwise arise as a result of affrication. It is therefore conceivable that the occurrence of / tësj/ and / dëzj/ instead
59
60
Laimute Balode and Axel Holvoet
of / tëwj/ and / dëŠj/ results from substitution rather than from regular phonetic development, but this cannot be proved. 3.3.3 A non-palatalized /l/ occurs, as opposed to the palatalized /lj/ of the standard language and most other dialects, when it is or has been followed by a mid or low front vowel (in the Lithuanian dialectologists’ transcription, this non-palatalized /l/ is rendered ·łÒ before a mid or low front vowel): SL
SA [«lje˜kjtji]
lÙekti
ł«ekc
meaning [«le˜ktës]
‘fly’ (inf)
As a concomitant of the lack of palatalization, /æ / and /æ˜/ were replaced by their back counterparts /a/ and /a˜/ (/a/ and /æ/ being in complementary distribution, as described in 1.5.2): l«edas
lãdas
‘ice’ (nom.sg)
A similar development can be observed when /l/ was originally followed by a nasal vowel continuing the sequence *eN. As described above (3.2.1), such a nasal vowel normally underwent narrowing into /i˜/; however, after /l/, which failed to undergo palatalization before a non-high front vowel, no /i˜/ could arise, as this vowel requires the preceding consonant to be palatalized; instead, a high centralized vowel arose, reminiscent of Russian y (/G/), and rendered by the corresponding Cyrillic letter in the Lithuanian dialectologists’ transcription: sáule˛
[´sa.uljæ.]
sául9
[´sa.ulG.]
‘sun’ (acc.sg)
3.3.4 In morphology, a characteristic feature of South Aukštaitian is the retention of the illative. It is a remnant of a series of local cases which are well attested in Old Lithuanian (traces can be found in Latvian as well), but have for the most part fallen into disuse in the standard language and the dialects. Four such cases can be reconstructed: –
the inessive (simply called ‘locative’), the only one to have been retained everywhere, corresponds to the IE locative, but has been renewed as to its form (usually an affix *-en, corresponding to the IE preposition *en, was added): e.g. mieste ‘in town’
The Lithuanian language and its dialects
–
(where the final -e results from the shortening of an original nasal vowel reflecting *-eN); the illative, being the lative counterpart of the inessive, and consisting in the original accusative with the affix -na added, e.g. miestan ‘into town’ (with -an from *-an-na);
–
–
the adessive, denoting proximity, and consisting of the original IE locative with the affix -pi added, e.g. Old Lith. Joniep (Germ. bei Johannes) ‘in (the Gospel of) John’ (lit. ‘near, close to John’); the allative, the lative counterpart of the adessive, denoting approach, movement towards a person or object, and consisting of the genitive with the affix -pi added (the modern standard language still has a few fossilized relics of this case): Old Lith. MS Lith.
Papiežauspi ‘to the Pope’ velnióp ‘to the Devil’
The adessive and the allative have fallen into disuse almost everywhere (except for fossilized relics), but the illative still occurs in some dialects as an alternative to constructions with the preposition ‘to, into’ (it may also occur in the standard language as used by those writers who have the illative in their native dialect): SL
SA
meaning
einu ˛i miesta˛
ainu miestan
‘I’m going to town’
3.3.5 In declension, South Aukštaitian has retained a number of archaisms. The old IE *i-stems and consonant stems have retained some inherited endings for which most dialects and the standard language have substituted analogical endings borrowed from other declensions: SL
Here the SA ending -i is the original ending of consonant stems, whereas -ui has been borrowed from the more productive *o-stems; nom.pl -es is the inherited IE ending for consonant stems, whereas -ys is the refashioned ending for *i-stems.
61
62
Laimute Balode and Axel Holvoet
3.3.6 In conjugation, on the other hand, the conditional shows a few innovative forms, refashioned after the pattern of the past tense: SL
SA
meaning
eı˜ˇciau eı˜tum
aitáu aitái
‘go’ (cond.1sg) ‘go’ (cond.2sg)
3.4 East Aukštaitian (rytu˛ aukštaicˇiu˛ tarme˙s) 3.4.1Common features of East Aukštaitian 3.4.1.1 With regard to the treatment of the original sequences /aN/ and /eN/ mentioned above (3.0), East Aukštaitian is characterized by consistent narrowing of the vowel, regardless whether the nasal sonorant is kept intact or disappears through the intermediary stage of a nasalized vowel: SL
3.4.1.2 As in the Šiauliai dialects of West Aukštaitian (cf. 3.1.2.2), the short high vowels /u/ and /i/ tend to be lengthened under stress; in some dialects of East Lithuanian, the opposition between original /u/ and /u˜/, /i/ and /i˜/ may be completely neutralized. In the latter case, they acquire circumflex tone; if lengthening is incomplete, dialects ascribe ‘intermediate’ tone (·sÒ) to these vowels: SL
EA
pùse˙
[Ápusje.]
mH`škas
[Ámjiwkas]
pu˜se˙ pusse˙ m«yškas misškas
meaning [«pu˜sje.] [´pu.sje.] [«mji˜wkas] [´mji.wkas]
‘half ’ (nom.sg) ‘wood’ (nom.sg)
The Lithuanian language and its dialects
3.4.1.3 As in South Aukštaitian, non-palatalized /l/ appears before mid and low front vowels and their reflexes, cf. 3.2.3. SL
EA
l«edas sáule˛
[«ljæ˜das] [´sa.uljæ.]
lãdas sául9
meaning [«la˜das] [´sa.ulG.]
‘ice’ (nom.sg) ‘sun’ (acc.sg)
3.4.1.4 As in some of the West and South Aukštaitian dialects, the opposition between /a/ and /æ/ is unstable word-initially. In the case of diphthongs, neutralization in favour of /ai/ is general, while the merger of /a(˜)/ and /æ(˜)/ into /a(˜)/ is restricted to the eastern dialects of East Aukštaitian: SL
EA
meaning
eı˜ti «ežeras
aı˜t ãžeras
‘go’ (inf) ‘lake’ (nom.sg)
3.4.1.5 In morphology, the retention of the illative is noted as in many South Aukštaitian dialects (cf. 3.2.4). 3.4.2The dialects of the Paneve˙žys region (paneve˙žiškiu ˛ tarme˙s) 3.4.2.1 Like the neighbouring Šiauliai dialect of West Aukštaitian, this dialect is characterized by a strong tendency towards stress retraction. The farther to the north, the stronger this tendency is. In the southern part of the region, stress is retracted from inflectional endings containing short vowels only (a), in the north it is also retracted from long circumflected vowels onto the stem (b):
3.4.2.2 Word stress plays a much more important part in determining the phonetic shape of a word than in the literary language and in West and South Aukštaitian. While the influence of stress on the quality and quantity of vowels in the Kaunas dialects of West Aukštaitian is relatively slight, it is enormous here, as well as in the neighbouring dialects of the Širvintos region. Not only is the distinction between long and short vowels in unstressed position virtually eliminated, but important qualitative changes occur as well.
63
64
Laimute Balode and Axel Holvoet
Long vowels are shortened in unstressed position: SL
EA (Pan.)
meaning
ãkys pelÙe
ãkis pèla
‘eye’ (nom.pl) ‘mouse’ (nom.sg)
Originally short vowels in inflectional endings are radically reduced (this is usually regarded as the most characteristic feature of this dialect). Only two vowels occur here: these are mid-high, lax, unrounded vowels noted /u/ and /i/ respectively; to the ear, they sound like something intermediary between /u/ and /o/, /i/ and /e/ respectively. /u/ represents any vowel following a non-palatalized consonant, /i/ represents any vowel following a palatalized consonant. This reduction of the original short vowel system must have originated in unstressed endings and subsequently have been generalized: SL
Note that in the transcriptions the apostrophe is used to mark that (1) when the vowel is whispered, its quality cannot be established, and (2) even when it is reduced to zero, the preceding consonant behaves, phonotactically, as if a vowel was present, e.g., a voiced obstruent is not unvoiced before final /s/. 3.4.2.3 The reflexes of sequences of the type /V + N/ (cf. 3.0) vary somewhat: alongside /uN/ the reflex /oN/ » (with a narrow /o/) also occurs.
The Lithuanian language and its dialects
3.4.3The dialects of the Širvintos region (širvintiškiu ˛ tarme˙s) 3.4.3.1 Vowel shortening and reduction in unstressed syllables is comparable to what can be observed in the Paneve˙žys region. The results of the shortening of long vowels and the diphthongs /ie/ and /uo/ differ somewhat from the reflexes noted in the Paneve˙žys region. Instead of the long mid vowels /e./ and /o./ and the diphthongs /ie/ and /uo/, the low vowels /æ/ and /a/ occur. The unstressed vocalism is thus reduced to a simplified system including only the high vowels /i/ and /u/, and the low vowels /æ/ and /a/, and is, thus, identical to the original Lithuanian (and Common Baltic) system of short vowels. The existence of a reduced vowel system in unstressed syllables in combination with a richer vocalism under stress is analogous to East Slavic akan’e, especially to that of the Belarusian type. It should be stressed, however, that the occurrence of /æ/ and /a/ in the place of /e˜/ and /o˜/ is not the result of opening (lowering) in unstressed position. Since /e˜/ and /o˜/ come from Common Baltic /æ˜/ and /a˜/ respectively, the dialects of the Širvintos region have retained the original vowel quality. Apparently quantity in unstressed position became unstable at an early date in this dialect, so that original long and short vowels merged here while becoming differentiated under stress. SL su¯n«elis žmog«elis puod«elis pien«elis
EA (Širv.) [su.« njæ˜ljis] [Šmo.«gjæ˜ljis] [puo«djæ˜ljis] [pjie«njæ˜ljis]
sun«elis žmag«elis pad«elis pen«elis
meaning [su«njæ˜ljis] [Šma«gjæ˜ljis] [pa«djæ˜ljis] [pjæ«njæ˜ljis]
3.4.3.2 Stressed short /i/ and /u/ undergo full lengthening and acquire circumflex tone: SL bùvo mìškas
EA (Širv.) [Ábuvo.] [Ámjiwkas]
bu˜va mı˜škas
meaning [«bu˜va] [«mji˜wkas]
‘be’ (pret.3) ‘wood’ (nom.sg)
3.4.4The dialect of the Anykšcˇiai region (anykšte˙nu ˛ tarme˙s) 3.4.4.1 The most striking peculiarity (shared with the Kupiškis dialect) is the deviant treatment of original short stressed and non-final /a/, whether occurring on its own or as the first component of a diphthong or diphthongoid sequence. In the standard language, original stressed short /a/ is usually lengthened (there is a series of morphologically determined exceptions), and the /a/ occurring as the first component of a diphthong with the acute tone is pronounced as half-long. In the dialect of the Anykšcˇiai region the reflexes are narrowed: an open half-long vowel /f/ appears in both cases. It is described by dialectologists as showing the ‘intermediate
65
66
Laimute Balode and Axel Holvoet
tone’ (·sÒ). From the phonological point of view, the half-long pronunciation is, of course, a variety of shortness, and the ‘intermediate’ tone is actually lack of tonal differentiation. SL n«amas dárbas
[«na˜mas] [´da.rbas]
EA (An.)
meaning
nfs.mas dfs.rbas
‘house’ (nom.sg) ‘work’ (nom.sg)
In some of the dialects of this region, the reflexes of short stressed non-final /æ/ are narrowed as well; they are also half-long, and are ascribed “intermediate” tone. n«eša [«njæ˜wa]
ne» sša
Æ
[~nje» swa]
‘carry’ (pres.3)
3.4.4.2 As in the neighbouring dialects of the Širvintos region, the vowel system is poorer in unstressed vowel systems than in stressed ones (it is reduced to the opposition of high and non-high vowels). The unstressed long mid vowels /e˜/ and /o˜/, as well as the diphthongs /ie/ and /uo/, do not occur in unstressed syllables: instead, the low vowels /æ/ and /a/ appear; they retain their quantity, however (i.e. they are pronounced half-long), whereas they are shortened in the dialects of the Širvintos region. SL žmog«elis te˙v«elis pien«elis
EA (An.) [Šmo.«gjæ˜ljis] [tje.˜vjæ˜ljis] [pjie«njæ˜ljis]
žmageslis teveslis peneslis
meaning [wma.sgje.lji.s] [tjæ.svje.lji.s] [pjæ.snje.lji.s]
3.4.4.3 Stress retraction is less general than in the dialects of the Širvintos region: it shifts from a short ending onto a preceding long vowel or diphthong; the latter is half-long and has ‘intermediary’ tone: SL žmonà su¯nùs
EA (An.) [Šmo.Ána] [su.Ánus]
žmfs.na sus.nus
meaning [sŠmf.na] [ssu.nus]
‘wife’ (nom.sg) ‘son’ (nom.sg)
3.4.4.4 In morphology the main peculiarity is the retention of the IE supine, a form related to the infinitive once used to describe the goal of a movement. It has disappeared from the standard language and most of the dialects. Note that the object of the supine is in the genitive, a feature which is sometimes inherited by the infinitive in those cases where it supplanted the supine, i.e. with verbs of motion.
The Lithuanian language and its dialects
SL
EA (An.)
eı˜sime ši«e no pjáuti go:fut.1pl hay:gen.sg cut:inf ‘We’ll go and cut the hay’
eismà ši«e no pjáutu˛ go:fut.1pl hay:gen.sg cut:sup
3.4.5The dialects of the Kupiškis region (kupiške˙nu ˛ tarme˙s) 3.4.5.1 The distinguishing feature of this dialect is the twofold treatment of /æ/ and /e˜/ according to whether they are followed by a palatalized consonant or not (as a rule, this is not a factor influencing the quality of vowels in Lithuanian). Where the standard language has /æ/ word-finally or before a non-palatalized consonant, the Kupiškis dialect has /a/; and where the standard language has /e˜/, it has /a˜/ under the same conditions. SL bè medùs bérnas katÙe
Note that the consonant preceding the vowel /a(˜)/ is not palatalized, which means that the change occurred before the rise of assimilatory palatalization. The treatment of /æ/ under stress when followed by a palatalized consonant is as in the dialects of the Anykšcˇiai region. 3.4.5.2 Unlike the dialects of the Anykšcˇiai region, those of the Kupiškis region have mid vowels in unstressed position; still, the vocalism of unstressed syllables is somewhat poorer than that of stressed ones, as the diphthongs /ie/ and /uo/ are not distinguished from /e˜/ and /o˜/; instead, half-long /e./ and /o./ appear: SL žmog«elis puod«elis te˙v«elis pien«elis
3.4.5.3 Stress is retracted from a short ending onto a long vowel or diphthong, as in the Anykšcˇiai region.
67
68
Laimute Balode and Axel Holvoet
3.4.5.4 In morphology, the dialects of the Kupiškis region share the retention of the supine with those of the Anykšcˇiai region.
˛ tarme˙s) 3.4.6The dialects of the Utena region (uteniškiu 3.4.6.1 The most characteristic feature of these dialects is that the Common Baltic long open vowels /a˜/ and /æ˜/ have not undergone narrowing into /o˜/ and /e˜/ either in stressed or in unstressed position. Instead, we find a long /a˜/, slightly shifted to the back, or an open /f˜/, with intermediate stages: SL žõdis
[«Šo˜djis]
EA (Ut.)
meaning
[«Ša˜djis]//[«Šf:djis]
‘word’ (nom.sg)
Similarly, the equivalent of the /e˜/ of the standard language is a much broader vowel, ranging from /e˜/ to /æ˜/: te¢˙vas
[´tje˜vas]
[´tje˜vas]// [´tjæ˜vas]
‘father’ (nom.sg)
The retention of original /æ˜/ and /a˜/ is undoubtedly connected with the fact that the treatment of short stressed /æ/ and /a/ is similar to that in the Anykšcˇiai and Kupiškis regions: these vowels have a half-long pronunciation, but are phonologically short. There is thus usually no merger of original /æ/ and /æ˜/, or /a/ and /a˜/. In unstressed syllables, original /æ˜/ and /a˜/ are represented by half-long /æ./ and /a./, without narrowing. The diphthong /uo/ is usually preserved in unstressed position, but /ie/ has coalesced with /æ˜/: SL
EA (Ut.)
meaning
te˙v«elis pien«elis
[tjæ.svjæ.lji.s] [pjæ.snjæ.lji.s]
‘father’ (dim.nom.sg) ‘milk’ (dim.nom.sg)
The half-long /i./ which can be seen in the ending /-i.s/ in the above examples reflects a morphological fact: the Utena dialect uses the ending /-i.s/ for masculine nouns of this declension class, whereas the other dialects have the alternative ending /-is/. 3.4.6.2 As in the dialects of the Vilnius region, before the vowels /æ/, /æ˜/ and /e˜/ and the diphthong /ei/ the consonants /s/ and /r/ are not palatalized. The low vowels /æ/, /æ˜/ are correspondingly replaced by their back counterparts /a/ and /a˜/:
The Lithuanian language and its dialects
SL
EA (Ut.)
meaning
s«enas r«etas
sãnas rãtas
‘old’ (nom.sg) ‘rare’ (nom.sg)
3.4.6.3 The stress retraction characteristic of the East Aukštaitian dialects thus far described is not observed here. 3.4.6.4 The nominative forms of the personal pronoun jìs, jì ‘he, she’, plural ji«e, jõs, are replaced by the forms ãnas, anà, ãnys, ãnos, whereas in the oblique cases the forms derived from the stem of jis are retained. The nominative forms ãnas etc. are actually forms of the demonstrative pronoun meaning ‘yonder’. It is related to the Slavic demonstrative pronoun on, ona, ono, which is also used to supply the nominative forms of the 3rd person personal pronoun. The Lithuanian development therefore probably reflects Slavic influence (though an(a)s occurs also in Žemaitian, cf. 4.1.11). 3.4.7The dialects of the Vilnius region (vilniškiu ˛ tarme˙s) 3.4.7.1 These dialects share with those of the Utena region, among other things, the lack of stress retraction (unlike the remaining East Aukštaitian dialects) and the lack of palatalization of /s/ and /r/ before mid and low front vowels or their reflexes. Here also, we find instances of the unnarrowed pronunciation of Common Baltic and Proto-Lithuanian /æ˜/ and /a˜/, i.e. of the occurrence of /æ˜/ and /a˜/ where the standard language has /e˜/ and /o˜/, but the reflexes of these sounds show quite considerable variation in the Vilnius region. They range from /æ˜/ and /a˜/ to /ie/ and /uo/. The original diphthong /ie/ itself is not merged with the reflex of /æ˜/, but kept intact. Among the peculiarities distinguishing the dialect of the Vilnius area from the remaining East Aukštaitian dialects, we must mention: 3.4.7.2 Like the South Aukštaitian dialects on which they border, the dialects of the Vilnius area are characterized by dzu¯kavimas (cf. supra, 3.2.2). 3.4.7.3 In morphology, a series of particularly archaic endings is found. There are a few archaic consonant stem endings already cited from the South Aukštaitian dialects, where they also occur: SL
EA (Vil.)
meaning
ãkmenys akmeñs
ãkmenes akmenès
‘stone’ (nom.pl) ‘stone’ (gen.sg)
69
70
Laimute Balode and Axel Holvoet
Extremely archaic endings, used for all or nearly all declensions, can be found in the dative and locative plural. In the dative plural, some dialects have the ending -mu, corresponding to Old Lithuanian -mus (the inherited ending), whereas all other dialects have -m(s) (with loss of /u/). In the locative plural, the Vilnius dialects have the inherited endings in -su, whereas all the other dialects have introduced endings in -se after the pattern of the locative singular, which usually ends in -e. SL
EA (Vil.)
meaning
laukuosè vilkáms
laukuosù // laukúosu vilkãmu
‘field’ (loc.pl) ‘wolf ’ (dat.pl)
3.4.7.4 A curious feature in morphology is the use of two different imperatives: alongside the imperative derived with the affix -k(i)-, which is known in all Lithuanian dialects and in the standard language, there is a second imperative which is said to express a command that has to be obeyed immediately. SL
EA (Vil.)
meaning
rašýk
1. rašýk 2. rãšai
‘write’ (imp.2sg)
The imperative in -k(i) is a Lithuanian innovation, unknown even to Latvian (it is thought to contain an exhortative particle comparable to Russian -ka), whereas imperatives of the type rãšai seem to continue an original optative (they are also found in Old Lithuanian, and they are the only ones to be used in Latvian).
4. The Žemaitian (Low Lithuanian) dialects (žemaicˇiu ˛ tarme˙s) 4.1 Common features 4.1.1 Probably the oldest isogloss setting Žemaitian apart from Aukštaitian is that of the so-called ‘Žemaitian sound law’. It is reflected in the following forms, where Aukštaitian has the affricates /tëwj/ and /dëŠj/, whereas Žemaitian has /t/ and /d/: SL
Žemaitian
meaning
jáucˇiai m«edžiams
jáute¯ mèdems
‘ox, bull’ (nom.pl) ‘tree’ (dat.pl)
The mechanism underlying this divergence is the following.
The Lithuanian language and its dialects
At a certain stage in the prehistory of Baltic, /j/ was lost between a consonant and a front vowel. The result of this can be seen in the following alternation: vertì ve«rˇcia
< *vertie < *vert-j-a
<
*vert-j-ie
‘turn’ (pres.2sg) ‘turn’ (pres.3)
In the 2nd person, /t/ remained intact because /j/ was dropped before /ie/. However, as was mentioned above (cf. 1.5.2), in modern Lithuanian ve«rˇcia is pronounced /«vjærjtëwj5/: after the palatalized /tëwj/ only /æ/ can follow. The occurrence of /æ/ in this position is obviously secondary: in Common Baltic, /j/ was still followed by /a/, which, not being a front vowel, did not cause the loss of /j/. In the Proto-Baltic dialect underlying the Žemaitic dialects, the passage of /a/ to /æ/ apparently occurred much earlier than elsewhere, in Common Baltic times. As a result, /j/ was lost here as well, which gave rise to a Proto-Žemaitian form *verte. Before high back vowels the clusters /tj/ and /dj/ remained intact and, in principle, yielded /tëwj/ and /dëŠj/, as in Aukštaitian. However, as the number of forms with affrication was smaller in Žemaitian, there was a strong tendency to eliminate them from inflectional paradigms altogether. In the genitive plural, for instance, some dialects have the regular affricates, while elsewhere /tj/ and /dj/ are substituted: SL
Žemaitian
meaning
jáucˇiu˛
[´jeutëwju]//[´jeutju]
‘ox, bull’ (gen.pl)
4.1.2 The reflexes of the diphthongs /ie/ and /uo/ in Žemaitian are quite different from those occurring in Aukštaitian. These is much variation within the Žemaitian dialects themselves: –
–
South Žemaitian has /i˜/ and /u˜/: SL
South Žem.
meaning
dúona píenas
du¢¯ na pýns
‘bread’ (nom.sg) ‘milk’ (nom.sg)
North Žemaitian has /ei/ » and /ou/: » dôuna » pêins »
–
West Žemaitian has /e˜/ and /o˜/: dôna pêns »
71
72
Laimute Balode and Axel Holvoet
4.1.3 The mid vowels /e˜/ and /o˜/ have passed to /ie/ and /uo/ almost everywhere in Žemaitian: SL
South Žem.
meaning
kója de¢˙ti
kúoja díeti
‘leg, foot’ (nom.sg) ‘put’ (inf)
4.1.4 Everywhere in Žemaitian the short high vowels /i/ and /u/ are replaced by somewhat lower, lax vowels intermediate between /i/ and /e/ and /u/ and /o/ respectively. Lithuanian dialectologists write them as /e/ » and /o/: » SL
Žem.
meaning
bùvo lìkti
bõva » l«e» kte
‘be’ (pret.3) ‘let’ (inf)
4.1.5 The diphthongs /ai/ and /ei/ are often monophthongized into /a˜/ and /e˜/: SL
Žem
meaning
vaı˜kas výrai reı˜kalas
vãks výra¯ rÖekals
‘child’ (nom.sg) ‘man’ (nom.pl) ‘need’ (nom.sg)
4.1.6 The Žemaitian dialects are the only ones to have partly retained the nasal resonance in the reflexes of original sequences of the type /V + N/ occurring before fricatives or word-finally. In Aukštaitian and in the standard language, long oral vowels have arisen here through the intermediary stage of nasal vowels. In some Žemaitian dialects, sequences of the type /V + N/ occur here: SL
Žem
meaning
t«a˛ k¢a˛sti spr¢e˛sti
tã.n. káncti spréncti
‘that’ (acc.sg) ‘bite’ (inf) ‘judge’ (inf)
In these dialects the former nasal vowels thus have the same reflexes as the sequences /V + N/ occurring before obstruents. This treatment is not general, however;
The Lithuanian language and its dialects
where a nasal vowel has undergone early shortening (e.g. in consistently unstressed endings), there is no trace of the nasal resonance. In the Žemaitian forms just cited, there is no narrowing of the vowel occurring before the nasal sonorant. This treatment is restricted to southern Žemaitian, however; most Žemaitian dialects have narrowing of /æ/ and /a/ to /o/ and /o/ » (which are also the reflexes of /i/ and /u/, see above): SL
4.1.7 Stress retraction is frequent in Žemaitian. A characteristic feature of Žemaitian stress retraction is that a secondary stress is left on the original stressed syllable, e.g. SL
Žem.
meaning
šakà pavažà
šàkà pàvažà // pàvàžà
‘branch’ (nom.sg) ‘runner of a sledge’ (nom.sg)
Note that in such polysyllables as pavaža, stress is retracted onto the initial syllable rather than onto the syllable immediately preceding the original place of stress; and that two secondary stresses are sometimes heard in words of this type by Lithuanian dialectologists. 4.1.8 Like Aukštaitian, Žemaitian has a system of two tones; their character is, however, quite different from that observed in Aukštaitian. In most Žemaitian dialects the opposition is not between sharp and drawn tone; under both tones, the first part is more prominent. The acute tone (which is falling in Aukštaitian) is replaced by a kind of ‘broken’ tone (Stoßton): after an initial rise of pitch and intensity, a glottal stop intervenes and the remaining part of the vocalic segment is much lower in pitch and intensity. For the circumflex tone (even or rising in Aukštaitian) this sudden fall in pitch and intensity does not occur, but it is nevertheless the first part of the vocalic segment that is more prominent, and in diphthongs it is clearly lengthened, e.g. lau˜kas is pronounced [«la.uks]. In some dialects, however, both segments seem to be lengthened to the same extent: [«la.u.ks]. Lithuanian dialectologists mark the Žemaitian tones in a different way from the Aukštaitian ones; ·ˆÒ is
73
74
Laimute Balode and Axel Holvoet
used for the acute (Stoßton), the mark of the circumflex ·«Ò is put over the letter denoting the first component: SL
Žem
meaning
de¢˙ti lau˜kas
dîete» lãuks
‘put’ (inf) ‘field’ (nom.sg)
In a general way, the distinction between the tones is easier to perceive in Žemaitian than in Aukštaitian, both in the case of vowels and of diphthongs; this is probably caused by the glottalization associated with the acute. As a consequence, there is no tendency towards loss of tonal distinctions, as there is in the case of long vowels and the dipththongs /ie/ and /uo/ in Aukštaitian. It is interesting to note that the Žemaitian tones are much more reminiscent of the Latvian broken (Stoßton) and drawn (Dehnton) tone than of the Aukštaitian falling and rising tone. 4.1.9 Žemaitian has a strong tendency to drop or to shorten vowels in final syllables. Originally short, unstressed vowels in endings are consistently dropped: SL
Žem
meaning
n«eša výras
nèš výrs
‘carry’ (pres.3) ‘man, husband’ (nom.sg)
Along with a short unstressed /a/, a preceding /j/ may also be lost: ve¢˙jas krau˜jas
víes krãus
‘wind’ (nom.sg) ‘blood’ (nom.sg)
Originally long vowels in endings are shortened: ãkys ž«e me˙
ãkis ž«e me»
‘eye’ (nom.pl) ‘earth’ (nom.sg)
4.1.10 In morphology, Žemaitian has retained a few archaic endings, but it also shows innovative tendencies. Aukštaitian, for instance, has retained the three original declension classes of adjectives; in Žemaitian, the separate declension of u-stem adjectives has been largely assimilated to that of the other declension classes.
The Lithuanian language and its dialects
In verbal inflection, Žemaitian also has some innovations. Verbs which in Aukštaitian have a present tense stem in -i- (replacing the thematic vowel -a) have been assimilated to the thematic class with a present tense stem in -a- (the same process can be observed in Latvian): SL
Žem
meaning
tìkime
tìkam
‘believe’ (pres.1pl)
Žemaitian lacks the special habitual past tense with a stem in -dav- characteristic of Aukštaitian. Instead, a special auxiliary *liuob(e˙)ti is used. Usually the present tense forms of this auxiliary are combined with the infinitive (sometimes also with the future tense) of the main verb: SL
Žem.
meaning
rašýdavau rašýdavai
liúobu rašýti liúobi rašýti
‘I used to write’ ‘you used to write’
It should be added that the special habitual past tense is an innovation of Aukštaitian, unknown also to Latvian, which likewise has recourse to an auxiliary verb (me¯gt) to express habitual meaning. 4.1.11 Like some East Aukštaitian dialects (see 3.3.5.4 above), Žemaitian uses añs and anà (originally a demonstrative pronoun) as personal pronouns for the 3rd person; this stem is, however, used to form the oblique cases as well, not only the nominative as in Aukštaitian. 4.2 Subdivision of the Žemaitian dialects The Žemaitian dialects are divided in three groups on the basis of the divergent treatment of the diphthongs /ie/ and /uo/ (see above, 4.1.2): South, North and West Žemaitian. 4.2.1South Žemaitian (pietu ˛ žemaicˇiu˛ tarme˙s) The diphthongs /ie/ and /uo/ appear as /i˜/ and /u˜/: SL
SŽem
meaning
dúona píenas
du¢¯ na pýns
‘bread’ (nom.sg) ‘milk’ (nom.sg)
75
76
Laimute Balode and Axel Holvoet
South Žemaitian is further subdivided into: 4.2.1.1 The dialects of the Raseiniai region (raseiniškiu˛ tarme˙s). These dialects border on Aukštaitian and have, in some respects, a transitional character. Sequences of the type /æ+N/ and /a+N/ are preserved as in standard Lithuanian (without narrowing of the vowel). 4.2.1.2 The dialects of the Varniai region (varniškiu˛ tarme˙s). These dialects have a greater number of Žemaitian features than those of the Raseiniai region. They have narrowing of the vowel in the sequences /æ + N/ and /a + N/: SL
SŽem (Var.)
meaning
lángas lénkas
lónks » lénks »
‘window’ (nom.sg) ‘Pole’ (nom.sg)
One also finds more or less regular monophthongization of the diphthongs /ai/ and /ei/ here: SL
SŽem (Var.)
meaning
dáiktas re˜Fkalas
dâ.kc r«e.kâls
‘thing’ (nom.sg) ‘need’ (nom.sg)
4.2.2North Žemaitian (šiaure˙s žemaicˇiu ˛ tarme˙s) The diphthongs /ie/ and /uo/ appear as /ei/ » and /ou/: » SL
NŽem
meaning
dúona píenas
dôuna » pêins »
‘bread’ (nom.sg) ‘milk’ (nom.sg)
North Žemaitian is furthermore characterized by the co-occurrence of the high vowels: /i/ and /u/, from original long /i./ and/u˜/ in unstressed syllables, and the mid-high vowels /e/ » and /o/, » from original short /i/ and /u/. In the standard language and most dialects, the result of the shortening of long /i˜/ and /u˜/ is not distinguished from original short /i/ and /u/: SL
NŽem
meaning
akìs (< *akis) akìs (< */aki˜s/ < *akins)
àkès » àkìs
‘eye’ (nom.sg) ‘eye’ (acc.pl)
The Lithuanian language and its dialects
The tendency towards loss of final vowels, characteristic of Žemaitic as a whole, is particularly strong in the northern dialects of North Žemaitian (near the Latvian border). Original short vowels are lost completely, while short vowels resulting from the shortening of original long vowels are pronounced voicelessly (whispered), or they may be lost completely: SL vaı˜ka˛
NŽem «
[ vai.ka.]
vã.k
a
meaning [«va.ka] ®
‘child’ (acc.sg)
North Žemaitian is further subdivided in an eastern subgroup (the Telšiai region) and a western group (the Kretinga region). 4.2.2.1 The dialects of the Telšiai region (telšiškiu˛ tarme˙s). 1. The broadening of original short /i/ and /u/ is not carried through everywhere, but is subject to certain positional restrictions, e.g. SL
Here the opening of /i/ is conditional upon the quality of the vowel in the ending. 2. Long vowels in pretonic syllables are shortened: SL
NŽem (Tel.)
meaning
lu¯pele˙
lupel¥
‘lip’ (dim.nom.sg)
3. Where the standard language has long oral /æ˜/ and /a˜/ developed from nasal vowels, the dialects of the Telšiai region always have narrowing; in some parts, the nasal resonance has been preserved and appears as /n/: SL
NŽem (Tel.)
meaning
k´a˛sti spr´˛esti
kôst » ¥ // kônst¥ sprêst » ¥ // sprênst » ¥
‘bite’ (inf) ‘judge’ (inf)
77
78
Laimute Balode and Axel Holvoet
4. The Žemaitic monophthongization of /ai/ and /ei/ occurs mainly word-finally: SL
NŽem (Tel.)
meaning
svecˇiai m«edžiai
svèt«e mèdê
‘guest’ (nom.pl) ‘tree’ (nom.pl)
4.2.2.2 The dialects of the Kretinga region (kretingiškiu˛ tarme˙s). 1. The broadening of original short /i/ and /u/ is carried though consistently. 2. The nasal resonance in former nasal vowels continuing /æN/ and /aN/ has disappeared; instead, a long oral mid vowel or a diphthong has arisen: SL
3. /t/ and /d/ have been generalized instead of the affricates /tëwj/ and /dëŠj/ in those inflectional paradigms where the Žemaitian sound law operates (see above, 4.1.1) 4. The dual has been preserved, as in some dialects of the Kaunas region, see above, 3.1.1.6. 4.2.3West Žemaitian (vakaru ˛ žemaicˇiu˛ tarme˙s) This is the dialect of the Klaipe˙da region, German Memelland. This region once belonged to East Prussia and was always exposed to a strong German influence. Most of the autochthonous Lithuanian-speaking population was evacuated to Germany in 1944. The dialect is, thus, virtually extinct. The diphthongs /ie/ and /uo/ are represented by /e˜/ and /o˜/: SL
WŽem
meaning
dúona píenas
dô.na pêns »
‘bread’ (nom.sg) ‘milk’ (nom.sg)
Such typical Žemaitian features as the broadening of original short /i/ and /u/ or the monophthongization of /ai/ and /ei/ are encountered here as well. Final vowels are subject to strong reduction: complete loss of original short vowels, voiceless pronunciation of shortened, originally long vowels. Assimilatory palatalization of consonants before front vowels is hardly perceptible, except in the case of /k/, /g/ and /l/. Perhaps this is due to German influence.
The Lithuanian language and its dialects
Note * The authors wish to thank Bonifacas Stundžia (Vilnius), who kindly commented on the draft version of this survey, but bears no responsibility for any shortcomings that might have escaped the authors’ notice.
References Gerullis, Georg. 1930. Litauische Dialektstudien. Leipzig. Lietuviu˛ kalbos atlasas I-IV. Vilnius 1977– Lietuviu˛ kalbos tarme˙s. Vilnius 1970. Zinkevicˇius, Zigmas. 1966. Lietuviu˛ dialektologija. Vilnius: Mokslas. Zinkevicˇius, Zigmas. 1978. Lietuviu˛ kalbos dialektologija. Vilnius. Zinkevicˇius, Zigmas. 1984–1994. Lietuviu˛ kalbos istorija. I–VI. Vilnius. Zinkevicˇius, Zigmas. 1968. Lietuviu˛ kalbos tarme˙s. Kaunas.
79
AUTHOR "Valeriy Cˇekmonas"
TITLE "Russian varieties in the southeastern Baltic area"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Russian varieties in the southeastern Baltic area Urban Russian of the 19th century Valeriy Cˇekmonas
1.
Introduction
Two variants of the Russian language with different social statuses coexisted in the southeastern Baltic area throughout the 19th century. The first can be called urban Russian and the second rustic Russian dialects. Urban Russian was the language of educated people in Vilnius (Pol Wilno, Rus Vil’na, Vil’no), Riga and to some extent in Tallinn (Ger Reval, Rus Revel),1 that is, in the capitals of what were to become the Baltic States. Each of these cultural centres had their own sociolinguistic situation depending on the peculiarities of their political and ethnic history. The introduction and spread of the Russian language in Lithuanian, Latvian and Estonian territories during the 19th century has never been thoroughly studied, least of all from a sociolinguistic point of view. The problem was touched upon by Semenova in her sketch of language contacts in 19th century Riga (1977: 192–215); her observations are refined and somewhat extended in this paper and new data about the history of the Russian language in Vilnius, Riga and Tallinn are considered.
2.
Russian in Vilnius
The East Slavic Orthodox population is known under the name of “Russians” in old Vil’na from the thirteenth century. Part of the town of Vil’na was known as the Russian quarter (Ruskij konec) up to World War I. The clergy of the oldest Vil’na churches and merchants families from Smolensk, Polock, Pskov, Novgorod and other East Slavic Lands of the Grand Duchy of Lithuania (GDL) were Orthodox. They called their everyday language rus’ka mova ‘Russian language’; it was a spoken variety of the language usually called Old Belarusian (Žuravskij 1984: 17; Mecˇkovskaja 1989). Polish supplanted the written form of this language in official life
82
Valeriy Cˇekmonas
during the 17th century. Polish became the main language of social life and contacts in Vilnius during the 18th century, while the rural population around the town continued to speak the local Lithuanian vernaculars.2 The history of the Lithuanian urban Russian (or Great Russian) language proper began in 1795 when Vilnius (Pol Wilno as it was known at the time) became the centre of the newly created Vil’na province (gubernija) after the third partition of the Polish Commonwealth. Russian was the language of the Tsarist administration and official institutions only until the first Polish insurrection in 1831. After the insurrection was suppressed, resolute measures were undertaken to Russify Polish-Lithuanian society and the new part of the Russian Empire, officially called the Northwest Region (Severo-Zapadnyj kraj), as a whole. The ideological grounds for the Russification of the former GDL’s population and lands were given as being a restoration and support of Orthodoxy in the “age-old Russian territories” (iskonno russkie zemli) which through the centuries had been suppressed in the Grand Duchy of Lithuania by Polish Catholicism. The Imperial Vil’na University (opened in 1801), which was the main centre of Polishness, was closed down; a Russian gymnasium was established in Vil’na; the Polish press in Vil’na was suppressed etc. The anti-Polish policy reached its peak after the second Polish-Lithuanian uprising in 1861–1863 when the use of the Latin alphabet, teaching in Polish and using Polish in “public places” were prohibited (the most recent analysis of the Russification policy based on historical sources is presented in Aleksandravicˇius & Kulakauskas 1996: 82–106). The Cyrillic alphabet was required instead of the traditional Lithuanian one for publications in the Lithuanian language all over the Empire from 1865 (Murav’ev 1887: 313–314). The population of Vilnius grew continuously during the 19th century from 17,500 in 1796 to 56,300 in 1811 and 154,500 at the beginning of the 20th century (Jurginis et al. 1968: 214–215). According to the All-Russia census in 1897, 2.1% of the permanent residents of Vilnius were Lithuanians, 20.1% Russians, 30.9% Polish, 40.0% Jews and 4.4% Belarusians (of a total of 154,500 residents). The population of Vilnius rose especially steeply in the last quarter of the 19th century. There were 86,668 people living in the town in 1875, 37,909 Jews among them (Vilenskaja gubernija, 1901: 1–3). Except for the clergy, the traditional Orthodox Vilnius residents did not call themselves “Russians” in the 19th century. They were simply “Orthodox”. It is not easy to establish how large the permanent Vilnius Orthodox population was in the group which consisted of Vilnius traditional Orthodox and Russians newcomers in the first half of the 19th century. A Polish historian and Vilnius dweller, M. Balin´ski, pointed out that only 70 of the 1912 children born in 1932, that is about 3.7%, were christened in the Orthodox churches. We can conclude with caution that this reflects the size of the Orthodox population of Vilnius, that is since Balin´ski (1935: 61, 64, 67) estimated the whole population of Vilnius as 36,000–50,000
Russian varieties: Urban Russian
people, there must have been 1,332–1,850 Orthodox. According to the data collected by the officers of the Russian General Staff, there were about 60,000 people in Vilnius in 1858, 7,385 of them (12.3%) were Orthodox (along with 110,926 Catholics and 29,767 Jews) (Koreva 1861: 314, 315, 317). For the first time we can learn from these data that Russians proper (Great Russian, Veliko-Rosijane, Russian newcomers) composed only 2.3% of the population of Vilnius. Thus, the proportion of Orthodox and Orthodox Great Russian in Vilnius was very small, they were 3% or so at the beginning of the century and 20% by the end. The growth of their number was due to some migration from inner Russia, stimulated by the needs of the bureaucratic apparatus, the police and the educational system. A theatre was set up in Vilnius at the beginning of the 19th century. Plays in Russian and Polish were staged there until the 1840s when it became a Russian-only theatre. Vil’na’s Russian press was represented by the official Litovskij Vestnik (1835–1915), Vilenskie Gubernskie Vedomosti (1838–1915) and Vilenskij Listok (1841–1915). Litovskij Vestnik is the most interesting of them. There was also a Polish newspaper Kuryer Wilen´ski in Vilnius from 1797. From 1835 its first title was Litovskij Vestnik. Oficial’naja gazeta and Kuryer Wilen´ski was the second one. The newspaper had not only two titles, it was actually bilingual Russian-Polish, with the Russian text published in the first column and the Polish in the second. It had ceased to be absolutely parallel by the beginning of the 1860s when a well-known Russian-Polish scholar, A. K. Kirkor, became its editor. Russian texts dominated in the paper at that time; only official news and advertisements were translated into Polish and all non-official materials were published in Russian. After the use of the Polish language in social life was prohibited in 1864, Vilenskij Vestnik. Gazeta oficial’naja, politicˇeskaja i literaturnaja became an official Russian monolingual paper. Thus, the history of this publication reflects the process and mechanisms of the introduction of the Russian language in Vilnius. A great number of short-lived Russian-language publications came into existence in Vilnius after 1863. The Vestnik Zapadnoj Rossii (1864–1871) and Sel’skie ˇctenija (1877–1878) may be the better known of them. The number of Russian publications grew until 1915, which reflected the strengthened position of the Russian language in Vil’na and Vil’na province as a result of the Russification policy. Vil’na really became the most western centre of the Russian culture, Orthodoxy and language by the end of the 19th century. As an outpost of Orthodoxy in this region, it was especially important to the Tsarist policy and closer to the Russian social, cultural and political spheres of interests than the economically more important Riga. There were a great many educational institutions (among them several state and private gymnasia, professional schools, pedagogical high school, two seminaries (one each for men and women)) in Vil’na. The town itself was, of course, the centre of a Military District and provincial administration.
83
84
Valeriy Cˇekmonas
Almost nothing is known about the spoken Vilnius urban Russian of the 19th century. The editors, authors and correctors of the Vilnius Russian periodicals were usually educated at central Russian universities and other high schools, and they were doing their best to ensure that the language of their publications was the same as in those of St. Petersburg and Moscow. Slight deviations from Standard Russian can be observed in the texts of Vilnius newspapers, but they are very rare. “The remarkable verdict: a true story” (Zamecˇatel’nyj prigovor. Byl’. Vilenskij Vestnik 1. 1863), a sketch by A. Eisimont (the very name shows the author was from the Severno-Zapadnyj kraj) contains such expressions as (centr plošcˇadi zanjat množestvom ‘the centre of the square was occupied by the great number of the’) lošadinyx promyšlennikov ‘horse-dealers:gen.pl’ (instead of konskix baryšnikov), tratit’ tysjacˇi na izbytki ‘to spend thousands on over-indulgenses’ (two slips at once: Rus izbytok ‘over-indulgens, excess’ is not used in the plural, and it is impossible to waste money izbytok ‘cˇto …ostaetsja v izliške; lišok, izlišek … ‘what is left as remainder, remainder’ in Russian). Some special words for local Lithuanian-Belarusian realia (e.g. pan, vocative pane, lajdak ‘sir, rouge’) occurred in a story Iskateli kladov (mestnyj ocˇerk) “The searchers of the hidden treasures (a local piece)” by T-v (Vilenskij Vestnik 3, 1863). It is worth mentioning that an expression (written in the Cyrillic alphabet) koza bjala, aby passak mjala ‘a she-goat is white, but it should have a dowry’ contains one word borrowed from Belarusian or Lithuanian (passak < Belr paság > dial. Lith p«asogà ‘dowry’) that is noted here for the first time, and one loan from Belarusian (aby, cf. Pol. ˙zeby ‘in order’). These traces of Polish or Belarusian influence on the Vilnius Russian publications are generally extremely rare. Some features of the Vilnius sociolinguistic situation were reflected in fictional works of Russian and Jewish authors. Baron Vasiliy Aleksejevich fon Rotkirx (1819–1891, pen-name ‘Teobal’d’) was one of the best known Russian writers in Vil’na in the second half of the 19th century. He wrote about everyday life in 19th century Vil’na (Vospominanija Teobal’da: 1890). In a novel Jacek Kryštalevicˇ, Vil’na’s jurodivyj’ ‘God’s fool’, for example (first published in Vilenskij Vestnik, 1888, N 259), he told the story of a legendary Vil’na ‘God’s fool’ and rhymester of the 1830s and 40s. The educated characters in the novel speak Polish while the speech of the countrymen and ordinary town folk is represented (in this and other novels) by a stylised Belarusian dialect, cf.: Sluxajt’e, xlopcy. ˇci vy bacˇili koli …malpu? …Tak heto malpa i jos’c’? … A lixo jaje golove, jakaja jana paskudnaja! …Kab jaje vovk zarezav! ‘Listen, chaps. Have you ever seen… an ape? …Is this an ape? … (liter.) I’ll luck for its head, how ugly it is! …May a wolf it devour!’ etc. (1996: 154). Teobal’d’s texts attest that Russian was not the everyday spoken language of the common people in Vil’na. L. O. Levanda (1835–1888) wrote the Jewish Russian feature-story (Goldberg 1900: 4). He published his sketches and novels in the first Russian-Jewish publications (Rassvet, Sion, Russkij Jevrej); the best of them were reprinted in central
Russian varieties: Urban Russian
Russian publications (Levanda: 1875). The characters of his works commonly speak “Jewish-German jargon”. This becomes clear when they are going to forward a petition to a Russian administrator: it turns out they needed to translate this petition into Russian. Everyone of them, the author remarks, “was internally, to himself aware it would not be easy to do, especially if you could not lay your hands on the Russian-German and German-Russian Schmid’s dictionary, this handbook of every Jew who begins to teach himself Russian” (Levanda 1875: 58). But it cannot be determined from the style of their direct speech that they are using another language than Russian. Some Jewish words can be found in Levanda’s prose (for instance: mexuten ‘bridegroom, fiancé’, xupa ‘bride’s croun’, gviry (Ydish gvirim) ‘reach people’, mašures ‘servant’ etc.), but their meanings are always explained and the language could be that of any of the educated Russian authors of that time. The same can be said about Levanda’s novel “The hot times” (1875) about Polish problems and life where the well-spoken Russian-speaking characters use some Polish words (usually spelled in Polish). Thus, Vilnius Urban Russian was the language of imperial administration, the press, literary works and the theatre. The very phenomenon of L. O. Levanda and other Jewish authors who wrote in Russian reflected the strengthened social position of Russian during the 19th century as one of the main results of Russification. There is no evidence to suppose that any social sections in 19th century Vilnius used the Lithuanian language and claim, as Semenova (1977: 213) does, that PolishRussian-Lithuanian trilingualism existed in the town. Moreover, it is doubtful whether there was any stable tradition of spoken Russian among the uneducated people here. The peculiarities of the local Polish, Belarusian and Jewish languages are not, as far as we know today, reflected in the Russian language of Vilnius publications. We cannot agree with Semenova (ibid.) that the Lithuanian language influenced Russian in 19th century Vilnius (and Lithuania generally). The official Russian language was obviously a source of lexical borrowings into the local Polish and Lithuanian dialects. By the end of the 18th century (Kurzowa 1993: 480) Polish authors from Lithuania (J. Chod’ko, A. Mickiewicz, F. Mickiewicz, the so-called Filomats) had used Russian loans such as bezporza˛dki (
85
86
Valeriy Cˇekmonas
wojenny (< voennyj ‘military’), leczebny (< lecˇebnyj ‘medicinal’) etc. These Russims were rather barbarisms used for the new (Russian) social and political realities than proper loan-words, and no one of them get introduced into standard Polish. Kurzowa (1993: 480) writes that there is another group of borrowed Russian words that came into use during the second, more intensive upsurge of Russification. The process of Russification was actually a growing tendency throughout the 19th century. It became most intense after 1863 when the January insurrection was suppressed. This second wave of Russification brought new words and phrases, often mixed language expressions and set phrases. In spite of their foreignness they were deeply rooted in everyday speech and were commonly used until the interwar period of the 20th century; most of these words refer to Russian institutions and social relations: kładowa (< kladovaja ‘larder, store-room’), konvert (< konvert ‘envelope’), objawiene (< objavlenie ‘advertisement’), odczot (< otcˇet ‘account’), odsrocˇka (< otsrocˇka ‘postponement’), osmotr (< osmotr ‘inspection’), ostanowka (< ostanovka ‘stop, station’), oprzedzelc´ (< opredelit’ ‘appoint’), podpiska (< podpiska ‘written undertaking’), powiestka (< povestka ‘summons, call-up papers’), rozpiska (< raspiska ‘receipt’), stracˇovac´ (< straxovat’ ‘insure’), ziemielny (< zemel’nyj ‘land (adj.)’) and others. Many of the Russian loans listed above are found in the Lithuanian standard language or dialectal speech, cf.: (current) bìrža ‘stock exchange’, peñsija ‘pension’, prokuròras ‘public prosecutor’ (dialectal and/or often obsolete) ˇcìnas ‘rank’, ˇcinau˜nas ‘official, of high rank’, kaziònas ‘formal, official’, traktie˜rius (traktie˜ris, traktie˜ra) ‘tavern’, rašk«ažnas (raškažnus) ‘luxurious’, konvér˜tas ‘envelope’, strõšnas ‘terrible’, saldõtas ‘soldier’. Zinkevicˇius (1990: 91–92), gives more Russian loanwords of the same kind, for example, mirãvas (< mirovoj sudja ‘Justice of the Peace), navabráncas (< novobranec ‘recruit’), pìsorius (< pisar’ ‘clerc’), rótnas (< rotnyj komandir ‘company commander’), slie˜davatelius (< sledovatel’ ‘investigator’), spráuninkas (< ispravnik ‘district police officer’), staršinà (< staršina ‘sergeant-major’), uje˜zdas (< ujezd ‘district’), võlastis (< volost’ ‘small rural district in Russia’), zemstvà (< zemstvo ‘elective district council in Russia’), zem ˜ skis (< zemskij ‘elective district council in Russia:adj’), ucˇ`ilišcˇa (< ucˇilišcˇe ‘school’), ucˇitelis (< ucˇitel’ ‘teacher’), grìpelis (< grifel ‘slate-pencil’), rucˇkà (< rucˇka ‘pen’), tetradka (< tetradka ‘copy-book’). The above listed words are historicisms or found in the Lithuanian literature of the time and can easily be identified as the 19th century’s borrowings. But it is not easy or sometimes almost impossible to distinguish between older and more recent post W.W. II Russicisms in other groups of lexemes. The problem is very delicate and invites further investigation. The Russian language did not, however, influence Polish or Lithuanian phonetics, grammar, or syntax. Generally, the role of Russian as a state language and as the language of the educated people in Lithuania and the Russification itself slightly effected Lithuania’s Polish and local Lithuanian dialects. Neither of the latter influenced the official Russian used in Vilnius as a centre of the Russian
Russian varieties: Urban Russian
administration and culture in this region. Bilinguism in Polish or Lithuanian and Russian or trilinguism in Polish, Lithuanian and Russian were not characteristic of certain social strata, but rather of local educated individuals. It is natural to suppose that the number of bilingual and trilingual people was growing rapidly after the Russification of educational establishments, but, again, this did not produce any evident effect on the Standard Russian used in Vilnius. Some remarks should be made about the impact the Russification policy has had on the Lithuanian rural population. The ground was laid for this policy by the Tsar’s manifesto from 1862, which proclaimed the necessity of establishing public primary schools (narodnyje školy) everywhere in the Empire (Polnoe sobranie zakonov…1865: XXXVII n 37873; Kulakauskas 1986). More than 100 such schools were opened in Vil’na province in 1862–1883 and 150 in 1864–1865 in Kovno provice (Tyla 1990: 50–51). The primary schools were to replace other forms of teaching which were not under official control. The private and church (monastery) schools were closed and prohibited at the same time in Vil’na and Kovno provinces. All subjects, except for religion, were to be taught in Russian in public primary schools (the General Governor, K. Kaufman, reduced the period of time for the teaching of religion “in local vernaculars” to one year in 1866 (Aleksandravicˇius & Kulakauskas 1996: 84)). A remarkable document was recently published by the Lithuanian historian K. Misius (1994). It contains an analysis performed by N. Segievski, the Curator of the Vil’na educational district (okrug) in 18843 of the efficiency and general results of Russification after the use of the Latin alphabet was banned in 1865. Segievskij acknowledges that “only a few of the Lithuanians and Zhmudins knew Russian” before 1865 and “Russian speech was not heard in all parts of Russian Lithuania and Zhmudz” (Misius 1994: 411). He maintains that a 20-year period is too short to Russify the whole of society, but that the general results of the measures undertaken are not so bad. “Students at gymnasia, non-classical secondary schools and pedagogical seminaries leave these schools with sufficient knowledge of Russian” Segievskij ascertains. The most interesting for us is the following information: The Director of the public schools of this province testified that a total of 3000 Lithuanian and Zhmudian boys in Kovno province were taught to speak and read, and where possible, to write in Russian fluently. According to the Director of the Shavli (current Lith Šiauliai — V. Cˇ.) gymnasium, in which students from country estates comprise a high percentage, the Zhmudian boys know enough Russian after primary schools to go on to the preparatory, first and even second classes. According to the report of the Inspector in charge in the Ponevezh (now Lith Paneve˙žys — V. Cˇ.) non-classical gymnasia, most of the students who show a good knowledge of Russian in the entrance examination were the peasant children who had been through the primary schools (Misius 1994: 415).
87
88
Valeriy Cˇekmonas
Of course, Sergievskij should be optimistic in the performance of his official duty but he was right in his general prognosis: “The more of the population who complete a course in Russian, the more the Russian language will be spread among the common people” (ibid.). The efficiency of the teaching of Russian in “Nickolaj’s schools” (named after the last Tsar, Nikolaj the Second) means that older men unlike their children, understood and often spoke Russian quite well.4 Knowledge of Russian spread among the Lithuanian gentry (Lith bajorai) as well, especially among those of them who were acknowledged as Russian nobles (russkie dvorjane) and became official employees. As Zinkevicˇius (1996: 286) writes pathetically, Everything was Russified, but the estates and churches were still Polish strongholds which treated Lithuanians as “Poles who spoke a different language”. The use of Lithuanian declined catastrophically. It seemed as if there would soon be nothing left except for the names of cities, farmsteads, rivers and lakes, and those too were Slavicized. In 1886 the Russian cabinet of ministers lost no time in proclaiming that Lithuanian was extinct and remained only as an object for scholarly research. They questioned whether the Lithuanian nation and language would survive or not.
The natural historical Polonisation of Lithuanian society, which started after the Lithuanians became Christian in 1386–1387, gave way to forced Russification in the last quarter of the 19th century. The number of bilingual Lithuanians with Russian as a second language had increased immensely by the end of the century, but the Russian language was always viewed by the Lithuanians as foreign (unlike Polish which was their own, being the language of the native gentry and the traditionally “native” Catholic church). Russian never spread spontaneously among the rural population and no local forms of Russian arose anywhere in ethnic Lithuania (again, unlike Polish which was adopted as “native” and started to spread among the rural population, giving rise to local varieties in southeastern Lithuania during the period from the end of the 19th century to World War II). Nothing is known about the properties of the spoken Russian of LithuanianRussian bilingual people. There is, however, a unique source of this and the written Russian of the semi-literate peasants as well. Petitions to the Ministers of Education (and sometimes even to the Tsar himself) were written and forwarded by the Lithuanian villagers asking them to abolish the ban on the Lithuanian press.5 Most of them were analysed by Merkys (1978: 136–197), but some were published originally by Kulakauskas (1993). The petitions (especially from southern Lithuania) are authentic documents witnessing that even the Lithuanian text published in the Cyrillic alphabet was perceived as foreign and incomprehensible, unsuitable for teaching religion (because the petitions are mostly about the lack of religious books). We can only imagine how the authors and the writers of the petitions did their best to produce a good Russian text and their results are interesting.6
Russian varieties: Urban Russian
First of all, the texts do not observe the standard orthography of the time (that is an objective indication of their authenticity); for instance poimenovanyx (= poimenovannyx ‘mention’), dotakoj stepeni (= do takoj ‘(up) to’), apotomu (= a potomu ‘because of that’), gramotni (= gramotnyje ‘literate’), ažydat’ (= ožidat’ ‘wait’), abjavit (= ob”javit’ ‘announce’), domoxozjajevy (= domoxoz’ajeva ‘farmers’), po nevole (= ponevole ‘against one’s will’), Božix (= Bož’jix ‘God’s’), katorogo (= kotorogo ‘what:gen.sg’), zanegramotnyx (= za negramotnyx ‘for illiterate’), Sijatelstvo (= Sijatel’stvo ‘(your etc.) Excellency’) (here and further on examples are from Kazlauskas 1993). The authors did not distinguish between Rus zaprešcˇat’ ‘forbid, prohibit, ban’ and vosprešcˇat’ ‘ibid.’, and zaprešcˇenie or zapret pecˇati ‘prohibition, ban’; they called vosprešcˇenie, vospret (ibid.: 484–485); prinuždeny is used instead of vynuždeny ‘forced’ etc. Sometimes, Polish forms are encountered: po litovsku, politovsku ‘in Lithuanian’ (cf. Pol. po litewsku ‘ibid.’). Some syntactic constructions show poor command of Russian as well: po slucˇaju zapreta litovskoj pecˇati (should be ‘po pricˇine/iz-za zapreta’ ‘because of the prohibition of the Lithuanian press’); my lišeny vsjakix svedenij po tom ‘we are lacking information about’(should be: o tom). But calques from Lithuanian make these texts sound strange, for instance: u nas do takoj stepeni proredeli Katexizisy ‘the Cathechisms (in Lithuanian — V. Cˇ.) became so rare’ (Rus proredeli ‘became rare’ is not correct being a calque of Lith išrete˙´ti 3. ‘(of something) to become more rare because of disappearing ‘ (LKŽ XI); ibo išcˇezli meždu nami litovskie knižki ‘for Lithuanian books are disappeared (literally) among us’ (Rus meždu nami in this context is a calque of Lith tarp mu¯su˛ ‘among us’); pritom my bedny ješcˇ’o i s toj storony sounds as literal translation from Lith vargingi esame iš tos puse˙s ‘we are poor, we suffer because of ’ (literally ‘from that side’ ); deti naši ne imejut sredstv dlja izucˇenija molitvy Bož’jej‘our children have not any means to learn God’s prayer’ (Rus sredstv in this context is a literal translation from Lith neturi priemone˙s ‘they have not any means’); soveršenno izcˇerpnulis’ litovskie knigi ‘the Lithuanian books are disappeared’ (Rus izcˇerpnulis’ for ‘disappeared’ is a translation of Lith išsise˙me˙ ‘(they are) exhausted’); byvalo prežde pecˇatalis’ ‘were published in former times’ is a literal rendering from Lith budavo ankšcˇiau spausdinami); eto obstojatel’stvo neblagopolucˇno otozvalos’ nanašem molodom pokolenii ‘these circumstances were unfavourable for our children’ (Rus otozvalos’ ‘is told upon’ is used for otrazilos’ ‘affected’ as translation of Lith atsilepe˙ ‘had an effect on’); my bedstvujem v umstvennom i duxovnom otnošenii ‘we are suffering intellectually and spiritually’ (Rus bedstvujem in this context is a literal translation from Lith mes vargome/skurstome ‘we suffer’ and ‘we are poor’); the whole sentence my…po nevole prinuždeny pribegat’ k pokupke pol’skix knig no iz nix nam ne beret oxota molit’sja tak kak ešcˇe ne ocˇen’ xorošo ponimaem po pol’ski ‘we have to buy Polish books against our will but we have no wish to pray using them because we do not yet understand Polish well’ was a literal translation of something in Lithuanian like mes … esame
89
90
Valeriy Cˇekmonas
priversti pirkti lenkškas knygas, bet iš ju˛ mums ne˙ra noro melstis, kadangi dar nelebai suprantame lenkiškai ‘idem’. The forced introduction of the Russian language through the primary schools into the lives of the Lithuanian peasants was one of the more effective methods of the Russification process. Knowledge of Russian in the primary-schools became a powerful channel for loanwords and the adaptation of Russian words for the new social and cultural notions. It meant the Russian language had begun to take over the role that the Polish language had previously played. The difference was that knowledge of Russian did not affect the national sentiments of Lithuanians. Functionally, they associated the Russian language with official use; they never used it within the family or between themselves. Generally, it was a learned second language with a narrow (official) sphere of use. The above data seem to show this quite clearly. The policy of Russification, we should not forget, was aimed against the Polish nation, and the Catholic religion, culture and language first of all and Lithuanian society was Russified as a part of Polish society. The Russian-Polish conflict favoured the rise of the Lithuanian national movement and enhanced the spread of the Polish language among the lower ranks of the population, but that is beyond the scope of our present topic.
3.
Russian in Riga
The history of Russian in Riga is superficially similar to what happened in Vilnius as Semenova (1977) shows: there was a separate Russian quarter (Russkij konec) in Riga in the 13th century with Russian streets, a church and storehouses…. The Russian quarter was burned out in 1677 and a new Russian commercial centre called the Moscow suburb (Vorstadt) had arisen as late as the 18th century when Russians were allowed to settle in Riga… (Semenova 1977: 197)
After the bulk of Livonia, with Riga, was ceded to Sweden in 1629 “the Swedish authorities sought to eradicate all that was Russian and there were almost no Russians during this time” (ibid.). According to the Treaty of Nystad of 1721, Livonia became a part of the Russian Empire and Riga became the centre of the province of Liflandia (later Lifljandskaja gubernija). Russian officials, soldiers, craftsmen and workers arrived to Riga in the 18th century. The Russian language, however, was used only in private life. A few of Riga’s better-educated Russians used German to communicate with the local officials and businessmen. German was so important that not only Russian but also German terms were used in state documents… (Semenova 1977: 198).
Russian varieties: Urban Russian
Thus, although both Vilnius and Riga have had an East Slavic population since the 13th century, the history of the Russian language proper (Great Russian) in these towns began after they were incorporated into the Russian Empire. The sociolinguistic situation in Riga, however, was quite different from that in Vilnius. In Riga there was a proper Russian population, because “Russians who came when the town was subdued by Peter the Great stayed and multiplied” (Gun 1803 I: 70). Russians comprised 10.3% of Riga’s population by the end of the 18th century (3,092 of around 30,000 inhabitants, Gun 1803 I:132). But in fact there could have been many more of them. Gun (ibid.) says the Poles “are great in number” in Riga estimating them to be 2,186 (that is, 7.3% of all inhabitants). To understand who these Poles might be we must take two things into account. Firstly, according to Semenova (1977: 198), Old Believers began to settle in Riga at the end of the 18th century.7 Secondly, Želtov (1876: 2) maintains that the Russian population of Riga is formed from two elements, namely those who came from the central Russian provinces and those from the Baltic Region’s neighbouring provinces e.g. Vitebsk, Kovno, Vil’na. According to Želtov, the latter are mostly Old Believers who have been resettled from the Russian State to the Northwest Region in former times. It is of great importance that the Russians from the central provinces are simply called “Russians” by the people of Riga while those from the Northwest Region are called “Polish” (pol’skie). The language of the “Polish” Russians has been slightly changed under the influence of Belarusian and Polish-Lithuanian dialects but it is, nevertheless, essentially Great Russian. (Želtov 1874: 3)
Bear in mind that there is nothing to indicate a mass resettlement of the Poles to Riga in the 18th century and what Gun calls Riga’s “Polish population” at the end of the 18th century can be considered to be Russians (mainly Old Believers) as well. This would mean that a maximum of 5,278 (3,092 + 2,186) or 17.6% of the people in Riga at that time were Russian speaking. According to Gun (1803: 124–126), there was only one public High School for “Russian youth” in Riga (Glavnoe narodnoe ucˇilišcˇe), two private (primary?) Latvian schools, two German high schools (Glavnoe gorodskoe ucˇilišcˇe pri sobornoj cerkvi and Imperatorskij Licej), two private (vol’nyje) schools and several (it is not said how many of them) primary schools. “Although different nationalities live in the town and its suburbs, each of them keeps its native tongue. Everybody here understands the three predominating languages, Russian, German and Latvian” (Gun 1803:71) observes. According to the data of the First All-Russia census of the population, the demographic structure of the population of Riga in 1897 was as follows:
91
92
Valeriy Cˇekmonas
Denomination
Native tongue %
Orthodox Old Believer Catholic Protestant Other
39,851 9,062 26,820 183,021 22,929
14.2 3.4 9.5 64.85 8.12
Total
282,230
100
% Great Russian, Belarusian Ukrainian Polish Latvian German Yiddish Other
45,452 13,415 127,046 67,286 16,922 12,109
16.1 4.7 45.01 23.8 5.6 4.79
282,230
These data show that the percentage of Russians was almost the same in 1897 as one hundred years previously. Latvians came to dominate during the 19th century and the German-speaking population was only a third of what it had been. The peculiarities of the Russian language as it was spoken in Riga have been described by Želtov (1874), who may, by the way, be viewed as a pioneer of Russian urban dialectology. His description is neither exhaustive nor very accurate because he was not a professional linguist. Some phenomena mentioned by him might be irrelevant in the sense that they were not common to the Riga Russian speech (being characteristic of the idiolects of the individuals studied). It is a unique description, however, and it deserves a more thorough analysis than it was given by Semenova, who paid most attention to the German loan-words listed in this work. It follows from Želtov’s observations that the most common and important peculiarities of Riga Russian speech were: ikan’e (a pronunciation of /e/ in the pretonic syllables as /i/): diržat’ ‘keep’, vis’olyj ‘fun, merry’, bilit’ ‘white-wash, bleach’ and the verbal forms -mšy (cˇytamšy) ‘have read’, which dominated. Ikan’e is characteristic of some of the central and southern Russian dialects as distinct from the strong jakan’e (a pronunciation of /e/ in the first pretonic syllable as /a/ independently of the quality of the next accented vowel) of the Pskov dialect and the different types of jakan’e of the western central Russian dialect (DARJA I: map 3); ikan’e is a feature of standard Russian pronounciation (or prostorecˇie ‘urban popular language’) as well. So, Riga’s ikan’e could be the central dialectal feature or the Russian urban popular pronounciation. Verbal forms with the affix -mšy are found sporadically in the western and northwestern Russian dialects; their area begins in the eastern part of the central Russian dialect and extends further southwards and eastwards (DARJA II: map 111). They are also usual in urban popular language (prostorecˇie) as well. Thus, both features under analysis seem to point in the same direction. Other important characteristics of the Riga speech are: the hard ž ˇc š šcˇ (cˇystyj ‘clean’, šcˇy ‘cabbage soup’ etc.); the genitive dative na gory ‘on the mountain’, k
Russian varieties: Urban Russian
sestry ‘to smb.s sister’, na ruki ‘on smb.’s hand’ (cf. stand. Rus na gore, k sestre, na ruke ‘ibid.’); the dative instrumental plural s rukam ‘hands:inst.pl’, s nogam ‘feet:inst.pl’ cf. stand. Rus s rukami, s nogami ‘ibid.’); the accented accusative singular feminine e.g. na gorú ‘on the mountain’, na zeml’ú ‘on the earth’ (cf. stand. Rus na góru, na zéml’u); s domu/gorodu/lavki ‘from home/town/shop’ along with iz domu/gorodu/lavki ‘ibid.’; plural masculine and neuter of the type glázy ‘eyes’, rógi ‘horns’, úxi ‘ears’ (cf. stand. Rus glazá, rogá, úši ‘ibid.’); generalization of the endings -ov, -jev in plural e.g. mestóv ‘places’, prjaslóv ‘a traditional device for spinning with spindle’, guljan’jev ‘out-door fete’, ožidan’jev ‘waiting’ (all gen.pl, cf. stand. Rus mest, prjasel, guljanij, ožidanij ‘ibid.’); pronouns like taja ‘that:nom.sg’, tuju ‘that:acc.sg’, jevonnyj ‘his’, jejnyj ‘her’, ixnij ‘their’ (cf. stand. Rus ta, tu, ego ee, ix ‘ibid.’); and finally, less characteristic, because they are widespread in Russian dialects, are forms like pik’oš ‘bake:2sg.prs’, sik’oš ‘slash:2sg.prs’, birig’oš ‘take care:2sg.prs’; išol, išla ‘he, she went’; becˇ ‘to run’. All the features listed point to the Riga dialect having a northwestern Russian component, and all of them are observed in the Baltic Old Believer dialects (see below). Several minor but very interesting peculiarities registred by Želtov are kólokolo ‘bell’ (stand. Rus kolokol); forms like v gorodú/rukavú ‘in the town/sleeve’ (cf. stand. Rus v gorode, v rukave); d’oržyš ‘keep:2sg.prs’ (stand. Rus d’eržyš ‘ibid.’). The first of them is registred in Novgorod, Tver’ provinces, in the northern Russian dialect and even in Siberia (SRNG 1978, vyp. 14). The other forms are very rare in the northwestern dialect, being found more often in the eastern part of the northern Russian dialect but nowhere do they form a solid area in Russian territory (DARJA II 1989, map 17). D’oržyš is even rarer — according to SRNG (1972, vyp. 8) only traces of similar forms (d’oržat’ ‘keep’, d’orž ‘expenses’) are found in the northern Russian dialect. The data above seem to be in excellent agreement with the historic and demographic data. They make clear that the Riga Russian spoken vernacular had a complex origin in the sense that it was not a variety of any Russian territorial dialect. The demographic history of the permanent Riga Russian population sketched by Semenova helps us, to some extent, to understand its history. The Moscow Vorstadt, “a suburb settled by the working people”, “was the only truly Russian oasis in Riga in the last quarter of the 19th century” (Semenova 1977: 206); this oasis presumably existed during the entire century. Judging from ikan’e and mšy-forms, returning settlers from Central Russia formed the nucleus of the Moscow vorstadt traditional population. It was constantly being replenished by newcomers from the Tver, Petersburg, Novgorod and Pskov provinces and the western region. They must have brought some dialectal features into Riga’s Russian speech. The Russian population of Riga was not, however, simply a conglomeration of speakers of different dialects. Russian migrants to Riga were predominantly craftsmen. The guild structures were traditionally preserved in Riga even under the
93
94
Valeriy Cˇekmonas
Russian administration. There is reason to believe that the development of a spoken Russian urban contact vernacular might have been aided by most of the permanent Russian population in Riga being organized in guild corporations, concentrated in the Moscow forshtadt. The central Russian vernacular or, maybe, a variety of standard Russian (prostorecˇie) probably lay at the basis of this contact vernacular. Other characteristics of it could originate in the northwest. Some forms may have been introduced into the contact vernacular from the northern Russian vernaculars, but such forms may not be common. Želtov mentioned some such forms that were encountered in the Russian speech, but they could be characteristic of the seasonal workers from the Jaroslavl’ province who came to Riga every year (Semenova 1977: 204). In any case, the very existence of the Russian spoken contact vernacular distinguishes the Riga sociolinguistic situation from that in the other Baltic cultural centres. Another specific feature of Riga’s sociolingustic history is the cultural domination of German almost to the last quarter of the 19th century. Semenova (1977: 212) writes that the local variety of the German language was the language of the privileged part of the society. The position of German in Latvia was similar to the role of French in Russian society, with the difference that German was distributed more widely and on a more democratic basis. Thus, German was the language of social authority in Riga; it was not persecuted by the Tsarist administration to the same extent as Polish. It was even cultivated by educated non-Germans. The German-speaking autochthons in Riga participated in the local Russian social life. Due to its elevated status, German was an important source of loans to other languages spoken in Riga. Semenova (1977) discussed loan words in the Riga Russian contact vernacular. The first Russian newspaper under the title “The Russian weekly edition in Riga” (Rossijskoe eženedel’noe izdanie v Rige) appeared in Riga in 1816. It existed only one year because 100 subscribers could not be found. The newspaper offers material for the description of Russian in Riga, especially the Russian of Germans who have learned it from the local Russians (Semenova 1977: 199). The editor-in-chief of the paper was a certain Albanus, a local German (ibid.). The language of his short-lived publication was influenced by his native tongue, although it was basically the Standard Russian language. Albanus’s initiative marks the introduction of standard Russian into the Riga society. There are some other indications that the Russian language of the local Germans bore specific features due to influence from German, but this variety of Russian had not a local as its basis. Theoretically, it is conceivable that a German-Russian contact vernacular bilingualism was quite common in everyday life, but the developing German-Russian bilingualism of the second half of the 19th century involved Standard Russian. Semenova (1977: 203) writes that the first signs of Russian social life in Riga go back to the second half of the 19th century when a Russian newspaper, “Riga
Russian varieties: Urban Russian
Herald” (Rižskij Vestnik) started up. Then public Russian organisations came into existence, Russian schools were established and the city’s administration gradually began to substitute Russian for German. The position of the Russian language was strengthened as the Russian administration was institutionalised and the Russian press established etc. The Russian population grew constantly during this period, reaching 20% of Riga’s inhabitants as early as 1860 (compared to 3% or so of the Orthodox in Vilnius at that time). In spite of the rapid changes during the second part of the 19th century, the German language continued to be used. Moreover, the local Germans (known as the Ostsee Germans) had a chance to stand up for their traditional rights and privileges in the local and European press (to the great displeasure of some Russian publicists and politicians.8 This, Semenova (1977: 210–211) writes, is why the local variety of Russian was a subject to considerable foreign influence in its spoken and partly in its written forms. There was a direct German influence in the intelligentsia and propertied classes while the workers, the lowest class of the day, experienced Latvian influence and experienced the German influence through Latvian. The local languages, on the other hand, never noticeably affected Vilnius Russian, especially in its written form. Just as Polish and Lithuanian borrowed a number of Russian words “the local German speech was slightly influenced by Russian as the state language and Latvian as the local everyday language” (Semenova 1977: 211). In the same way as the Lithuanian dialects adopted many Russian and Polish lexemes, a large number of German and Russian words were introduced into the Latvian dialects, but it is not easy to distinguish between this kind of borrowing and direct loans from the Russian dialects. So much for the language of 19th century Riga. To conclude this overview we turn to the Russian province of Èstljandia that included most of present-day Estonia.
4. Russian in Tallinn Tallinn (Est Tallinn, Rus Revel’, Tallin, Ger Reval) was incorporated into the Russian Empire in 1710. The town retained its traditional rights and administration because its inhabitants surrendered to Peter the Great’s troops. It became an important navy base and a fortress of the Empire, but the Russian-speaking military garrison was a closed microsociety in an actually German-speaking town in the 18th century. The Russian population in Tallinn varied from 17.9% in 1820 to 19% in 1836 and 1844 to 11.3% in 1871, and to 10.2% by the end of the 19th century, while the number of Estonians was continuously increasing — from 34.8% in 1820 to 68.7% in 1897 (Istorija Tallina 1972: 41, 338). The local Germans dominated at the beginning of the 19th century (43.9% in 1820), and became the second largest
95
96
Valeriy Cˇekmonas
group towards the end of the century (17.5% in 1897), and they continued to decrease in the 20th century. Thus, unlike the situations in Vilnius and Riga, Tallinn’s population became steadily more Estonian in the 19th century into the first quarter of the 20th century. The development of educational institutions and the press in Tallinn reflects the dynamics of the sociolinguistic situation in the town. The Tallinn gymnasia (established in 1631) and Dome school (recorded in 1319 for the first time) were kept and supported by the The Knighthood of Estland; these were privileged establishments with German as the language of instruction. The first Russian Alexander gymnasium in Tallinn was founded in 1871. The first Russian district school existed in Tallinn from 1805 to the 1840s. The Tallinn primary schools, which began to open from 1804, taught in German and Russian. Russian began to be an obligatory subject in all schools from 1820 on (Pshenichnikov 1910: 21). An imperial order was issued in 1837 that, in three years of time, persons from the Ostsee provinces would be forbidden to teach in schools if they could not teach in Russian, but the order was suspended until 1845 (ibid.: 22). From 1850, office work in the Baltic provinces was to be done in Russian. “Istorija Tallina” (1972: 91–100) reports that it was not easy to realise this order because the students at the German schools did not know the state language well enough. By order of the Minister of Education, Russian was to be used as the obligatory and main language of instruction at state and private schools from 1887 onwards. In other words, the Tallinn population did not begin to be systematically Russified until the end of the 19th century. Russian obviously played a minor role in the social life of the town before the beginning of Russification and the situation did not change immediately after 1887. The history of the Tallinn press in the 19th century seems to support the notion that Russian had a marginal position in the town. According to “Istorija Tallina” (1972: 101–102), the only newspaper published in Tallinn until 1852 was the German “Revalsche Wöchentlich Nachrichten”. At that point, the official Russian Èstljandskie Gubernskie vedomosti began to appear (1853–1917). A special Estonian issue “Maa Valla Kuulutaja” of this newspaper was published as well. Two leading German daily newspapers — “Revalsche Zeitung” (1860–1915) and “Revaler Beobachter” (1879–1915) were published in the second half of the 19th century. The Russian Revelskij Listok appeared during 1878. Revelskie Izvestija was published for a longer period (1892–1917).9 These facts alone point to the modest role of Russian in 19th century social life in Tallinn, in spite of its status as a state language. It was the language of 10% or so of the permanent inhabitants of Tallinn at that time. Some rich Russian merchants and owners presumably knew German as a traditionally important language but the local German population did not generally know Russian, and German-Russian bilingualism was not common. The local Estonians and migrants from the Estonian inland learned German or Russian in primary school and the Estonian population was probably the most bilingual. It is easy to imagine their spoken Russian language:
Russian varieties: Urban Russian
because of the structural differences between Estonian and Russian it must have been heavily influenced by their native dialects, but the Estonian pronunciation could not, of course, be reflected in the Russian spelling. After 1887 we would expect Estonian-Russian bilingualism to spread, and that would become Tallinn’s main sociolinguistic feature in the first quarter of the 20th century. The second bilingual group would be the Russified local Germans.
5.
Conclusions
Summarising the above observations on the history of the urban Russian language in the southeastern Baltic area in the 19th and early 20th centuries we can say that: –
–
–
–
– –
it was introduced into the three main administrative and cultural centres of the area — Vilnius, Riga and Tallinn — primarily as the language of the Tsarist administration; efforts were made to Russify the educational and cultural life in these towns. The process in each case depended on the political situation in the corresponding ethnic areas; in spite of the Russification policy, bilingualism in Russian and the autochthonous languages did not become widespread in the Baltic until the end of the 19th century; the Russian language in local publications did not deviate noticeably from the Standard Russian of the day; only in some Riga Russian newspapers which were edited by the local Germans was the influence of German obvious; Russian was the main source of lexical loans for all the local autochthonous languages and dialects; the Russian language was not creolised in any case; that is, the original population was not Russified in any area and the local Russian dialects were not developed on the local substratum; in other words, Russian did not spread from the main towns of the area over the Lithuanian, Latvian and Estonian countryside. There were other reasons why the rural Russian dialects and population made their appearance in the Baltic area.
Notes 1. We shall use the historical names of the towns, Russian provinces etc. when it is of importance for our presentation; in all other cases the current names will be used. 2. This population was Slavicizied in the 19th century; the problem will not be discussed in our paper. 3. Kopija s zapiski, predstavlennoj G. Vilenskomu General-Gubernatoru pri pis’me G. Popecˇitelja Vilenskogo Ucˇebnogo Okruga, ot 25 Aprelja 1884 g. za N 2415 (Misius 1994: 411).
97
98
Valeriy Cˇekmonas
4. I know this from my own experience, because such persons often served as my interpretors in Suvalkija (West Lithuania) when I was not understood by younger people during my first trips to Lithuania as a student. 5. It is worth noting the prohibition on using the Latin alphabet was always treated as the “ban on the Lithuanian press” by the peasants and this tradition was adopted by the Lithuanian historiography and cultural conscience. 6. Unfortunately, the analysis below is not exhaustive because it is based on the texts published by Kulakauskas in his simplified transliteration (without jat’, (“) at the end of the words and discerning between two kinds of (i) of the originals which were inaccessible for me). All the other characteristics of the texts, as Kulakaukas informed me personally, were included. 7. Semenova does not mention the source she relies on. 8. Samarin (1890: 108–109), for instance, said that the Russian administration’s introduction of the Russian language into schools and in official use was unfruitful: “All the Baltic intelligentsia without exception are opposed to the intensification of teaching the Russian language”. 9. The authoritative Brockhaus & Efron’s Encyclopaedic Dictionary (Ènciklopedicˇeskij slovar’, XXVI, 1899, 417–419) informs us that Russian Revel’skij Gorodskoj Listok was published throughout 1890, Kolyvan’ throughout 1891 and Revel’skie Izvestija appeared from 1893 on. None of these publications was available to me.
References A. B. 1898. Rasporjaženija grafa M. N. Muraveva po delu narodnogo obrazovanija v SeveroZapadnom krae v 1863–1865 godax. Vitebsk. Aleksandravicˇius, Egidijus; Kulakauskas, Antanas. 1996. Caru˛ valdžioje. Lietuva XIX amžiuje. Vilnius: Baltos lankos. Balin´ski, M. 1935. Opisanie statystyczne miasta Wilna. Wilno. DARJA II 1989. Dialektologicˇeskij atlas russkogo jazyka. II. Morfologija. Moskva:Nauka. Ènciklopedicˇeskij slovar’, 1890–1907. Izdateli F. A. Brokgauz i I. A. Efron. Sankt-Peterburg. Goldberg, B. A. 1900. L. O. Levanda kak publicist (po slucˇaju sorokaletnego jubileja vozniknovenija russko-evrejskoj pecˇati). Ocˇerk. Vil’na. Gun, Otton. 1803. Topograficˇeskoe opisanie goroda Rigi s prisovokupleniem vracˇebnyx nabljudenij. Socˇinennoe v 1798-m godu i Ego Imperatorskomu Velicˇestvu posvjašcˇennoe Ottonom Gunom… S nemeckogo perevedeno Vasil’em Džunkovskim. cˇ. 1–2. V Sanktpeterburge. Istorija Tallina, 1972. Istorija Tallina (s nacˇala 60-x godov XIX stoletija do 1970 g.). Sostavitel’ Rajmo Pullat. Tallin: Eesti raamat. Jurginis, J., Merkys, V., Tautavicˇius, A. 1968–1972. Vilniaus miesto istorija. T. 1–2. Vilnius, Mintis. Koreva, A. 1861. Materialy dlja geografii i statistiki Rossii, sobrannye oficerami general’nogo štaba. Vilenskaja gubernija. Sostavil A. Koreva. Sankt-Peterburg. Kulakauskas, Antanas. 1986. “Laikinu˛ju˛ taisykliu˛ Vilniaus, Kauno, Gardino, Minsko, Mogiliovo ir Vitebsko guberniju˛ liaudies mokykloms” parengimas ir svarstymai de˙l lietuviu˛ kalbos vartojimo Lietuvos pradžios mokyklose. Lietuvos TSR aukštu˛ju˛ mokyklu˛ mokslo darbai. Istorijos serija 25, 37–45. Vilnius. Kulakauskas, Antanas. 1993. Penki 1882–1883 m. Rytu˛ Lietuvos valsticˇiu˛ kolektiviniai prašymai de˙l lietuviu˛ spaudos lotiniškuoju raidynu leidimo. In Lietuviu˛ atgimimo istorijos studijos. 4. Liaudis virsta tauta, 479–489. Vilnius: Baltoji varnele˙.
Russian varieties: Urban Russian
Kurzova, Zofia. 1993. Je˛zyk polski Wilen´szczyzny i kresów północno-wschodnich. WarszawaKraków. Levanda, L.O. 1875. Gorjacˇee vremja. Roman iz poslednego pol’skogo vosstanija. Sanktpeterburg. Levanda, L. O. 1875. Ocˇerki prošlogo. Rasskazy L. O. Levandy. Sanktpeterburg. LKŽ, 1978. Lietuviu˛ kalbos žodynas. T. XI. Vilnius: Mokslas. Mecˇkovskaja, Nina. 1989. Ruskymi slovami a slovenskym jazykom: O jazykovom soznanii Franciska Skariny. Russian Linguistics 13, 245–256. Merkys, V. 1978. Nelegalioji lietuviu˛ spauda kapitalizmo laikotarpiu (ligi 1904 m.) Politine˙s jos susiku¯rimo aplinkybe˙s. Vilnius. Misius, Kazys. 1994. Apie lietuviu˛ spaudos draudima˛ ir kataliku˛ dvasininkijos vaidmeni˛ ji˛ ˛igyvendinant. In Lietuviu˛ atgymimos istorijos studijos. 7. Atgimimas ir kataliku˛ bažnycˇia, 411–434. Vilnius: “Kataliku˛ pasaulio” leidykla. Murav’ev, Mixail. 1887. Tri politicˇeskie zapiski o S.-Z. kraje (zapiska 14 maja 1864 g.). In Sbornik statej, razjasnjajušcˇix pol’skoe delo po otnošeniju k Zapadnoj Rossii sostavil, S. Šolkovicˇ. Vtoroe izdanie. Vil’na. Polnoe sobranie zakonov Rossijskoj Imperii. 1862g. XXXVII n 37873. SPb, 1865. Pšenicˇnikov, P. G. 1910. Russkie v Pribaltijskom krae. (Istoricˇeskij ocˇerk). Riga: Izd. russkogo nacional’nogo kluba “Russkaja beseda”. Samarin, Ju. F. Okrainy Rossii. Tom os’moj. Izd D. Samraina. Moskva:Tipografija Mamontova. Semenova, M. F. 1977. Iz istorii jazykovyx otnošenij v gorode Rige. In Kontakty latyšskogo jazyka, 192–215. Riga: Zinatne. Slovar’ russkix narodnyx govorov (SRNG). 1972. Vyp. 8; 1978. Vyp. 14. Leningrad: Nauka. Tyla, Antanas. 1990. Slaptas lietuviu˛ mokymas 1862–1905 metais. In Lietuviu˛ atgimimo istorijos studijos 1. Tautine˙s savimone˙s žadintojai: nuo asmens iki partijos, 47–66. Vilnius: Sietinas. Vospominanija Teobal’da. 1890. Cˇas’t’ 2. Vilenskie vospominanija. Vil’na: Tipografija M. R. Romma. (Teobal’d. 1996. Jacek Kryštalevicˇ, vilenskij jurodivyj. Publikacija P. I. Lavrinca. Vilnius 4(149):149–157). Zinkevicˇius, Zigmas. 1990. Lietuviu˛ kalbos istorija. IV. Lietuviu˛ kalba XVIII-XIX a. Vilnius: Mintis. Zinkevicˇius, Zigmas. 1996. The history of the Lithuanian language. Vilnius: Mokslo ir enciklopediju˛ leidykla. Žuravskij, Arkadij. 1984. O kanceljarskom jazyke Velikogo knjažestva Litovskogo. In Lietuvos istorijos metraštis. 1983, 17–32. Vilnius.
99
AUTHOR "Valeriy Cˇekmonas"
TITLE "Russian varieties in the southeastern Baltic area"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Russian varieties in the southeastern Baltic area Rural dialects Valeriy Cˇekmonas
1.
Historical background
The rural Russian dialects in the Baltic area were spoken by Russian peasants who migrated illegally from the Russian inland during the 17th–19th centuries. Most of them were so-called Old Believers. The name is associated with the reform of the Russian Orthodox Church pioneered by the Tzar Aleksej Mixajlovicˇ (1629–1676). It was launched by Patriarch Nikon (1605–1681) in 1653 (Kapterev 1996 [1912] I: 106–307; Solov’ev 1991: 116–118; Zen’kovskij 1995: 185–374). Some of the clergy and many common people refused to accept the reform with its changes in rites, services and traditional sacred texts. This provoked a large-scale split (Rus. raskol) in the Russian Orthodox Church and the traditionalists were proclaimed by the church and state authorities to be Raskol’niki (dissidents). The dissidents referred to traditional Orthodoxy as the Old Faith and they started to call themselves Old Believers (Rus. starovery). In the course of time the Old Believers have, however, disintegrated into several denominations, all of which moved away from traditional Orthodoxy. The only old rite they actually preserved was the making the sign of the cross with two fingers (known as dvoeperstie). The Old Believers began at once to be persecuted by church and state powers (Kapterev 1996 [1912] II: 366–420) (a full account of the repressive laws and measures against Old Believers that followed their anathematization in 1667 is presented in Ènciklopediceskij slovar’ 1899: 295–303). The persecutions were particularly severe from 1667–1762, which some writers (e.g. Prozorov 1933: 4) refer to as a “period of martyrdom” in the Old Believers’ history. As a result of these persecutions the Old Believers fled en masse from their homelands from the 1660s onwards. Most of them went into the deserted northeastern territories of Russia and Siberia, but many people from the western and northwestern Russian territories fled to Livonia and the United Polish-Lithuanian Commonwealth. They sought religious and personal freedom in Livonia, Kurland and Poland and were welcomed by the
102 Valeriy Cˇekmonas
local landlords. When the Baltic area was incorporated into Russia, a stream of Old Believers moved westward from Livonia to the Polish-Lithuanian Commonwealth (Zavarina 1969:159–164; Zavarina 1977: 71–74, 84; Iwaniec 1977: 68). The Baltic Old Believer population continued to increase during the 17th–19th centuries until the abolition of serfdom in 1861. They sought religious and personal freedom in Livonia, Kurland and Poland. Catherine II changed the policy for dealing with the Old Believers. A general amnesty was announced in 1762 and a succession of repressive laws were repealed. For example, the Old Believers were allowed to settle in towns from 1782 and to be elected to public organisations from 1785. Due to the new policy, the Old Believers were not especially persecuted in the territories of the Polish-Lithuanian Commonwealth, which was finally incorporated into the Russian Empire in 1795. Catherine II’s policy was continued under Alexander I (1801–1825). The Old Believers were legalised during his reign. It is known, for example, that several new meeting houses were opened during this time in what is now Lithuanian territory (Prozorov 1933: 146) and the community in Riga was granted a degree of official recognition by a governor-general. Due to this Riga became a centre for Old Believers of the Fedoseev persuasion, which is professed by the bulk of the Baltic Old Believers. The first Old Believers’ chapel (or meeting house) in Vilnius was opened in 1830. A chapel had existed in Kaunas from the very beginning of the 19th century (Prozorov 1933: 140). The first urban groups of Old Believers are recorded as appearing in many towns during this period (not only in the Baltic area, but in Russia as well). Usually, they were merchants or craftsmen. Although the number of urban Old Believers grew throughout the century, they comprised only a small part of the Old Believer population in the Baltic area. The Russian Orthodox rural population in the Baltic was, of course, very small up until the 1840s. As a component of the Russification policy, the tzarist authorities initiated large scale resettlement of the rural Orthodox population from Russia to the Baltic provinces (gubernias) in the second half of the 19th century. The resettled Orthodox communities grew particularly rapidly after the Polish uprisings in the provinces of Vilnius and Kaunas. But in spite of the efforts exerted by the authorities, the Old Believers always dominated in the Russian population even there. For example, there were 730 Old Believer villages in the province of Kaunas in 1862 (that is, before Orthodox communities began to be settled there (Koreckij 1863: vi–vii)). In 1907, the total number of such villages (of various size) reached 999, but 754 of them were not entirely populated by Old Believers (Stankevicˇ 1909: lxviii). It is worth mentioning that these data were taken from the secret official sources. Orthodox settlers usually came in small groups or even individual families to live among the population who spoke another language or together with Old Believers and they mixed with the local people in some cases (ibid.).
Russian varieties: Rural dialects 103
Zavarina (1977, 1986) has investigated the history of the Russian population in Latgalia. On the basis of inventories of Latgalian country properties she concludes that the main now-existing centres of the Russian population in Latgalia existed already in 1760 (1986: 20). “The thought of settling down on favourable terms abroad, reliable concealment (from their Russian masters — V. Cˇ.), tolerance and the Polish landowners’ organisation of the escape routes undoubtedly stimulated the flight of the Russian peasants from their native soil” (ibid.: 24; Stankevicˇ 1909: iii). Judging from archive documents about captured persons, Zavarina concludes that among the Russian migrants from the second half of the 18th century until the beginning of the 19th century there were both Orthodox and Old Believers: “the Orthodox dominated, but because they settled mainly among the Russian Old Believers who lived there in large settlements, some Orthodox fell under the Old Believers’ influence” (ibid.: 23). The documents analyzed by Zavarina show that “the bulk of the Russian population in Poland (including Latgalia) and, in the 1760s, in Vidzeme as well were from Novgorod, Pskov and some from the Smolensk provinces” (ibid.: 21). These statements could serve as a general model of the origin of the Russian rural population in the whole Baltic area with some corrections for particular territories. The history of Lithuanian Old Believers was discussed in Nemcˇenko 1963:63–70. The most peculiar situation was on the Russian-Estonian ethnic border where the local Baltic Finnish population was becoming Russified as a consequence of Russian colonisation (Murnikova 1970: 5, 231). Scholars agree that most of the Baltic Old Believers came from the Pskov and Novgorod regions. This supposition is based on three sets of facts. In the first place, there are direct historical indications that Old Believers from these areas fled to contemporary Livonia (Livland), Kurlandia (Kurland) and Poland, and were then resettled further west (Zavarina 1968: 16; Iwaniec 1977: 82). Some of the most ancient Old Believer centres in these areas are found in the Old Believers’ chronicle called Deguckij Letopisec (1990). Secondly, all, or almost all, of the Baltic Old Believers are of the Fedoseev persuasion. The initiator, Feodosij Vasiljev was born in the village Kresteckij Jam of Novgorod gubernija. So, the Fedoseev persuasion “came into being and was initially spread in the Pskov-Novgorod area… Gradually, Fedoseev teaching was recognised by some of the population in the neighbouring Petersburg, Tver’, Jaroslavl’, Moscow and some other provinces” (Zavarina 1986: 16). Old Believers coming from other areas, especially frequent in the 19th century, were never more than a fraction of the Pskov and Novgorod stream (Zavarina 1986: 16, 21–23; Potašenko 1993: 55). Thirdly, the characteristic features (see below) of the Baltic Old Believers’ dialects have shown them to belong to the northwestern Russian Pskov dialect.
104 Valeriy Cˇekmonas
2.
Demographic data
It is not easy to estimate the size of the Old Believer population in the Baltic. They themselves have had no interest in making themselves known. The official Russian statistics shed no light on this matter because the Russian authorities always tended to hide the real number of Old Believers in the Empire (Prugavin 1904: 11; Kirilov 1913: 3, 5; a short overview of the problem is presented in Pozdeeva 1995: 6–8). Some historians admit that at least 100,000 Old Believers (1.1% of the whole population) lived in the grand duchy of Lithuania in 1791 (Korzon 1897: 320). The real numbers may be much higher given that of about 80,000 Russians in Vilna province in 1825 more than 85%, about 60,000 of them (Potašenko 1993: 58) were said to be Old Believers. The official Russian sources reported 18,000–19,000 Old Believers in Vilna province (which later was divided into Vilna and Kovno (Kaunas) provinces). An interpretation of data from 1860 gives the same number (ibid.: 98). According to V. Milovidov (1979: 29), there were at least 200,000 Old Believers in present-day Lithuania, Latvia and Estonia during the last quarter of the 19th century. The First All-Russian census showed considerably fewer Old Believers in the Baltic provinces: OldB Lifljandija Kurljandija Latgalija Vilna gub. Kovno gub. Èstljandija total
Zavarina (1986: 40–41) pointed out that there were about 67,000 Old Believers in Latgalia just twenty years before the census. According to Potašenko (1993: 97–101), the official figure of 60,000 Old Believers in Vilna and Kovno provinces corresponds to at least 100,000 in reality (compare Prugavin’s (1904: 11 assumption that the official figure of 2,137,738 Old Believers of the First census in 1897 may be interpreted as 20 million in reality). So, in spite of the official data, there were probably more Old Believers than Orthodox in the Baltic area in the 19th century. However, Old Believers were spread unevenly over the Baltic territory. The highest concentration of them was observed in Latgalia where about 45% of the total Russian population of Kurland province lived. Most Old Believers in the Vilna and Kaunas provinces lived in the eastern, northeastern and central parts of present-day Lithuania. In what was to become Estonia, the Old Believers’ villages
Russian varieties: Rural dialects 105
were concentrated predominantly along Lake Peipsi’s west coast and nearby areas (Murnikova 1960, 1970). This is still the basic distribution pattern. From the above facts, it is clear that the Russian peasant population, most of whom were supposedly Old Believers, comprised the second largest national group (after the autochthonous Lithuanians, Latvians and Estonians) in the Baltic area during the 19th century. They lived among the autochthons for several centuries and co-operated with them economically and in everyday life. This would in itself be enough to develop intensive language contacts. After the Baltic region was incorporated into the Russian Empire, even more favourable conditions were created for such contacts (because the local population was forced to learn Russian). The tzarist authorities tried to make use of Old Believers in their Russification policy by offering them the most favourable terms for farming (Sbornik…, 1886: 305–307; Stankievicˇ 1909: xxi–lv). But despite their socio-linguistic importance and demographic situation, Old Believers did not stimulate the developing contact between Russian and the local languages. They slightly influenced surrounding Lithuanian, Latvian and Estonian dialects but did not became bilingual themselves. Indirectly, by the very fact of their presence, they assisted the Russification policy. Typologically, the situation is not common and it needs to be explained. The Old Believers’ linguistic inertia can be accounted for in several ways. They sought isolation because of their origin. This was aided by their being scattered throughout the local population, though unwillingly in many cases. Furthermore, Old Believers were religious fanatics. One should agree on the whole, that The lives of the surrounding Lithuanians, Poles and Latvians were insignificant for a peasant Old-Believer’s religious consciousness. A Russian settler sought to preserve his religious and national identity in a foreign environment among the adherents of a different creed. Responding with the strengthening of his inner spiritual integrity under the pressure of everyday life, he sought support and found it in the faith of his parents, in the ancient rites, written records and books. (Markelov 1990: 169–170)
It should be added that Old Believers were on good terms with Catholics and Protestants, but tended to avoid contact with Orthodox people. In other words, they were isolated in a spiritual and religious sense, but obviously they had considerable contact with speakers of other languages in their everyday and economic life. Old Believers did not seek to acquire the values of the imperial Orthodox culture under Russia and the new Standard Russian. Linguistically, they lived in isolation even in large towns and because of that they preserved their spoken dialect and the Church Slavonic written language in a situation that would seem to be especially favourable for language contacts. It is known they avoided sending children to the Russian primary schools and did not co-operate with the Orthodox population in cultural and social life. Generally, the split with the Orthodox Church was always a
106 Valeriy Cˇekmonas
living reality for Old Believers; they and other Russians represented two opposing religious and cultural traditions. It is significant that Old Believers were “foreigners” for Russian statistics in the Baltic territories almost until 1860 and Raskol’niki (dissidents) for the Orthodox Church until 1905. Generally, Old Believers kept to their traditional values and were not Russified in spite of the efforts of the authorities. This was why the two Russian groups in the population — the Orthodox and the Old Believers — did not converge into a single group that could be of great consequence for the local people. The Old Believers’ voluntary isolation and their open hostility towards the Orthodox newcomers prevented the latter from becoming rooted in the new places and led to their coming back to Russia during World War I.
3.
Linguistic geography of the Old Believer dialects and their sociolinguistic characteristics
The Baltic Russian dialects were ignored by linguists for a long time. Research on them can be said to have begun with a collection of Russian folklore in pre-war Estonia (Murnikova 1968: 229). (A good deal of data was collected from 1928–1940 but was never, to my knowledge, studied by linguists.) No large-scale investigation of the Old Believer dialects was initiated until after World War II. Four conferences were held on the subject between 1962 and 1974 and the papers presented at them were published (Materialy konferencii 1962; Dialektologicˇeskij sbornik 1968; Trudy pribaltijskoj dialektologicˇeskoj konferencii 1970; Dialektologicˇeskij sbornik 1974). Material for a first bibliography was collected (Šulene 1963), and a considerable body of work on the lexicon, somewhat less on morphology and syntax and some on phonetics appeared in various academic editions. Semenova proposed a special programme for collecting lexical and syntax data for a regional atlas of the Old Believers’ dialects (1968: 5–13). The publications and presumably the field investigations had ceased by 1980. Sinica, a researcher of the Old Believers’ lexicon in Latgalia, deposited her manuscript on the subject at the Moscow library in 1983 (Lënngren 1994: 165). As far as I know, only Zincˇenko-Lënngren has continued these studies after the 80s, concentrating on problems of dialectal semantics (Lënngren 1994). The main results of the investigation in Latgalia were summarised as a tentative classification of the Old Believer dialects by Semenova (1964, 1972). The data from Lithuania and Estonia were not discussed. There is no overall map of the Old Believers’ settlements in the Baltic region although it is possible to list the villages in the former Baltic republics which were visited. Nemcˇenko’s (1963) article on the distribution of the rural Russian population in Lithuania based on the data of general census of 1959 is unique of that time. The main drawback of this account is
Russian varieties: Rural dialects 107
its failure to distinguish between the old Slavic population and post-W.W. II migrants from the Soviet republics. According to it, in two districts (Zarasai and Jonava) the Russian population comprised more than 10% of the total population; in eight districts, 5–10% of the population were Russians; eleven districts had 3–5% Russians, and eleven had 1–3% Russians. In the rest of the twenty-two districts the proportion of the Russian population was not more than 1% (Nemcˇenko 1963: 70). Most of the settlements with the largest Russian populations are concentrated in the Zarasai district: 19.18% of the whole population there is Russian and they live in more than 250 villages. Russians live in more than 50 settlements in the Rokiškis district and make up 9.1% of the whole population. …No fewer than 30 settlements can be counted in the Kaunas district. (ibid.: 71)
Dialects investigated during that period were called either Old Resident (Rus starožil’cˇeskie) or Old Believer although it is most likely that all of them were actually the Old Believer dialects in Lithuania and Latvia. There are no explicit indications in the published material that the dialectologists have found any Old Resident Orthodox Russian dialects or even any indication that such dialects comprised a significant part of the Russian Baltic dialects. This could be considered as one of the most important results of the post-war studies. In any case, practically nothing is known about the dialects of the 19th century Orthodox Russian migrants who seem to have vanished into thin air after 1914. This matter has never been the subject of scientific inquiry as far as I know. It is known, however, that a great many people were moved to Russia when the battlefronts of World War I approached the Baltic region. Lithuanians and Old Believers came back afterwards back, unlike, one can assume, the 19th century Orthodox migrants. Actually only Latgalia (including Daugaupils, Ilu¯kste, Re¯zekne, Kra¯slava, Dagda, Malta, Vil¸ani, Ludza, Zilupe, Karsava districts, see Figure 1) was investigated in Latvia. In Lithuania, the Rokiškis, Zarasai, Ignalina and Švencˇionys districts, which geographically can be treated as a southern extension of Latgalia were the main areas of interest (Figure 1). Nemcˇenko investigated the vernaculars of the largest Old Believer villages in the centre of Lithuania (Jonava district). In Estonia the dialects of several villages on the Estonian-Russian border and in the Tartu district were described. Unfortunately, no map of the investigated localities was ever published, but the authors cannot be faulted for this since it was prohibited.
108 Valeriy Cˇekmonas
T
Krs
P
Rz Ldz A
D
Zl
Dg
Kr
Rok Zr
Sh
J
Lnt
Vln
Figure 1.The Baltic Region (a schematic map). J = the central Lithuanian zone (Jonava district); Vln = Vilnius; Sh = Švencˇionys; Lnt = Lyntupy (Belarus); Rok = Rokiškis; Zr = Zarasai; D = Daugaupils; Kr = Kraslava; A = Ambel¸i; Dg = Dagda; Zl = Zilupe; Ldz = Ludza; P = Preil¸i; Krs = Ka¯rsava.
In the areas mentioned above, some dialects were more thoroughly covered than others. The full description of some Jonava village dialects was the subject of Nemcˇenko’s dissertation which was published in 1958, 1960, 1961 and 1963. The lexicon of the Latgalian dialects is comparatively well documented due to the work of Sinica (see bibliography in Lënngren 1994: 163–164). Novgorodov (1958) reports on the syntax of an area in Latgalia while Šulene (1963, 1965, 1968) studied the Zarasai district in Lithuania (where data were collected in 10 villages). Interesting phonetic observations of some dialects in Estonia were made by Xejter (1968, 1970, 1974) and Murnikova (1962). “Materials for the dictionary” of the Baltic Russian Old Resident dialects by Nemcˇenko, Sinica and Murnikova (1963) was the chief result of the studies on lexicon of the varieties spoken in Preil¸i (Latvia), Jonava (Lithuania), Jõgu and Tartu (Estonia) districts (Materialy 1963 further on). The transcribed texts in this work
Russian varieties: Rural dialects 109
provide data to futher our understanding of the phonetics, morphology and syntax of the dialects. This is of especial importance because the results of investigations into other levels were not generalised in a similar way. As I have learned from field-workers of the time, all the investigators attempted to interview speakers of the dialects who preserved their traditional speech and could not speak any other languages. The socio-linguistic description of the dialects was not treated as being of importance. As exception to this is an article (Sivickene 1979) referring to Russian-Lithuanian bilingualism in eastern Lithuania during the Soviet period. Many publications, however, contain interesting data about the influence of the surrounding languages on the Russian dialects, particularly in Estonia. The idea of combining the data from some dialects from Lithuania, Latvia and Estonia into a single vocabulary indicates that the principal goal in the studies of the Old Believer dialects was to find out what they have in common. This aspect may have been important because of the interest in the genesis of the Baltic Old Settlers. Lënngren (1994) follows this tradition as well. Nonetheless, the differences between the Old Believer dialects are of primary importance for studies of their contact with other languages. The nature of these contacts depended on the sociolinguistic situation in each area, but this was not taken into account at the time. The following remarks are therefore a program for future studies rather than an account of wellestablished facts. The linguistic geography of the Baltic Russian dialects is such that most of them are insular dialects, surrounded by Lithuanian, Latvian and Estonian vernaculars. There are, however, some Old Believer frontier dialects in the northeast of Latgalia and eastern Estonia which are not insular but rather constitute an extension of the northwestern Russian dialect. Both kinds of dialects could be classified as to whether they are in contact with Slavic or non-Slavic languages and further according to which languages are involved. The insular dialects are those spoken by the population living in single Russian villages or in clusters of such villages in central and northeastern Lithuania (Rokiškis, Zarasai, Ignalina, Švencˇionys) and in the many areas in Latgalia. There are, presumably, similar villages in the Võru and Tartu districts in Estonia, but their distribution is not precisely described. As far as I know, the Russian and local populations live together in the same villages in most other areas of Lithuania. According to the local people’s accounts, Old Believers usually sought to create microsettlements in mixed villages (“they keep together”) even if they were a small minority (of several families), but that is now a thing of the past. The Russian dialects in the Švencˇionis-Lyntupy area on the LithuanianBelarusian border and all along the Latvian-Belarusian border in Latgalia are defined as insular dialects in contact with the Belarusian vernaculars. Some insular Old Believer dialects are in contact with both Belarusian and Polish, e.g. those among the Belarusian or Belarusian and Polish speaking population in the Vilnius
110
Valeriy Cˇekmonas
Insular OB dialects in contact with the Slavic languages
in contact with Belarusian
in contact with the non-Slavic languages
in contact with in contact Lith., Lat., Est. with Polish
Non-insular OB dialects in contact with the Slavic languages
in contact with Belarusian
in contact with the non-Slavic languages
in contact with in contact Latvian and Estonian languages with Polish (?)
in contact with Belarusian and Polish
district and among the Polish speaking population in the Jonava and Kaunas districts in Lithuania, but no such Russian dialect was specially investigated. It is not clear if the Old Believer dialect exists among the Polish-speaking population in Latgalia because nothing is known about the Polish dialects in this area. Along the Lithuanian-Belarusian border (Švencˇionys-Lyntupy zone) Old Believers’ villages are scattered among the Belarusian ones. To our knowledge, there are no villages here with mixed Belarusian-Russian populations. The situation along the Latvian-Belarusian border is unknown to me. According to Karski (1903, 1918), there is a Belarusian population in the Daugaupils-Zarasai-Ilu¯kste area in Latvia all along the present Latvian-Belarusian border and in most of the Ludzu district. There are Russian Old Believer dialects here as well, but the socio-linguistic situation as a whole was not investigated in this area. It is felt that the Slavic population in the east part of the Ludzu district are Russian, not Belarusian. Since the Russian-Belarusian language frontier is not defined, the sociolinguistic situation remains unclear in this part of Latgalia. The frontier Russian dialects on the Latvian-Russian and Estonian-Russian borders should be considered separately from the insular dialects. First of all, some border vernaculars are represented on the maps of DARJA (1986, 1989) as part of the Russian inland dialects. Since the population on the border includes both Orthodox and Old Believers the contacts between the Russians and the local population may be more intense. The Russification of the local population here caused dialectal peculiarities that are unknown in other areas. In the next section, sociolinguistic and geographical aspects are discussed.
Russian varieties: Rural dialects
4. Linguistic characteristics of the Old Believer dialects 4.1 Phonetics All of the Old Believer dialects are characterised by the strong akan’e-jakan’e (or neutralisation of /o/ and /a/ into /a/ in pre-tonic syllable irrespectively of the quality of the following accented vowels, that is pronunciation like vódy ‘waters’, b’édy ‘troubles’, but vadá ‘water’, b’ad’á ‘trouble’, vadý ‘water:gen.sg’, b’adý ‘trouble:gen.sg’ etc.); this is emphasised or can be learned from the transcribed examples in every publication. According to the traditional classification, strong akan’ejakan’e is of course a fundamental feature of northwestern Russian Pskov dialects (DARJA I 1994: maps VI, 1, 3) and definitely supports the notion of Baltic Russian provenance from the historical Pskov territory and the southern part of Novgorod. Since no traces of okan’e (that is, an absence of the above mentioned neutralisation and pronunciation lika vodá ‘water’, b’edá ‘trouble’ etc.) were observed anywhere in Latvia and Lithuania, it is conceivable that either Old Believers from okan’e-dialects never migrated to the Baltic or they were only a small group and were totally assimilated into the Pskov group. In any case, the main stream of migrants must have been from those dialects that had featured akan’e by the end of the 17th century. Traces of ikan’e (a pronunciation like bidá ‘trouble’, bid’é ‘trouble:dat.sg’ etc.) and jekan’e (a pronunciation like b’edá, b’ed’é etc.) (along with the strong jakan’e) were noted in Latgalia (Nemceva & El’sberg 1970: 38–39), but it is obviously a recent innovation. Semenova admits the dissimilative akan’e (that is, the coexistence of v6dá ‘water’ — but vadé:dat.sg, vadý:gen.sg, vadój:inst.sg, vadú:acc.sg) is characteristic of the “Russified Belarusians” in Latgalia living along the border with Belarussia’. Nothing has been published on this kind of dialect, however. It is not known whether the Belarusians there were Russified during the Soviet period (which is possible) or earlier, which is more interesting. All the Baltic Russian dialects with akan’e are reported to have a reduced (weak) (6) ((;) after the soft consonants) in unstressed syllables except for pre-tonic open syllables at the beginning of a word (a type of agaród – v6garódi ‘kitchen-garden – in a kitchen-garden’) and open final post-tonic ones (a type of rabóta ‘a work’ – rabót6l ‘he worked’) cf.: g6lavá ‘head’, žáv6r6n6k ‘sky-lark’, avtór6k ‘Tuesday’, v;s’aléj ‘merrilier’, vós;m ‘eight’ etc. (Materialy 1963). Such combinations are characteristic of the southeastern Pskov dialect (DARJA I 1994: Maps 1, 3, 9, 17, 33). The Jonava dialect, with its soft palatalised (cˇ’) as in Standard Russian (Nemcˇenko 1958: 157) is unlike all the other Baltic Russian dialects, which are characterised by the hard (cˇ) as in the Russian Pskov dialect (cf. ˇc’alón6k ‘shuttle’, ˇc’ík6t’ ‘tick’ J — ˇcalón6k, ˇcýk6t’ L, E). (Here and henceforth the examples marked as J(onava), L(atgalia) and E(stonia) are drawn from Materialy dlja slovarja 1963.)
111
112
Valeriy Cˇekmonas
If Nemcˇenko’s records are accurate, this would mean that the Jonava dialect is unlikely to stem from Pskov ones. Nemcˇenko noted the affricative pronunciations of (t’ d’) as (c’ dz’) in the Jonava dialect remarking that these (c’ dz’) are different from the true affricates of Polish and Belarusian (1958:156). Materials from the other areas show only (t’ d’). Furthermore, I have found sporadic (s” z”) in one text from the Varakl¸a¯ni district in Latgalia (Semenova 1972: 18) and indications of sporadic lisping soft s” z” in the Dagda district (Novgorodov 1958: 353). Otherwise only the palatalised (s’ z’) are found in the material. The facts need to be explained because the pre-palatal (s” z” t” d”) are characteristic featurs of the mainland Pskov dialect. As I know from my experiences during the summer of 1997, the “normal” Pskov (s” z” t” d”) are found in the Old Believer dialects in the northeast of Lithuania, but they never have been noted there previously. It is possible the dialectologists were not expecting to observe these consonants and failed to notice or classify them in this and other areas. There are some other phonetic features that confirm or support the idea that the Baltic Old Believer dialects have a Pskov background: v (f) < u at the beginning of the word: vgadac’ ‘guess right’, vbic’ ‘kill’, vdaric’ ‘to hit’ (Nemcˇenko 1958: 154); fpirát’ ‘rest against’ (J, L), vjéx6la ‘she is away’, fp’órs’a ‘he set against’ (Novgorodov 1958: 352), cf. stand. Rus ugadát’, ubit’, udárit’ etc.; v-prothese before the back vowels: vul’ ‘bee-hive’ (stand. Rus ulej), vótrub’je ‘bran’ (stand. Rus ótrubi) (J, L), vútr6m ‘in the morning’, vóstryj ‘sharp’, vúžyn ‘supper’ (Novgorodov 1958: 356), vók6r6ka ‘ham:gen.sg’ (Semenova 1972: 18), vóz’;ra ‘lake’, vóstr6f ‘island’, vós;n’ ‘autumn’, vókny ‘windows’, vócˇyn’ ‘very’, vóstryj ‘sharp’, vútra ‘morning’, vútka ‘duck’, vugól ‘corner’, vušát ‘tub’, vušól ‘he went away’ (Nemcˇenko 1958: 157), cf. stand. Rus útrom, óstryj, úžin, ókoroka etc.; 6 < u in post tonic closed syllables: róz6m ‘intellect’, paz6xa ‘bosom’, žól6t ‘acorn’, vótr6b’ja ‘bran’, výs6š6t’ ‘to dry’ (Nemcˇenko 1958: 154), b6šavát’ ‘to rage’, b6rakí ‘beetroot’ etc. (Novgorodov 1958: 352); although pázuxa (J, L, E), rozum (L); u < y near labials in unaccented syllables: mužukí ‘men’ (J, L, E), puzurí ‘bubbles’, muzukánt ‘musician’ (Novgorodov 1958: 352); reflexes of the so-called “second pleophiny”: stalóp ‘pole’, ˇc’alón6k ‘shuttle’, m6lan’já ‘lightning’, žól6c’ ‘bile’ (Nemcˇenko 1958: 154); ˇcalón6k (L), ˇc6lanók (E), kór6m ‘forage’, v’ar’óx ‘top’ (J, L, E), cf. stand. Rus stol;, celnók, mólnija, želcˇ’, korm, verx. Even though the phonetics of the Baltic Old Believer dialects is not described consistently, the data from the published materials together with the corresponding maps from DARJA I 1986 (Maps VI, 1, 3, 9, 12, 16–17, 60, 91–92) point to their originating in the area of the Pskov Dialect. They were obviously a scattered continuation of the old Pskov dialects from northwestern mainland Russia. Some very important Pskov dialect features are notable by their absence. First of all,
Russian varieties: Rural dialects
cokan’e (or pronunciation of the etymological (cˇ) as (c)) in the Pskov dialects has disappeared quite recently (Cˇekmonas 1997: 182–187). The Old Believers left while it was flourishing in their homeland. Yet, cokan’e has not been noted in their dialects in Lithuania and only slight traces of it were found in other areas (Novgorodov 1958: 353; Semenova 1958: 597–599). Characteristic phenomena such as the loss of the distinction between hissing and hushing sibilants or voiceless and voiced consonants are much more rare in the Baltic Russian dialects than in mainland Pskov. (x) in place of (s) is noted in several words (xvísnut’ ‘whistle’, opojáxat’ ‘gird’, výplexnut’ ‘splash out’) in the central Lithuanian Russian dialects (Nemcˇenko 1968: 81) while it is observed more often in the Pskov dialect (Gluskina 1962: 32–33, B’jornflaten 1997: 16–25). Furthermore, Semenova noted only traces of specific core Pskov dialect forms such as mexát’ ‘mix’ (< mešat’) and k’afcá ‘bobbin’ (< céfka) in the central Baltic dialects. All these phenomena are better represented in the Russian-Estonian border dialects (Xejter 1968: 18–19). A comparison of even the incomplete and inconsistent phonetic data recorded from the Old Believer dialects in Latgalia seems to suggest that they could be nonhomogeneous (see Table 1). In the Ka¯rsava district, some villages right on the border with the Pskov region were investigated; the data recorded were used in DARJA. The dialects of these villages must be seen as mainland Russian ones. They differ from the other Latgalian dialects investigated by some essential features (having bilabial (w), soft (cˇ’); (s”, z”) along with (s’, z’); (c’, dz’) along with (t’, d’)). The rest of the Russian dialects under comparison differ slightly among themselves as well. There are several possible explanations for these facts. The phonetics of the Russian Old Believer dialects has not been examined thoroughly and some features have simply been noted. It does not apply to cokan’e; the almost total absence of it suggests that “Russification” must have taken place after resettlement to the Baltic and because of that the Baltic Pskov dialect may look like a paler version of the mainland one. The strong akan’e-jakan’e and lack of cokan’e seem to be the main features of Baltic Russian while the others are of less importance. The Baltic Russian dialects may be a kind of contact vernacular (koine) or, at least, were in the process of developing into such vernaculars. The data suggest that most of the Old Believers probably came from the southern and southeastern areas of the Pskov dialect, but their homeland cannot be pinpointed more precisely on this basis. 4.2 Morphology The morphology of the Baltic Russian dialects seems to have been better studied than their phonetics. Obviously the following features are common to all the Baltic Russian dialects: –
-y in the forms of the dative and prepositional cases (Novgorodov 1958: 359; Nemcˇenko 1958: M’urkxejm 1970: 111) as k žený ‘to smb.’s wife’, k vadý ‘to the
113
114
Valeriy Cˇekmonas
Table 1.Phonetic features in the Old Believer dialects of Latgalia LATGALIA District
Dagda
Vyški
Ambel¸i
a in pretonic syllables: 1st
strong akan’ejakan’e
the same but strong akan’ewith some devi- jakan’e ations
2nd
6/;
Preil¸i
Ka¯rsava
Phonetic characteristics
a in post-tonic ? closed slb.
6/;, sporadic a (str’ápat’ ‘to cook’)
usually a (gólas 6 (abéd6t’ ‘dine’) (usually a) ‘voice’) and a?
u: in post-tonic 6 (výr6gal slb; ‘scold’)
u
?
?
?
v, sporadic b. w the same before cons. and in word final position
(v only?)
(v only?)
w (transcribed as v)
(t’, d’)
(t’, d’ usually, c’, dz’ rarely)
in pre-tonic slb 6 (b6rakí ‘beet- u root:pl’) v
c, ˇc
both hard
c, ˇc/cˇ’
ˇc, (c?)
t’, d’
(t’, d’)
(t’, d’)
(t’, d’)
s’, z’
(s’, z’)
(s’, z’)
(s’, z’)
-dn-, -bm-
-nn-, -mm-
?
-nn-, -mm-
(-nn-)
-nn-, -mm-
šš, žž
štš, ždž
šcˇ/š’cˇ’, žž
štš, ždž (occ. ždž’?)
(štš, ?)
šš (occ. š’š’), žž
(c, ˇc’)
s’, z’, occas. s”, z”
Dagda (Novgorodov 1958), Viški (Nemceva & El’sberg 1970), Ambel¸i (Lënngren 1994), Preil¸i (Sinica 1964) and Ka¯rsava (Zverkovskaja 1985); the symbols (not words) in brackets show that the examples are taken from the transcribed text (not from the special descriptions) in the corresponding articles.
– – –
nagam, rukam ‘with, to smb. legs, hands’ for dative and instrumental plural, cf. (DARJA II map 41; Obrazovanie 1970: 394); deverbative forms like žýfšy, ušóccy ‘to live, to leave: 1–3sg-pl.pf’ as predicatives, cf. (DARJA II map 111); in J and some Latgalian dialects, verbal forms such as v’az’ó ‘he, she carries, drives’, l’úbi ‘he, she likes, loves’ (without -t in the forms of the 3rd person singular of the I and II conjugations) occur alongside the more common ones with -t (Nemcˇenko 1958: 167; Novgorodov 1958: 364; Sivickene 1988: 131; Lënngren 1994: 53). All these forms are found in the Western Russian mainland dialects (DARJA II: maps 2, 41, 111, 79);
Russian varieties: Rural dialects
–
the ending -i (-y) in the nominative plural as s’óly ‘villages’, balóty ‘swamps, mires’, glázy ‘eyes’, rógi ‘horns’, bakí ‘sides’, dróvy ‘fire-wood’, xl’avý ‘cattlesheds, cf stand. Rus s’óla, bolóta, glazá, rogá, boká, drová, xlevá.
It is not easy to compare the data published on the subject because of their incompleteness, but it seems that there are some other traits in morphology that could be seen as common to all dialects investigated (e.g. pronoun forms such as jon ‘he’, janá ‘she’, janó ‘it:3sg.neut’; tój ‘that:3sg.masc’, tája ‘that:3sg.fem’, tóje ‘that:3sg.neut’). However, the ending -ov/-ev (as in s6pagóf ‘high boots’, m’astóf ‘place’, rúžjef ‘rifle’ (all gen.pl), cf. (Novgorodov 1958: 359; Lënngren 1994: 52) is not noted in the central Lithuanian J dialect (Nemcˇenko 1958: 160–161). The morphology of the Baltic Russian dialects is generally more homogeneous than their phonetics. The comparative forms of adjectives with -óše, -ože (e.g. legóše ‘lighter, easier’, dolóže ‘longer’, cf. stand. Rus. légcˇe, dlinnée/dól’še) are worth special consideration. They are noted in J and southern Latgalia. In J, “the forms with -ej are often used in comparative forms of adjectives: svetlej ‘brighter’, krasivej ‘more beautiful’…” but “the forms with -óše, -óže are used in this function as well”: dolgój – dolóže ‘long – longer’, tolstój – tólšcˇe ‘thick – thicker’, mélkij – melóše ‘shallow – more shallow’, krépkij – krepóše ‘strong – stronger’, gládkij – gladóše ‘smooth – smoother’, tónkij – tonóše ‘thin – thinner’, sladkij – sladóše ‘sweet – sweeter’, rédkij – redóše ‘thin, rare – thinner, rarer’, l’ógkij – l’ogóše ‘light – lighter’, lóvkij – lovóše ‘adroit – more adroit’ (Nemcˇenko 1958: 162–163). Forms like kul’túrne ‘more cultural’, poméne ‘less’ (after the model bogácˇe ‘richer’, lúcˇe ‘better’) occur here sporadically (ibid). In Latgalia (Dagda district), forms with -ej (glybéj ‘deeper’, uzéj ‘narrower’) are common, while those with -oše “are derived from a limited number of adjectives”: lavóš6 ‘more adroit’, dalóž6 ‘longer’, tan’óš6 ‘thinner’, p6m’akóš6 ‘softer’ etc. (Novgorodov 1958: 361). Obrazovanie (1970: 413) pointed out that forms with -oše are only characteristic of the western part of the Pskov dialect. Data from map 59 (DARJA II) show -oše as the only affix used in the small western strip of this dialect, while it is used along with -ee (-ej) in the south and east (up to a line from Novoržev–Toropec–Dvina River). Thus, the comparative affix -oše serves as a reliable indication that the central Lithuanian Russian dialects come from the southeastern section of the Pskov Dialect. It also shows that at least the southern Latgalian dialects are related to this part of the Pskov dialect (the data from other zones were not reported). Some features of the Dagda district dialects in southeastern Latgalia described by Novgorodov suggest that they originate or have connections to the southern Pskov region. Among these are the forms with -u in the genitive singular (e.g. s frontu ‘from the battle-front’, na kan’u ‘on the horse’ etc. cf. stand. Rus s frónta, na koné) and with -of, -ef in the genitive plural (e.g. s6pagóf ‘high-boots’, zéml;f
115
116
Valeriy Cˇekmonas
‘lands’, the both gen.pl, cf. stand. Rus sapóg, zemél’ with zero morpheme), which occur widely here (Novgorodov 1958: 358–359; Lënngren 1994: 52; compare Zverkovskaja (1985: 178)). Both morphology and phonetics taken together clearly indicate a background in the southeastern part of the Pskov Dialect. Both suggest that the Russian dialects of the Baltic areas and sub-areas may have an admixture of other dialects, but it is hardly possible to unravel the origins of the mixed dialects on this basis. 4.3 Syntax Three major works were devoted to the syntax of the Baltic Old Believer dialects (Novgorodov 1958; Nemcˇenko 1960 and Šulene 1964). The latest short account is by Lënngren (1994: 53–60), based on data from one dialect (that from the area around Ambel¸i, situated between Preil¸i and Kraslava). This seems to be an attempt at generalisation for all Latgalian dialects (without saying it explicitly). The works mentioned cover different traits. The most extensive of them is Šulene’s unpublished dissertation on the vernaculars near Zarasai, which are representative of all the Baltic Russian dialects in many respects. My own experience suggests that syntactic peculiarities such as the following are the primary characteristics of Pskov and Russian Old Believer dialects as compared to Standard Russian: 1. Forms with -šy, -fšy, -cˇy (-cy), -mšy (originally adverbial participles, often called “a new perfect”) are widely used with a predicative function: narot sabráfšy ‘people gathered’, doma astafšy ‘remained home:1–3sg-pl’; ani ražžókšy kast’ór ‘they have made a fire’; ana prišóccy ‘she has come’ (Dg, 302), v jego naga slamafšy ‘his leg is fractured’ (J, 63), p’os sarváfšy ‘a dog broke loose’ (Zr, 319). a. These forms are usually derived from perfective verbs (in all dialects), and rarely from imperfective ones in Dg (štop ani etu kukuruzu séjafšy ˇcastej ‘if they soud this maize more often’, 303) and Zr (ani tut žýfšy ‘they lived here’; ani v izbý roššy ‘they grew out in a izbá’, 319). b. The -šy and -fšy forms of transitive verbs are used in Dg: (aný) zdelafšy lar’ok ‘(they) have made a stall’; ˇcorta výv’afšy ‘they have expelled a devil’; on vz’ómšy vnucˇku javonuju ‘he has married his granddaughter’ (303); in J, they are plausible (ja byla vyslasfšy pasylku ‘I had posted a parcel’; ja byl kusocˇek almazu pryv’ošcˇy ‘I had brought a bit of diamond’, 63) and are noted in Zr. I believe the usage of forms derived from transitive verbs as intransitive in constructions like v jego naga slamafšy ‘his foot is broken’; platje parvavšy ‘a dress is torn’, pugofka atarvavšy ‘a button is torn off’ (J, 63) is common everywhere in the Baltic. c. The -šy/-fšy forms are used with byl (-a, -i) in the pluperfect function: pešn’ú byli vz’afšy ‘they had taken an ice-pick’; byli pradelafšy darogu ‘they had built a road’; byla …kupifšy, byl prin’oššy ‘I had bought…, brought’ (Dg, 304), byli
Russian varieties: Rural dialects
pajexaccy ‘they had gone’, dom byl zgarefšy ‘a house was burnt down’ (Zr, 323), malad’ož byla sobrafšy ‘the young people had gathered’, fcˇeras’ syn byl prijexafšy ‘my son had come yesterday’ (J, 63). In Dg, constructions with budu, budet…+ -šy/-fšy -forms are encountered (budu n’a spafšy ‘I shall have not been sleeping (by that time)’, budit…porišoccy ‘he will have come (by then)’, budit pritamifšy ‘he will have been tired (by that time)’ (303) and -šy/-fšy forms are used in the subjunctive mood (byli b zakurifšy ‘would have smoked:1–3pl.subj’, byl by pripisafšy ‘would have get registred:1–3sg.subj’, byl by koncˇyfšy školu, ‘would have left school:1:3sg.subj’, 303). This was not noted in other dialects (but it does not mean, I believe, they do not exist over there). A tendency to develop a whole paradigm with the verbal -šy/-fšy -forms is beginning to emerge: Person, number
perfect
pluperfect and subjunc- future tive mood
1–3sg
(jest’) kupifšy, astafšy ‘have, has bought, left’ (see below)
byl (by), byla (by), kupifšy, astafšy ‘had (would have) bought, left’
1–3pl
(jest’) kupifšy, astafšy ‘have bought, left’
byli (by) kupifšy, astafšy budut kupifšy, astafšy ‘had (would have) ‘will have bought, left’ bought, left’
budu, budiš, budit kupifšy, astafšy ‘shall, will have bought, left’
The “normal” past forms with -l (kupil ‘bought:sg.masc’ kupil-a ‘bought:sg.fem, kupil-i ‘bought:pl; kupil by ‘bought:sg.masc.subj’ etc.) and “normal future” (that is, the pres forms of the perfective verbs like kupl’u ‘I shall bye’, kupiš ‘you will by:2sg’, kupit ‘he, she will bye’; and the constructions byt’ ‘to be’ + inf like ja budu, ty budeš, jon/jana budet kasit’ ‘I shall, you will:2sg, he/she will mow’) are more frequent in all dialects than the constructions with the -šy/-fšy forms. The contexts of preferences are not clear nor are their functions completely understood. According to Novgorodov (1958: 303), the -šy/-fšy forms are used in “clusters” by dialect speakers, that is, these forms might not be used for a long time during an interview, but once somebody uses them, they seem to trigger more of them. Finally, the -šy/-fšy -forms with full pronominal endings are found, “although not very often”, in Dg (300–301): ves’ stok praroššyj ‘the whole stack is sprouted’, xlep zam’orššyj ‘the crops are frosted over’, daroška abroššaja ‘a path is overgrown’, ja kak ras byl prijexafšyj z Dagdy ‘I have just come from Dagda’ etc.; they are not noted in the other Baltic Russian dialects and are very rare in the Russian mainland dialects (Kuz’mina & Nemcˇenko: 295–297).
117
118
Valeriy Cˇekmonas
2. The predicate is often expressed in the passive voice by short (nounal) passive participle forms: pojas dlinnyj p6tpajax6n ‘a long girdle is put on’; v’odry astavliny ‘the buckets are left’ (Dg, 302; J, 62; Zr, 410). a. In these constructions, agreement in gender and number is often lacking: ani byli v g’armaniju zabrana ‘they were taken away to Germany’ (Dg, 302, 319–320) (if agreeing with ani ‘they’, zabrany:nom.pl would be expected), plat’e takóje býla pašýt ‘such a dress was made’ (the agreement requires plat’je…pašýta:nom.sg.neut); rádiva býla pastávlen ‘a radio-set was put down’ (Lënngren 1994: 59) (should be rádiva… pastavlena:nom.sg.neut), a tebe dadena agarot ‘the kitchen-garden is given to you’ (not daden:nom.sg.masc, as expected), drovy nasecˇena ‘firewood is choped’ (not nasecˇeny:nom.pl), zavod pastrojena ‘a factory is built’ (not postrojen:nom.sg.masc) (Zr, 419), tam gorat rasstrojena ‘a town is sprung up over there’ (not rasstrojen:nom.sg.masc), tam skamejecˇka pastavlen ‘a small bank is put down over there’ (if co-ordinated, should be pastavlena), jon ˇcustvujet, ˇcto sestra jego vtoplen ‘he feels, that his sister is drowned’ (not sestra…vtoplena as axpected) (J, 62). b. Passive participle predicates can be used impersonally with v (u) + genitive (most frequently used with pronouns): v men’a kopana, narvana ‘I have it dug out, picked’; v nas atsážana ‘we have it planted’, v men’a tavaru vz’ata ‘I have taken some goods’, v nas tr’apak nábrana ‘we have got a lot of clothes’, bat’ka zaxaronena ‘(smb’s) father is buried’, seni atkryta ‘an innter porch is opened’ (Zr, 314, 411, 419; Dg, 302; J, 63). c. Passive participle predicates can be used with byl, byla byli ‘was, were’: ani byli v g’armaniju zabrana ‘they were taken away to Germany’ (Dg, 302), byl pastrojena dom ‘a house was built’, byla dadena jamu žarabec ‘a stallion was given to him’ (Zr, 422). d. Passive participles as predicates may also be in full (pronominal) form: kartoška uže davno vykapanaja ‘potatoes have been lifted long ago’, sapagi ploxa ššytyje ‘the high-boots are badly made’ (J, 62), ješcˇ’o lošad’ zan’ataja ‘the horse is busy’, tam toža kast’ola pabitaja ‘there is a ruined (Catholic) church over there as well’ (Zr, 310, 311), etad dom jašcˇo ne byl pastroinyj ‘this house has not been built yet’, zemli…pabrasanyji ‘the farm-land is deserted’ (Dg, 301). These pronominal forms are encountered more often than those with -šyj (-aja, -šyi). Generally, forms containing past participles comprise a paradigm that is isomorphic to that of the -šy/-fšy forms:
Russian varieties: Rural dialects
Tense, person passive perfect 1–3sg or v+1-:gen.sg (v men’a… etc.)
1–3pl or v+1–3gen.pl
passive pluperfect and passive future subjunctive mood
(jest’) zabran(a), byl (by), byla (by) zabrany(j) ‘am, is taken’ zabran(a) ‘had been taken (would been taken)’
budu, budiš, budit zabrana ‘shall, will have been taken’
(jest’) zabran(a,y)
byl (by), byla (by) zabran(a) ‘had been taken (would have been taken)
budit zabran(a), budut zabrany ‘will have been taken’
(jest’) zabrany ‘are taken’
byli (by) zabrany ‘had been taken (would be taken)’
budim, budite, budut (zabrany) ‘shall, will have been taken’
(jest’) zabran(a,y) ‘are taken’
byl (by), byla (by) zabran(a), byli (by) zabrany
budit zabrana, budut zabrany ‘will have been taken’
It is not clear to what extent this paradigm is actually realised in the dialects, because the forms of 1–2sg have not been noted by dialectologists; so, it can be treated as a programme for further investigations along these lines. The function participle forms in the Northwestern and Northern dialects were interpreted in the same way (Kuz’mina 1975: 255, 1982: 272). We can call these shadow paradigms since it is possible to use l-forms instead of the perfect, pluperfect and subjunctive mood and the “normal” future instead of the “participle future” without changing the meaning (that is, to transform ja kupil(a)/ja kupifšy/v m’ane kuplina ‘I have bought’; ja byla by kupila/ja byla by kupifšy ‘I would have bought’; v jago zabrali/v jago zabtafšy/v jago zabrana ‘smth. has been taken from him’; v jago zab’arut/v jago budut zabrafšy ‘smth. will have been taken from him’ etc.). Their identity is obvious in both the Baltic and the Northwestern Russian dialects. Constructions of the type u psa ubežena ‘the dog run away’, tak byla žyta ‘it has been lived in this way’ (that is, with passive participles from intransitive nonreflexive forms) are absent (or not noted) in the Old Believer dialects. This is of great importance for the localisation of their Russian mainland base. According to Kuz’mina & Nemcˇenko (1971: 107–114, Map 2), this sort of construction is not known south of a line between (approximately) Opocˇka, Velikie-Luki and Xolm, that is, in the southern and eastern part of the Pskov dialect area (Obrazovanie 1970: 396). This fact was never taken into account in previous investigations. 3. Comparative constructions with the subject of comparison being expressed with za + accusative are common to all the dialects under discussion: jon krasivej za
119
120 Valeriy Cˇekmonas
teb’a ‘he is more beautiful than you’, bol’šyj za vas ‘bigger than you’ (J, 77), ana starš6 za Vovu ‘she is older than Vova’, bol’š6 za brata ‘taller than (smb.’s) brother’, niž6 za vas ‘shorter than you’, latyskij jazyk legcˇ6 z6 n’ameckij ‘the Latvian language is more easy than German’ (Dg, 294–95; Lënngren 1994: 57), ty d’užej za men’a ‘you are more stronger than I’, los’…bol’še za karovu ‘an elk is bigger than a cow’, pomalože za dedušku ‘younger than (smb’s) grandfather’ (Zr, 173–175). “Contaminated” word-groups with the genitive as a standard of comparison such as lucˇˇc6 za toj lampy ‘more better than a lamp’, vyš6 za vaš6va rost6m ‘taller than you’ are noted in Dg (Novgorodov 19581: 295). Both constructions are common to the southern part of the Pskov region, as can be seen in the materials of the Pskovskij oblastnoj slovar’ (1994: t. 10, 283). The comparative with za is familiar to the other East Slavic languages, but “it is almost unknown in the Russian dialects” (Proxorova 1991: 57). 4. A construction of the type priexat’/vernucca/vyjti … z (s) ‘come back, come out’ + genitive of a noun (domu, lavki, gorada… ‘house, shop, town:gen.sg’) is used in our dialects (J, 71; Lënngren 1994: 56; Zr, 152). It is widely distributed in the northwestern Russian dialects, but is lacking in the southeastern part of the Pskov dialect area, that is, in the Seliger-Torzok group of dialects (Obrazovanie 1970: 400). This fact suggests the eastern limits of the Pskov dialect to which most of the Baltic Old Believer dialects can be related. Other syntactic features are those shared by all the northwestern Russian dialects (and, in some cases, the Northern dialects as well) and those specific to the Baltic Old Believer dialects. The foremost of them are the following: 5. The preposition v (u) is used with the instrumental plural of nouns to express the meaning of the Standard Russian u kogo ‘whose, who has/have’: v kupcax vsegda deneg mnoga ‘the merchants always have much money’, v r’ab’atax tam vsevo našli ‘it was found a lot of things that belonged to children’, kupili v pomešcˇikax ‘have bought from the landowners’ (J, 75), f starikax sprašyv6l ‘asked the old people’, f xaz’ainax žyli ‘lived with their masters’ (Dg, 295, 296–297), v l’ud’ax ‘between people’, v kabanax ‘boars have smth.’, i v litovcax …gavar’at tak ‘and so Lithuanians say as well’ (Zr, 97). This construction is characteristic of some of the western dialects (Kuz’mina & Nemcˇenko 1964: 324) and is known throughout the Pskov region (Gomonov 1962: 82; Pskovskij oblastnoj slovar’: 1976: t.3 12–13). 6. Genetivus existentialis (or the subjective genitive) is noted in all the Baltic Russian dialects: jest’ (u nas) takix starikov, l’udej, gribov, vs’akaj ryby ‘there are old men, people, mushrooms, all kind of fishes here’ etc. (J, 65; Novgorodov 19582: 253; Zr, 372–374; Lënngren 1994: 60). It is a common feature of the northwestern and northern Russian dialects (Kuz’mina 1993; Kuz’mina and Nemcˇenko 1976; Russkaja dialektologija 1965: 194).
Russian varieties: Rural dialects
7. In Dg and Zr, jest’ ‘is’ can be used in a predicate with the verb in the present, past or future tense or in the participle and -šy/-fšy forms as well (Novgorodov 19581: 305): jest’ žyv’o starik ‘the old man is living’, on…jest’ prijexal ‘he has come’, (oni) jest’ palucˇut ‘they will have got’, jest’ prišoccy ‘have/has come’, abrublin jest’ ‘is chopped off’ (Dg), jest’ fundament stoit ‘the foundation is built’, syn jest’ ubityj ‘(smb’s) son is killed’ (Zr, 335). This is also a characteristic of most of the northwestern and northern Russian dialects (Russkaja dialektologija 1965: 194). 8. The nominative is used with (n’a) nada ‘it is (not) wanted’: vypivka nada ‘a drink is wanted’, nada sila ‘strength is necessary’, ne nada kur’atina/cˇužyje/lotka ‘chicken-meat is/strangers are/a boat is not needed’ (Zr. 43, 391, 392), nada r’amont/dagl’at ‘be in need of repare, care’, nada v’adro/jablaki ‘a bucket is, apples are wanted’ (Dg, 320; Lënngren 1994: 60) n’a nada i klad ‘even a treasure is not needed’, den’gi n’a nada ‘money is not necessary’ (J, 65). The construction is known everywhere in the western central Russian dialects and is widely encountered in the northern Russian dialect as well (Obrazovanie 398). 9. The same forms are used for the accusative and nominative plural of animate nouns (lošadi, karovy d’aržali ‘(they) keep horses, cows’, cf. lošadi, karovy:nom.pl) in all the dialects under discussion. The following characteristics could be viewed in one or another way as being specific to the Baltic Russian dialects. 10. The non-prepositional dative is normally used to express the experiencer in all the dialects that have been investigated: mne (jamu) zub/galava balit ‘I have (he has) a tooth-ache, a head-ache’, /gar’acˇka byla ‘(smb.) was in fever’ (Zr, 56), jamu naga raspuxla ‘his leg is swollen’, vam apitit pajavicca ‘you will have an appetite’ (Dg, 323), mne uže i nogi zabaleli ‘my feet already hurt’ (J, 66). The possessor may also be expressed in the same way: nam koni byli ‘we keep horses’, mne želudok xarošyj ‘I have a good stomach’ (Zr, 55), nam tož6 muki netu ‘we have no flour either’, matka pam’orla et6mu kuznecu ‘this smith’s mother is dead’ (Dg, 323), jemu ne bylo vrem’a ‘he had no time’ (J, 66). The use of the dative for the experiencer is found in the northwestern Belarusian dialect (DABM, map 223), and is encountered in some vernaculars in the Smolenks region (Proxorova 1991: 45–46, Map 3), but is not characteristic, as far as I know, of Russian dialects. 11. The locative with v for acc. of direction: byl v Rigi ujexacˇˇcy ‘I have/he, she has come to Riga’ (cf. Rus v Rigu), v l’asu byli pašoccy ‘they have been in the forest’ (Dg, 297) (cf. Rus pošli v les ‘came in the forest’), pajexacˇˇci v Rigi ‘he, she has/they have come to Riga’, pajd’om v l’asu za gribami ‘let us go mushrooming in the woods’ (Zr, 78–79) J. Such constructions are not noted in Russian or other East Slavic dialects. 12. The adverb rovna ‘straight, exactly’ is used as a preposition with the meaning of do ‘up to’: klever obsejano rovna balota ‘clover is sown up to the bog’, rovna sažalki
121
122 Valeriy Cˇekmonas
atmerina ‘it is measured off up to the pond’, vada rovna mostika padnimalas’ ‘water rose up to the little bridge’ (Zr, 270–271); zales rovn6 gorla v vadu ‘he dipped in water up to the neck’; pastrigla volasy…rovna glas ‘she cut her hair up to her eyes’, sum’ot rovna galavy ‘a snow-drift is head-high’ (Dg, 293), vady rovna kalena ‘water is knee-deep’ (J, 71). Rovna is not used in such way in the East Slavic dialects. 13. The construction s + instrumentalis for the name of an object to manipulate such as s chulkami abuvacca ‘to wear stockings’ (literal. ‘with stocking’), ja s platocˇkam byla zakutafšy ‘I have put a kerchief on my head’, rubit’ s toporom ‘to use an axe to chop smth.’ is recorded in Zr, 154–155; Šulene (1963: 31). The construction is believed to be supported by the Lithuanian language or borrowed from it. A similar phenomenon is known in other Slavic languages, but it is not a characteristic of the Russian dialects (Kuz’mina 1985; Proxorova 1991: 61–62 and maps 14–15). 14. A very peculiar use of net in negations is observed in Dg: ana net takaja sil’naja ‘she is not so strong’, on net rodstvennik ‘he is not smb’s relative’ (311–312), cf. stand. Rus ona ne sil’naja, on ne rodstvennik. M. A. Novgorodov (1958: 312) considers it to be a calque of the Latvian construction of the type vin¸š nav spe¯cı¯gs ‘he is not strong’. 15. Word order where the possessive genitive precedes dependent elements e.g. žany atec ‘father-in-law’ (liter. wife’s father, cf. stand. Rus otec ženy) is noted in Dg (313–314), but I have observed it in Eastern Lithuania as well. Such word order is usual in the northwestern Belarusian dialect and in the Polish vernaculars in Lithuania. Thus, the syntax, phonetics and morphology of the Baltic Russian dialects all indicate that they originated in Pskov. The syntax alone appears to have been influenced by the Baltic languages (see below). 4.4 Lexicon The lexicon of the Baltic Old Believers received a good deal of attention from the dialectologists. Literature on this subject is surveyed in Lënngren (1994). Semenova (1972: 14) attempted to trace lexical differences between individual dialects in Latgalia. Materialy (1963) presents data demonstrating lexical differences (however not discussed in an explicit form) between the main groups of the Old Believer dialects. The relationship of one group of semantic lexemes (denoting the state and condition of a person) to the Russian mainland dialects was analysed by Sinica (1970: 188–203). Understandably, all the data collected has not been analysed from this point of view. The lexicon of the Pskov dialect and of the other Russian dialects had not been studied at that time so any linguistic geographical interpretation of the findings in Baltic Russian could only be provisional.
Russian varieties: Rural dialects
The main problem was, however, that there was no procedure or parameters for such an interpretation. The most popular method was one-to-one comparison of semantic groups of words or lexical sets noted in the Baltic Russian dialects with those of mainland Russian dialects. The data obtained are, of course, of interest. For instance, Sinica (1964: 55) has shown that words denoting natural things and phenomena in one of the Latgalia dialects are, for the most part the same as in Standard Russian…. The lexicon of our dialect is basically that of the northern Russian dialect…. All the groups under investigation resemble the Pskov and Novgorod lexicon. …Of the southern words, stávka ‘a pound’, xmára ‘cloud’, rakíta ‘broom’, stéžka ‘path’, sosónnik ‘pinery’, kriníca ‘a well’, pozavcˇóra ‘the day before yesterday’ are used. … All this attests to the northern Russian basis of our dialect.
Sinica’s account is not very clear in itself. From about 30 “northern Russian words” 10 words or so are “of the Pskov-Novgorod lexicon” as well (as rog ‘cape’, rožók ‘cape’, zen’ ‘shadow’, glýža, glýžina ‘clod’, ves’ólka ‘a kind of birch’, závtirje, nazávtrie ‘tomorrow’, naútrie tomorow, in the morning’). Several of the “northern Russian words” (bolotína ‘a small bog’, vixór’ ‘whirlwind’, glušína ‘a kind of birch (with rough leaves)’, ves’ólka ‘rainbow’) are known in the Pskov dialect as well (POS 1967–1995). So the idea that the Latgalian dialect has a northern Russian dialect basis is not justified. The most serious question is, however, how the northern dialectal basis with okan’e acquired the strong akan’e-jakan’e. Lënngren (1994: 61–79) compared some semantic groups of the Ambel¸i (Latgalia) Russian vernacular to the Pskov dialects, in a continuation of the studies on the Baltic Old Believer lexicon. It is not clear from the text of the book if the units of Ambel¸i dialect only (collected by the author) are investigated or if the comparison is based mainly on Sinica’s materials from the Preil¸i district. This is important because if Lënngren operates only with her own data she has established the lexical identity of the two dialects. The genetic relationship between the lexicons of mainland Russian and Latgalian Russian dialects is not the principal focus of her multipurpose analysis; she starts by assuming the latter are based on the Pskov dialects. The potential of the lexical data to more precisely trace the genesis of the Baltic Russian dialects has not been exhausted, however. For example, the principle word for ‘flail’ in the Baltic Russian dialects is prív’;s’ (m6latít’ priv’az’ám ‘to tresh using flails’ (Materialy 1963: 253; Zavarina 1986: 139–140; Rixter 1970: 177–178). Etymologically prív’az’ is obviously derived from priv’ázyvat’ ‘to bind’. In this meaning, cep is known in the dialects which are in contact with Belarusian ones (Sinica 1963:59) and I can confirm this with my own observations in Švencˇionys district (on the Belarusian border to Lithuania). In Jonava cep is used for the ‘(short) swinging rod of a flail’ (Materialy 1963). The absence of the lexeme cep for ‘flail’ in the Pskov region was noted by some ethnographers even in the 19th century (Nikiforovskij 1895:442).
123
124 Valeriy Cˇekmonas
The DARJA material on this subject made it absolutely clear that priv’az’ is the only word for ‘flail’ in the main part of the Pskov dialect proper (Kormakova 1991: 149). The names for ‘a swinging rod’ built from the root cep-/kep- (cf. cep in J) are used in Pskov, Novgorod and some of the northern Russian dialects (ibid. 151). Finally, a name in Baltic Russian for the ‘long handle of a flail’ cavin(k)a (< *cevina > *cev; ‘a rod’ etymologically) (Materialy 1963: 340) with k’avina as a variant (Semenova 1958: 598) corresponds to the forms cev’jó, cevína, kevjó, kevína, tevjó, tevína ‘the same’ in the western part of the Pskov (including the Gdov district) and Novgorod dialects with the east-north border almost along a line from the Lovat’ River–the Il’men’ Lake–St. Petersburg (Kormakova 1991: 150; B’jornflaten 1993: 9–10). Evidently, the word for flail and its components alone help us to trace the background of the Baltic Old Believer dialects with the certainty comparable to some phonetical and morphological traits mentioned above.
5.
Language contacts
The Baltic Russian dialects fall into three groups as far as their contacts with other languages go. The first comprises two frontier strips of ethnic Finno-Slavic confusion: an area in southeastern Estonia of what are known as “Pecˇora Half Believer” (Rus poluvercy) or Setu settlements (Pecˇora Half Believers are the local Orthodox Estonians) and (2) the Russian-Estonian villages northwest of Peipsijarve, where “Ijzaku Half Believers” are located (Murnikova 1960: 46–47, 1963: 120–121, 1970: 3). Dialects whose speakers have a long tradition of bilinguism in Russian and the local languages represent the second and smallest group. All the other Baltic Russian dialects fall into the third group which can be qualified as the dialects with superstratum influence. The words “ethnic Finno-Slavic confusion” in the above paragraph refer to two phenomena. Speakers of the Russian dialect have adopted or are in the process of adopting the Estonian language (or dialect) in the Ijzaku area. The Ijzaku Russian dialect has presumably in its turn arisen on the Vodish substratum; it has been heavy influenced by Estonian as well (Murnikova 1960: 45–47, 1963: 120–123; Xejter 1968: 14–15). In the Pecˇora district, the Setu speech is merging with Russian. Substratum-adstratum relationships are observed in both cases. The Setu Russian speech has not been investigated. The peculiarities of the Ijzaku Russian dialect are described in a succession of works by Xejter (her dissertation on the subject was not available to me). It is not surprising that all levels of this Russian dialect are heavily influenced by the Finnish substratum, especially its phonetics (Xejter 1979). The Russian dialect of the Mehikoorma district represents the second group of Baltic Russian dialects. Mixed marriages and conversion to the Lutheran denomination occurred here. “Not only the adult population, but the children worked here
Russian varieties: Rural dialects
as labourers in the rich Estonian farms” (M’urkxejn 1973: 2). “These circumstances favoured the development of bilingualism in the Russian milieu” (ibid.; M’urkxejn 1971: 3–6) and nowadays (that is, in 1973) “there is no sharp distinction between the communicative functions of the Russian and Estonian languages”; even “in the ‘very Russian’ families, the members of the families often speak Estonian between themselves” (M’urkxejn 1973: 2). The Mehikoorma dialect is greatly influenced by the Estonian language as a result of this kind of contact. The influence is noted on all levels, from phonetics to the lexicon (M’urkxejn 1968, 1973; Kuz’mina 1974 and Nemcˇenko 1974 on syntactic calques from Estonian). It must be emphasised that the Russian dialects in contact which are so permeable for alien influence in both Ijzaku and Mehikoorma areas are spoken by the Orthodox population. Nothing of the kind has happened to the Baltic Russian dialects of the third group, which represents the Old Believers, and we must agree with Lënngren (1994) that this fact has not been sufficiently emphasised in previous studies. No phonetic traits were reported that could be considered as a result of foreign influence on the Old Believer dialects. In morphology, there is one phenomenon suspected to be of this sort and it is a variation in the gender of nouns. Sivickene (1979) has described both a change of neuter to feminine in nominative singular of nouns and an expansion of the new feminine paradigm and variation in gender of the borrowed words. Both phenomena occur in all areas of Lithuania. The first means using most historical neutra as feminine (e.g. ad’ajála ‘blanket, quilt’, arúdija ‘tool’, vóz’era ‘lake’, górla ‘throat’ žála ‘a sting’, m’ésta ‘a place’, vr’em’a ‘time’ etc.). Sivickene listed 50 words or so, not only in nominative singular (e.g. majá t’éla stáraja ‘my calf is old’), but in the other grammatical cases as well (at’éc zanimál xaróšuju m’estu:acc.sg ‘smb’s father occupied a good post’; v adnój m’és’ci:loc.sg ‘in one place’). Thus, prósa ‘millet’ has changed its gender almost completely (only the dative singular of masculine was noted), while the other words fluctuate. Also, many borrowed words which are masculine in Standard Russian are declined (with the wide range of fluctuations) as feminine in the Old Believer dialect (arganízma ‘organism’, g’aktára ‘hectare’, gípsa ‘gypsum’, grípa influenza’, kas’t’óla ‘Polish Roman-Catholic church’, kilam’étra ‘kilometre’, lítra ‘litre’, tórfa ‘peat’, céntra ‘centre’ etc., more than 40 words are cited). Analogous facts are reported from Latgalia (Novgorodov 1958:357). The neuter gender is absent in Lithuanian and Latvian which suggests that these languages may have influenced the changes. But it is hardly a direct influence like Estonian has on the Ijzaku Russian dialect which resulted in a “confusion of genders” like adín dacˇká ‘one sister’, mat póm’er ‘smb.’s mother died’, tri s’astrá ‘three sisters’ etc. (Xejter 1968: 18). If this were the case, the words vóz’era ‘lake’ d’éla ‘matter’, vr’ém’a ‘time’, pól’a ‘field’, s’éna ‘hay’, ím’a ‘name’, mucˇ’énija ‘torture’, for example, would become masculine, because their Lithuanian counterparts are of that gender (e˜žras, reı˜kalas,
125
126 Valeriy Cˇekmonas
laı˜kas, lau˜kas, šie˜nas, var˜das, kankinimas respectively). The trend is to change neutra to feminine in the mainland Russian dialects (Vysotskij 1948: 83; Cˇagiševa 1968) and the words stáda ‘herd’, séna ‘hay’, déla ‘matter’ are cited as the commonest examples of such change (Russkaja dialektologija 1965: 100). Outside influence of the Baltic languages could stimulate the further development of this trend in the Baltic Russian dialect but it is unlikely direct “Baltic pressure” given these facts. Gender change in borrowed words designating new cultural phenomena may be due to these words having been borrowed relatively recently not from Russian, but from Lithuanian or Latvian. Their gender may be an example of hypercorrection in these circumstances. The failure to distinguish the plural and singular of the 3rd person of verbs noted in the dialects of the east Lithuanian districts may be a result of Lithuanian influence. The phenomenon was described by Sivickene (1974) who has cited more than 30 examples of type ni l’ac’íc’… kamarý ‘the gnats are not (liter. is not) flying’; travú p6dajót … žénš’cˇiny ‘women give (liter. gives) grass’; aní vaz’m’ót … ‘they will take’; aný … nibaícca ‘they do not (liter. they does not) afraid’; sabáki b’ažýt i lájit ‘the dogs are (liter. is) runing and barking’ etc. She remarks that “this interference is observed in a situation of incomplete bilingualism: not all the speakers of the Russian dialect in the Soviet Republic of Lithuania speak Lithuanian. Many of them are passive bilinguals, that is, they understand the other language, but cannot speak it” (Sivickene 1974: 98). It is worth emphasising that, according to Sivickene, “some people of the middle and older generations do not understand Lithuanian”, but “texts with singular instead of plural forms are recorded from these persons as well” (ibid.). But failure to distinguish these forms is not a common phenomenon for any group of speakers. It occurs “very inconsistently and is a rarity compared to the many common forms of plural. It is observed in a spontaneous speech only and requests for repetition elicit only the common forms” (ibid. 97). These data suggest that the failure to distinguish the plural and singular of the 3rd person of verbs is quite a new feature, caused by developing bilingualism which is obviously unusual to the Old Believers. The lexicon of the Baltic Russian dialects was not as impenetrable to foreign influence as the other levels. Nemcˇenko (1968) has a long list of words that he believes to be loan words into central Lithuanian Russian dialects. There are many Lithuanian words among them. Such words as uškur’ ‘a married man who lives in his wife’s house’ (< Lith užkury˜s), burbúl’ ‘small ball’ (< Lith. bur˜bulis ‘bubble’), vága ‘furrow’ (< Lith vagà), lúpa ‘lip’ (< Lith lu¯´ pa) may be older loan words because they are known in other languages and dialects (Lauch’ute 1982: 10, 31, 34). Words like kupst ‘hummock’ (< Lith kùpstas), vapsvá ‘wasp’ (< Lith vapsvà) are known in Polish and Belarusian dialects in adapted forms (cf. Belr vósapa, asva ‘wasp’, Belr kups’ó, Pol kups´cik ‘hummock’, ibid.: 29, 31).
Russian varieties: Rural dialects 127
The following words are noted only by Nemcˇenko: gír’a ‘pine forest’ (< Lith girià), kilbúk ‘gudgeon’ (< Lith kilbùkas), medátka ‘marigold’ (< Lith dial. made˜tka), šalbabón ‘a kind of haricot’ (< Lith dial. šabalbónas), sklip ‘a strip’ (< Lith skly˜pas), kumetína ‘farm labourers’ house’ (< Lit kumetýnas, dial kumetýne), greb’ázd ‘pole (on a roof)’ (< Lith dial. grebe˙´zda), brunkl’a ‘a small plank used to tether a cow’ (< Lith dial brunkly˜s (-is)), alíva ‘a kind of lubricating oil’ (< Lith alyvà), nágini ‘bast shoes made from leather’ (< Lith nãgine˙s), kápla ‘mattock’ (< Lith kaply˜s), lísva ‘bed (in kitchen-garden)’ (< Lith lýsve), ámar ‘small insect vermin’ (< Lith ãmaras), dálba ‘crowbar’ (< Lith. dalbà), kar’úki ‘small sledge’ (< Lith dial. kariùkai), rúnkol’ ‘swede’ (< Lith ruñkelis), lótra ‘drunkard’ (< Lith lãtras), kréicˇi ‘dowry’ (< Lith kraitis), lína ‘rope, line’ (< Lith. lýnas), žal’úka ‘a kind of potato’ (< Lith. žaliùkas), dýl’a ‘squared beam’ (< Lith. dile˙˜). The phonetic form of these words is minimally adapted, and it is of interest that none of these words were included into Materialy (1963). There are obvious Lithuanisms-barbarisms in Nemcˇenko’s list as well e.g. pádanga ‘tire’ (Lith padangà), val’díba ‘administration’ (Lith valdýba), taríba ‘Soviet’ (Lith tarýba), plítin’a ‘brickworks’ (Lith ply˜tine˙), pénin’a ‘dairy’ (Lith píenine˙), medelín ‘nursery garden’ (Lith medelýnas), víršinik ‘chief ’ (Lith vir˜šininkas), pírminik ‘chairman of a kolkhoz’ (Lith pìrmininkas), v’adéjas ‘(he) manager’ (Lith vede˙jas), v’adéja ‘(she) manager’ (Lith vede˙ja), s’an’únas ‘seniorman’ (Lith seniu¯˜ nas), nóvadas ‘inspector’ (Lith núovadas), gírnik ‘forester’ (Lith gìrininkas), amatókas ‘handcraftsman’ (Lith amatiniñkas), g’agužín’a ‘May-feast’ (Lith gegužìne˙), vákaras ‘evening’ (Lith vãkaras), vajdínimas ‘play’ (Lith vaidìnimas), ikurtúvi ‘house-warming’ (Lith ikurtuve˙s), spálvy ‘colours’ (Lith spalvos), trintuk (tyrantúk) ‘eraser’ (Lith trintùkas), súgerimas ‘absorbtion’ (Lith sugerìmas), skaitl’úki ‘counting frame’ (Lith skaitlùkai), skaicˇávimas ‘counting’ (Lith skaicˇiãvimas), dajlíraštis ‘calligraphy’ (Lith dailýraštis), pajšíba ‘drawing’ (Lith paišýba), nósin’a ‘handkerchief ’ (Lith nósine˙), pátajsa ‘correction’ (Lith pataisà), títnagas ‘flint’ (Lith tìtnagas), darbaden’a ‘work-day’ (Lith darbódienis), pažiméjima ‘certificate’ (Lith pažme˙´jimas), rušávimas ‘sorting’ (Lith ru¯šiãvimas), balsávimas ‘voting’ (Lith balsãvimas), bal’sovat ‘vote’ (Lith balsúoti), pósedis ‘sitting’ (Lith póse˙dis), susirinkímas ‘meeting’ (Lith susirinkimas). Mostly they are the Lithuanian words for Soviet realia (for instance, val’díba, taríba, pírminik, v’adéjas, v’adéja, darbaden’a, pósedis, susirinkímas, bal’sovat’), school-life (e.g. spálvy, trintúk (tyrantúk), súgerimas, skaitl’úki, skaicˇávimas, dajlíraštis, pajšíba) and various other new phenomena. Most of the loans into Russian dialects given by Nemcˇenko could be treated as Polish in origin. Some of them are originally Lithuanian words (e.g. rojst ‘marsh (with growing bushes)’ (< Lith raı˜stas), l’úny ‘guagmire’ (< Lith liu¯´ nas), rapúha ‘toad’ (< Lith rupu¯že˙), víksva ‘sedge’ (< Lith viksvà), šešók ‘polecat’ (< Lith še˜škas), pašúra ‘shed’ (< Lith pašiu¯´ re˙), šúly ‘posts, poles’ (< Lith šùlas:sg), pakúla ‘tow’ (< Lith pãkulos:pl), žabárý ‘long switches’ (< Lith žãbaras, žabaraı˜:pl’), šakáli ‘splinters’ (< Lith šakaly˜s, šakaliaı˜:pl), klúmpi ‘wooden shoes’ (< Lith klùmpe˙,
128 Valeriy Cˇekmonas
klùmpes:pl), dirván (derván) ‘long-fallow land’ (< Lith dirvónas), pl’ant ‘highroad’ (< Lith pléntas), kump’ák ‘gammon’ (< Lith kumpis), ˜ káry ‘a kind of cart or sledge’ (< Lith dial. kariùkai)), some are Belarusian (e.g. vav’órka ‘squirrel’, rézgoviny ‘a traditional home-made device to carry hay’, drapák ‘cultivator’, búsel ‘stork’ etc.); all of them are used in the local Polish dialects. The other words from Polish, which are known over all territories of the former Grand Duchy of Lithuania are of the sort špak ‘starling’, pac ‘rat’, gának ‘porch’, kómen ‘chimney-stack’, l’úfta ‘ovendoor’, tynk ‘plaster’, kl’ámka ‘latch’, skrýn’a ‘trunk’, pras ‘flat iron’, knot ‘wick’, lancúg ‘chain’, drot ‘wire’, šrub ‘screw’, uxnál’ ‘a long nail to shoe a horse’, gúma ‘rubber’, papéra ‘paper’, rajtúzy ‘long pants’, kavnér ‘collar’, kapšúk ‘tobaccopouch’, kápa ‘bedspread’, dyván ‘carpet’, firánka ‘curtain’, kort ‘a kind of fabric’, bavélna woolen cloth’, lazánki (lázanki) ‘a kind of home-made noodles’, cepelíny ‘zeppelins (a kind of dish made of potato dough)’, kind’úk ‘tripe, chitterlings’, loj ‘suet’, árfa ‘winnowing-machine’, cibúl’a ‘onion’, ágrast ‘goose-berries, goose-berry bush’, bruk ‘cobble-stone road’, šróty ‘small shot’, tarták ‘sawmill’, gospodýn’a ‘hostess’, škaplérnik ‘amulet’, šváger ‘brother-in-law’, švagérka ‘sister-in-law’, bankárt ‘natural baby’, baxúr ‘teenager’, šl’úb ‘wedding ceremony’, páterri ‘prayers’, lajdák ‘a dissipated person’, šcˇíryj ‘frank, sincere’, (a)škarádnyj ‘greedy’, fájnyj ‘fine’, rudýj ‘red, red-haired’, brúdnyj ‘dirt’, retovát’ ‘save’, fundovát’ ‘treat’, kirmaš ‘fair’, rožáncy ‘rosary’, krejda ‘chalk’, kráta ‘grating’, pl’áma ‘spot’, brud ‘dirt, filth’, pádla ‘carrion’, škápa ‘cupboard’, armáta ‘cannon’, bizun ‘lash’, cibúk ‘chibouk’. The origin of these words in Polish is not discussed here. Nemcˇenko’s list of foreign elements in the central Lithuanian Russian dialects is the longest of all that have been published in the works on loan words in Baltic Russian. In Marcˇenko’s (1970) account of loan words in the Russian dialects of the Zarasai district only two words of Lithuanian origin are mentioned (sklip ‘lot’ and pašúra ‘shed’), all the other being from Polish. Sinica (1964) found only ten loans from Latvian among more than 1200 dialectal words. Nemceva (1968) analysed pairs of synonyms in Latgalian dialects. They are usually made up of one original Russian word and its Polish or Belarusian (but not Latvian) counterpart, e.g. zolápópel ‘ashes’, kryl’có-gának ‘porch’, luk-cibúl’a ‘onion’, kuznéc-koval’ ‘smith’, pógreb-sklep ‘cellar’, truba-kómin ‘chimney’, útka-kácˇka ‘duck’, kukúška-zazúl’a ‘cuckoo’, svád’ba-veselje ‘wedding’, kl’úkva-žuravína ‘cranberries’ etc. There are only xor’ók-šešók ‘polecat’ (< Lith še˜škas, cf. Lat. sêsks, sêska, sêskus) and morkóvkabarkan ‘carrot’ (< Lat barka¯ns, barka¯ni)); both the Baltic words in these pairs could have been borrowed straight from Polish. Nemceva emphasises that the data were obtained from old people (ibid.: 83). Thus, Nemcˇenko’s account that many words borrowed from Lithuanian are commonly used in J and some other dialects suggests a rather unusual situation. His data may be interpreted in the following manner. The Lithuanian loan words are new to Russian dialectal speech and they were obviously elicited from people who
Russian varieties: Rural dialects 129
are familiar with Lithuanian. (Nonetheless, all these words were listed as genuine loans in Lauch’ute (1982: 86–91).) The other Lithuanian loan words may be words known to some, especially younger, speakers of the dialect but not to all of them and not commonly used words. His information may also be true of the informants living in Lithuanian settlements with Russian families as a small minority. These data are good (Nemcˇenko was born in an Orthodox family in the Jonava district) but obviously do not reflect an overall, typical and traditional situation in the Baltic areas with a concentrated Old Believer population. The majority of the Russian dialects in Estonia are not exceptions in this respect (Murnikova 1962; Materialy 1963). Sinica (1964) found only 10 Latvian loan words among 1200 dialectal words in the Russian dialects of Latgalia. In the Ijzaku and Mehikoorma areas, however, dialectal Russian speech “is marked by an abundance of borrowing from the Estonian language” (Xejter 1968: 18; M’urkxejn 1973: 2–5) which is natural for a dialect with heavy substratum influence (see above). All the authors agree that the bulk of loan words in the Lithuanian and Latvian Russian dialects are from Polish. This reflects the importance of Polish as the main cultural dialect in the Grand Duchy of Lithuania. The fact is of considerable interest because it shows the Lithuanian and Latgalian Old Believers were not linguistically isolated from the languages surrounding them. Polish was a sacred language for the Catholic population, and it was “a noble language” for the rural population of the Grand Duchy of Lithuania. Old Believers had their own scale of socio-linguistic values. Old Church Slavonic was a sacred language for them, and a kind of “Slavonic-Russian bilinguism” was possible in their milieu (Murnikova 1970: 7). The influence of the written Church Slavonic language on the Old Believer dialect needs to be investigated. Old Believers did not usually have any knowledge of the less socially prestigious Baltic languages, it must be admitted, and their dialects were never saturated with loan words from them. They could use Polish as the most prestigious language in contacting with the outer world. Unfortunately, this aspect of the sociolinguistic situation in the Old Believer communities was not investigated directly and it is not certain any information can be obtained now. The data on syntax shed light on the Old Believer language contacts. On the one hand, all the peculiarities of the southeastern part of the Pskov dialect are preserved in them (see features 1–12 above). But, on the other hand, some very specific innovations, which can be treated as loans from Lithuanian and Latvian, are noted in them as well (features 9–14). Of them, the experiencer dative (as in mne galava balit ‘I have a headache’) can theoretically be a native feature because, as was mentioned above, there are traces of it in some Smolensk vernaculars and it is possible it may have existed in the other northwestern Russian dialects. The construction s tufl’ami abuvacca ‘put on smb.’s shoes’, s ˇculkami ad’avacca ‘wear the stockings’, s platocˇkam zakutafšy ‘to have put a kerchief on one’s head’, rubit’ s taparom ‘to chop smth. using an axe’ (Zr) is less likely to be native rather than borrowed.
130 Valeriy Cˇekmonas
Since the other dialectologists did not mention it, however, this construction may not be characteristic of all Baltic Russian dialects and because of that it may be quite recent in Zr. I could not find out whether this construction is encountered in the speech of elderly illiterate people. If so, it cannot be borrowed (calqued) directly from Lithuanian because the older people have never been bilingual or known Lithuanian well enough to be able to introduce a foreign construction into their native speech. Constructions with net ‘no, not’ (such as ana net sil’naja ‘she is not strong’ Dg) are not encountered in the Lithuanian territory and may be considered as not very old, dated to the first part of the 20th century, when Russian-Latvian bilinguism became possible. Constructions like rovna gorla/ozera = standard Rus do gorla/ozera ‘up to one’s throat/lake’, noted in Dg, Zr and J, are probably earlier loans (cf. Lith lìgi ‘up to’ < ligùs ‘flat, even’; Lat lı¯dz ‘ibid’ etymologically related to lı¯gt ‘to even’, and combinations like Lith lìgi gerkl˜˙es/e˜žero ‘up to one’s throat, lake’, lat lı¯dz ezeram ‘up to lake’). But, again, the very existence of earlier syntactic borrowing contradicts the fact that bilinguism was not usual for the Baltic Old Believers. Actually, syntactic borrowing (calques) is not possible without knowledge of a foreign language. The contradiction could be resolved by the assumption that the loans in question may not have been introduced directly from the Baltic languages but rather from local Polish speech, which would, as it was shown above, influence the Old Believers’ language. Almost nothing is known about the Latgalian Polish dialects. The Lithuanian ones commonly use constructions with the preposition s ‘with’; Pol. równo ‘up to (stand. Pol do)’ (like jeden kon’ równo szyi utonł ‘a horse was sunk up to its neck’) is noted in the Vilnius district by Šulene (1964: 271). Further, the experiencer dative is common in the Polish language (cf. mnie głowa boli ‘I have a headache’). Generally, all the above-listed (and some unlisted) peculiarities of the Baltic Russian dialects’ syntax also exist in the Lithuanian Polish dialects. So, as in the case of lexical loans, the local Polish dialect could be a mediator for the Baltic languages’ influence on the Russian Old Believer dialects. 5.1 Old Believer dialect influence on the surrounding dialects The possibility that the Old Believer dialects may have influenced the dialects around them has not been investigated until now. Theoretically, such an influence is quite conceivable. Since Old Believers did not mix with the local population but had every-day contact with them, only lexical borrowing from the Old Believer dialects can have occurred. But these loan words cannot easily be distinguished from loan words from the Russian language proper since the lexicons of these two varieties are basically similar. The local Baltic population mainly borrowed terms for social and cultural realia from the Russian language. The Old Believers themselves needed this kind of term and they could acquire them more easily and faster
Russian varieties: Rural dialects
than the non-Russian population. They may have acted as a bridge for loans from Russian into the local Baltic languages, but it is difficult to find evidence for this. As far as we know from published sources, no characteristic dialectal words that could be considered as loan words from the Old Believer dialects into local ones have been recorded. The Latvian Latgalian vernaculars are full of Slavisms (Endzelin 1899; Semenova 1959: 596). Rek¸e¯na (1962) has shown that more than half of the 1500 dialectal words used in the Kalupe Latvian vernacular are loan words from Slavic languages. The same picture is revealed in her thorough investigation of trade terminology in some southern Latgalian dialects (Rek¸e¯na 1975:539–566). But all the studies show that the loan words could be classified as Slavic, Russian, Polish, or presumably Belarusian, but in no single case is there evidence for a direct loan from the Old Believer dialect. The same can be said about Slavic loans into the Lithuanian dialects. According to the data in the “Atlas of the Lithuanian Language” (LKA I 1977), for example, there are no Russian loan words in the Lithuanian dialects whose geography is confined to the areas with the Old Believer population.
6. Concluding remarks During a summer expedition in 1996 I was able to make 12 hours of good tape recordings in 14 villages with Old Believer populations in northeastern Lithuania, the Zarasai, Ignalina, Švencˇionis and Rokiškis districts. It was the first linguistic field investigation of these dialects for more than 30 years. I believed it made little sense to resume the investigation because all of these villages have been visited many times in the past. I did not hope to find any speakers of the traditional dialects. But the situation turned to be more interesting than expected. There are quite a number of villages where Old Believer dialects are still spoken, although they are only the remains of previously large settlements. Their inhabitants are all elderly. The traditional closeness of the Old Believer communities is a thing of the past. According to our informants, the traditional way of life in the communities started to change after World War II when the Old Believers’ children began to attend state schools. In ethnic Lithuanian territories, many of them attended Lithuanian schools. As the young people moved into the towns mixed marriages gained acceptance. These changes have affected the speakers of the traditional dialects. Few are dialect speakers in the proper sense of the word. Many more of them are bilingual in Russian and Lithuanian with varying proficiency in their second language; some even speak Lithuanian better than Russian. In almost every village you can meet some so to say “Russified” speakers whose language can be defined as Standard Russian with some dialectal features. This is due to having lived in Russia for a long period or close contacts with post-war migrants from Russia. The main language in
131
132
Valeriy Cˇekmonas
the mixed families of the younger generations is inevitably Lithuanian if one of the spouses is Lithuanian by nationality. The results of the influence of languages in contact are reflected variously in individual ideolects. Some of the speakers, for example, better preserve strong akan’e-jakan’e, others avoid using the perfect participle, e.g. pryšóccy, pam’óršy etc. Present-day Old Believer dialects can be said to be represented by a multitude of ideolects that are modified in more than one way. Generally, the Old Believer dialects have been in close contact with the surrounding languages — Lithuanian, Standard Russian (in towns), more rarely Polish (in southern Lithuania) in the second half of the century. In other words, what did not occur during several centuries of isolation has happened in the course of the last fifty years. The current sociolinguistic situation in the villages visited is presumably more similar to that in Ijzaku and Mehikoorma than to what it was at the time of the visit.
References B’jornflaten, Jan Ivar. 1993. Pskovskie govory v obšcˇeslavjanskom kontekste. In Norvežskie doklady na IX-om s”ezde slavistov, Bratislava, sentjabr’ 1993 g., 7–22. Oslo. B’jornflaten, Jan Ivar. 1997. Opyt lingvogeografii Pskovskoj oblasti. In Pskovskie govory. Istorija i dialektologija russkogo jazyka, 8–30. Oslo: Solum forlag. Cˇagiševa, V. I. 1968. K voprosu o kategorii srednego roda v mestnyx govorax. Pskovskie govory II: 146–159. Pskov. DARJA I, II. 1986/1989. Dialektologicˇeskij atlas russkogo jazyka. Vypusk I. 1986. Fonetika. Vypusk II. 1989. Morfologija. Moskva: Nauka. Dialektologicˇeskij sbornik. 1968. Dialektologicˇeskij sbornik. (Latvijas Valsts Universiteta¯tes zina¯tniskie Raksti, N92.) Riga, Zvaigzne. Dialektologicˇeskij sbornik. 1974. Dialektologicˇeskij sbornik. Materialy IV dialektologicˇeskoj konferencii po izucˇeniju govorov i jazykovyx kontaktov v Pribaltike. Oktjabr’ 1972. Vilnius. Ènciklopediceskij slovar’. 1899. T. XXVI, kn.51. Sankt-Peterburg. Èndzelin, I. 1899. Latyšskie zaimstvovanija iz slavjanskix jazykov. Živaja starina. God 9, vypusk III, 285–312. Sankt-Peterburg. Èstljandskaja gubernija. 1905. Pervaja vseobšcaja perepis’ naselenija Rossijskoj Imperii, 1897. T.XLIX. Sankt-Peterburg. Gluskina, S. M. 1962. Morfonologicˇeskie nabljudenija nad zvukom [cˇ] v pskovskix govorax. Pskovskie govory I: 26–58. Pskov. Gomonov, I. T. 1962. Predlogi v pskovskix narodnyx govorax. Pskovskie govory I. Trudy pervoj pskovskoj konferencii 1960 goda, 77–95. Pskov. Iwaniec, E. 1977. Z dziejów staroobrze˛dowców na ziemiach polskich. Warszawa. Kapterev, N. F. 1996 (1912). Patriarx Nikon i car’ Aleksej Mixajlovicˇ. T. I–II. Moskva: Izdatel’stvo Spaso-Preobraženskogo Valaamskogo monastyrja (Rotaprintnoe vosproizvedenie izdanija: Sergiev Posad, 1912). Karmakova, O. E. 1991. Izucˇenie kombinatoriki v nazvanijax cepa i jego cˇastej. In Sovremennye russkie govory, 148–163. Moskva. Karskij, E. F. 1903. Belarusy. T. 1. Sankt-Peterburg.
Russian varieties: Rural dialects
Karskij, E. F. 1917. Ètnograficˇeskaja karta belorusskogo plemeni. In Trudy Komissii po izucˇeniju plemennogo sostava naselenija Rossii. Vyp. 2, izd. 2. 1918. Petrograd. Kirillov, I. A. 1913. Statistika staroobrjadcˇestva. Moskva: Izd. Žurnala “Staroobrjadcˇeskaja mysl’”. Koreckij, P. I. 1863. O raskol’nikax v Kovenskoj gubernii. In Pamjatnaja knižka Kovenskoj gubernii na 1863 g. Otdel II, 1–45. Kovno. Korzon, T. 1897. Wewne˛trzne dzieje Polski za Stanisława Augusta. T. 1. Kraków/Warszawa. Kovenskaja gubernija. 1905. Pervaja vseobšcˇaja perepis’ naselenija Rossijskoj Imperii, 1897. T. XVII. Sankt-Peterburg. Kurljanskaja gubernija. 1905. Pervaja vseobšcˇaja perepis’ naselenija Rossijskoj Imperii, 1897. T. XIX. Sankt-Peterburg. Kuz’mina, I.B. 1975. O meste socˇetanij tipa (jest’) postavlen, byl (budet) postavlen; (jest’) vstavši, byl (budet) vstavši v grammaticˇeskoj sisteme sovremennyx russkix govorov. In Obšcˇeslavjanskij lingvisticˇeskij atlas. Issledovanija i materialy, 1973: 219–258. Moskva: Nauka. Kuz’mina, I. B. 1982. Nekotorye itogi istoricˇeskogo izucˇenija russkix pricˇastij. In Obšcˇeslavjanskij lingvisticˇeskij atlas. Issledovanija i materialy, 1980: 263–274. Moskva: Nauka. Kuz’mina, I.B. 1974. O vozmožnosti vzaimodejstvija sintaksicˇeskix sistem govorov raznyx jazykov na funkcional’nom urovne (postanovka voprosa). In Dialektologicˇeskij sbornik, 1974:86–92. Vilnius. Kuz’mina, I. B. 1985. O nekotoryx sintaksicˇeskix javlenijax vostocˇnoslavjanskix jazykov (po materialam OLA). Obšcˇeslavjanskij lingvisticˇeskij atlas. Materialy i issledovanija, 1982. Moskva: Nauka. Kuz’mina, I. B. 1993. Sintaksis russkix govorov v lingvogeograficˇeskom aspekte. Moskva: Nauka. Kuz’mina, I. B., Nemcˇenko E.V. 1971. Sintaksis pricˇastnyx form v russkix govorax. Moskva: Nauka. Kuz’mina, I. B., Nemcˇenko, E. V. 1964. O razlicˇitel’nyx javlenijax russkix govorov v oblasti predložnyx slovosocˇetanij. Izv AN SSSR OLJA t. 23, vyp. 4, 324–325. Kuz’mina, I. B., Nemcˇenko, E. V. 1976. Glagol’nye konstrukcii s subjektivnym genetivom v sovremennyx russkix govorax. In Obšcˇeslavjanskij lingvisticˇeskij atlas, 1976: 214–238. Moskva: Nauka. Laucˇjute, Ju. 1982. Slovar’ baltizmov v slavjanskix jazykax. Leningrad. LKA I. 1977. Lietuviu˛ kalbos atlasas. I. Leksika. Vilnius: “Mokslas”. Lifljandskaja gubernija. 1905. Pervaja vseobscˇaja perepis’ naselenija Rossijskoj Imperii, 1897 g. T. XXI. Sankt-Peterburg. Lënngren, Tamara. 1994. Leksika russkix staroobrjadcˇeskix govorov (na materiale, sobrannom v Latgalii i na Žitomiršcˇine) (Acta Universitatis Upsaliensis. Studia Slavica Upsaliensia 34). Uppsala. Marcˇenko, E. Z. 1970. Zaimstvovannye slova v russkix starožil’cˇeskix govorax Litvy (po materialam russkogo govora Zarasaiskogo rajona Litovskoj SSR). In Trudy 1970: 155–174. Markelov, G. V. 1990. Deguckij letopisec. Drevlexranilišcˇe Puškinskogo doma. Materialy i issledovanija, 166–179. Leningrad: Nauka. Materialy konferencii. 1963. Russkie govory v Pribaltike (sbornik statej). (Lietuvos TSR Aukštu˛ju˛ mokyklu˛ darbai. Kalbotyra, VIII.) Vilnius. Materialy. 1963. Nemcˇenko V. N., Sinica A. I., Murnikova T. F. Materialy dlja slovarja russkix starožil’cˇeskix govorov Pribaltiki. Pod redakciej M. F. Semenovoj (Ucˇenye zapiski Latvijskij gosudarstvennyj universitet. Ucˇ enye zapiski 51, vypusk 8 A. Posvjašcˇ aetsja V Meždunarodnomu s”ezdu slavistov), Riga. Milovidov, V. 1979. Sovremennoe staroobrjadcˇestvo. Moskva: Nauka. Murnikova, T. 1970. Ob izucˇenii russkix govorov v Èstonii. In Trudy 1970: 3–12. Murnikova T. 1968. Ob izucˇ enii russkix govorov v Èstonii. In Trudy pribaltijskoj dialektologicˇeskoj konferencii 1968 g, 3–12. Tartu.
133
134 Valeriy Cˇekmonas
Murnikova T. 19681. Russkie fol’klornye i dialektologicˇeskie materialy v Literaturnom muzee g. Tartu. In Trudy 1970: 227–237. Murnikova, T. F. 1960. Russkie govory v Èstonii. Latvijas Valsts Universiteta¯tes zina¯tniskie Raksti XXXVI sejums, 6 A laidiens: 44–52. Murnikova T. F. 1962. Opisanie russkogo govora ostrova Pijrisaara. In Trudy po russkoj i slavjanskoj filologii, 345–363. Tartu (Tartu Riikliku ülikooli toimetised. Vihik 119), Tartu. M’urkxejn, V. V. 1968. Nabljudenie nad sistemoj sklonenija imen sušcˇestvitel’nyx v odnom iz russkix govorov Èstonskoj SSR. In Trudy 1970: 108–119. M’urkxejn, Vera. 1971. Fonetiko-fonologicˇ eskoe i morfologicˇeskoe opisanie russkogo starožil’cˇeskogo govora Mexikoorma Èstonskoj SSR. Avtoreferat KD. Moskva. M’urkxejn, Vera. 1973. Èstonskie leksiceskie zaimstvovanija v odnom iz russkix govorov Èstonskoj SSR. Sovetskoe finno-ugrovedenie IX, N 1, 1–9. Nemceva, L. I. 1968. Slova-dublety v starožil’cˇeskix govorax na territorii Latvijskoj SSR. Dialektologicˇeskij sbornik 1968: 82–95. Nemceva, L. I., El’sberg, I. Ja. 1970. Kratkij fonetiko-fonologicˇeskij ocˇerk odnogo iz russkix starožil’cˇeskix govorov Latvijskoj SSR. In Trudy 1970: 34–46. Tartu. Nemcˇenko, E. V. 1974. Iz nabljudenij nad sintaksisom russkogo govora o. Mexikoorma Tartusskogo r-na Èstonskoj SSR (k voprosu o pronicajemosti sintaksicˇeskoj sistemy govora pri vozmožnosti inojazycˇnogo vlijanija). Dialektologicˇeskij sbornik 1974: 125–132. Nemcˇenko, V. N. 1958. Foneticˇeskie i morfologicˇeskie osobennosti govora russkogo naselenija Jonavskogo rajona Litovskoj SSR. Vilniaus Valstybinio Universiteto Mokslo darbai XXVI. Kalbotyra I: 147–173. Vilnius. Nemcˇenko, V. N. 1963. Russkie starožily Litvy i ix govory. Vilniaus Valstybinio Universiteto Mokslo darbai. Kalbotyra VII: 66–86. Nemcˇenko, V. N. 19681. Dialektnaja leksika russkix starožil’cˇeskix govorov Litvy, obšcˇaja s leksikoj litovskogo jazyka (k voprosu o kornevyx dialektizmax). Dialektologicˇeskij sbornik 1968: 36–72. Nikiforovskij, N. Ja. 1895. Ocˇerki prostonarodnogo žitja-bytja Vitebskoj gubernii. Sank-Petersburg. Novgorodov, M. A. 1958. Iz nabljudenij nad fonetikoj i morfologiej russkogo starožil’cˇeskogo govora Dagdskogo r-na Latvijskoj SSR. Ucˇenye zapiski Daugavpilsskogo pedagogicˇeskogo instituta N1: 351–368. Daugavpils. Novgorodov, M. A. 19581. Iz nabljudenij nad sintaksisom govora russkogo starožil’cˇeskogo naselenija Dagdskogo r-na Latvijskoj SSR. Ucˇenye zapiski Daugavpilsskogo pedagogicˇeskogo instituta N1: 291–334. Daugavpils. Novgorodov, M.A. 19582. Govor russkogo starožil’cˇeskogo naselenija Dagdskogo rajona Latvijskoj SSR. KD, Daugavpils (after Šulene 1964: 374). Obrazovanie. 1970. Obrazovanie severnorusskogo narecˇija i srednerusskix govorov (po materialam linvisticˇeskoj geografii). Avtorskij kollektiv: K. F. Zaxarova, V. G. Orlova, A. I. Sologub, T. Ju. Stroganova. Nauka: Moskva. Potašenko, G. 1993. Iš sentykiu˛ Lietuvoje istorijos (1795–1915). Vilnius (typescript). Pozdeeva I. 1995. Russkoe staroobrjadcˇestvo i Moskva v nacˇale XX v. In Mir staroobrjadcˇestva. Vyp. 2. Moskva staroobrjadcˇeskaja, 6–40. Moskva: Rossijskoe universitetskoe izdatel’stvo. Programma i tezisy mežvuzovskoj dialektologicˇeskoj konferencii po izucˇeniju russkix govorov v Pribaltijskix respublikax (1–3 okt., 1962). Vilnius, 1962. Proxorova, S. M. 1991. Sintaksis perexodnoj russko-belorusskoj zony: areal’no-tipologicˇeskoe issledovanie. Minsk: Izd. Universitetskoe. Prozorov, I. 1933. Istorija staroobrjadcˇestva. Kaunas (rotaprint). Prugavin, S. 1904. Staroobrjadcˇestvo vo vtoroj polovine 19 v.: Ocˇerki iz novejšej istorii raskola. Moskva.
Russian varieties: Rural dialects
Pskovskij oblastnoj slovar’. 1994. Vypusk 10. Izdatel’stvo Sankt-Peterburgskogo universiteta: Sankt-Peterburg. Rekena, A. 1962. Govor Kalupe. Avtoreferat kand. dissertacii. Riga. Rek¸e¯na, A. 1975. Amatniecı¯bas leksika daža¯s Latgales dienvidu izloksne¯s un ta¯s sakari ar atlilstojašajem nosaukumiem sla¯vu valoda¯s. Rı¯ga¯: Zina¯tne. Rixter, E. V. 1970. Zametki o terminax sel’skoxozjajstvennyx orudij i žilišcˇa u russkix zapadnogo Pricˇudja. In Trudy 1970: 175–187. Russkaja dialektologija. 1965. Pod red. R. I. Avanesova i V. G. Orlovoj. Vtoroe izdanie. Moskva: Nauka. Sbornik…. 1886. Sbornik pravitel’stvennyx rasporjaženij po vodvoreniju russkix zemlevladel’cev v Severo-Zapadnom kraje. Izd. 2, po 20.IX.1885 g. Vil’na. Semenova, M. F. 1959. Po povodu dvux foneticˇeskix javlenij russkix i latyšskix govorov Latgale. In Ra¯kstu krajums velı¯jums akad. prof. dr. Ja¯nim Endelı¯nam. Rı¯ga¯. Semenova, M. F. 1964. Slavjano-latyšskie etnolingvisticˇeskie otnošenija. Ucˇenye zapiski Daugavpilsskogo pedagogicˇeskogo instituta t. 10, vypusk 6: 33–38. Semenova, M. F. 19641. Russkie govory Latvijskoj SSR. Latvijas PSR Zina¯tn¸u Akade¯mija. Valodas un literatu¯ras institu¯ta Raksti VIII: 293–296. Semenova, M. F. 1968. O regional’nom atlase russkix starožil’cˇeskix govorov Pribaltiki. Dialektologicˇeskij sbornik 1968: 5–13. Semenova, M. F. 1972. O russkix starožil’cˇeskix govorax Latgalii. In Russkij fol’klor v Latvii. Sost. I. D. Fridruk, 5–28. Riga. Šulene O. 1963. K voprosu o bespredložnyx socˇetanijax v russkom govore Zarasajskogo rajona Litovskoj SSR. Materialy konferencii 1962: 15–36. Šulene, O. 19631. Materialy dlja bibliografii po russkim govoram v Pribaltike. Materialy konferencii 1963: 87–96. Šulene, O. 1965. Sintaksis russkogo govora Zarasajskogo rajona Litovskoj SSR. Kand. diss.; avtoreferat kand. diss. Vilnius. Šulene, O. N. 1968. Nekotorye osobennosti otricatel’ nyx predloženij v russkix govorax Pribaltiki (po materialam russkogo govora Zarasajskogo rajona Litovskoj SSR). Dialektologicˇeskij sbornik 1968: 117–130. Sinica, A. I. 1963. Sel’skoxozjajstvennaja leksika v govore russkogo starožil’cˇeskogo naselenija Prejl’skogo rajona Latvijskoj SSR. Kalbotyra VIII: 45–70. Vilnius. Sinica, A. I. 1968. O slovare Vyšek i Moskovskoj. Dialektologicˇeskij sbornik 1968: 24–35. Sinica, A. I. 1964. Leksika govora russkogo starožil’cˇeskogo naselenija prejl’skogo rajona Latvijskoj SSR. Kand diss. Moskva (after Semenova 1964: 37). Sinica, A. I. 1970. Leksika, oboznacˇajušcˇaja sostojanie cˇeloveka, v govore “Vyški” Daugavpiskogo rajona Latv. SSR (po voprosniku dlja OLA). In Trudy 1970: 188–203. Sivickene, M. 1979. O dvujazycˇii sel’skogo russkogo naselenija Lit SSR. In Socialine˙s lingvistikos problemos (Lietuviu˛ kalbotyros klausimai XIX): 59–66. Vilnius: Mokslas. Sivickene, M. 19791. Lingvogeograficˇeskaja xarakteristika lituanizmov russkix govorov Litvy. In Aktualiosios kalbotyros problemos. Moksline˙s konferencijos pakvietimas ir teze˙s, 184–186. Vilnius. Sivickene, M. 1988. Russkie govory Litvy v svjazi s ix isxodnoj lokalizaciej. In Pskovskie govory v ix prošlom i nastojašcˇem. Mežvuzovskij sbornik naucˇnyx trudov, 126–134. Leningrad. Sivickene, M. 1969. Nekotorye voprosy izucˇenija leksicˇeskix zaimstvovanij v russkix govorax Litvy. In Voprosy teorii i istorii jazyka. Sbornik statej, posvjašcˇennyx pamjati B. A. Larina, 208–213. Leningrad: Leningradskij gosudarstvennyj universitet. Sivickene, M. K. 1968. Ob osobennostjax projavlenija nekotoryx morfologocˇeskix cˇert v russkix govorax Litvy. Dialektologicˇeskij sbornik 1968: 112–116.
135
136 Valeriy Cˇekmonas
Sivickene, M. K. 1974. Ob odnom slucˇae morfologicˇeskoj interferencii (na materiale russkolitovskogo dialektnogo kontaktirovanija). Filologicˇeskie nauki 6: 95–100. Moskva. Solov’ev, S. M. 1991. Socˇinenija. Istorija Rossii s drevnejšix vremen. Kn. 4. Tt.11–12. Moskva: Mysl’. Stankevicˇ, A. 1909. Ocˇerk vozniknovenija russkix poselenij na Litve. Vilenskij vremennik, Kn. IV. Russkie poselenija Kovenskoj gubernii. Vil’na. Trudy 1970. Trudy pribaltijskoj dialektologicˇeskoj konferencii 1968 g. Tartu. Vilenskaja gubernija. 1900. Pervaja vseobšcˇaja perepis’ naselenija Rossijskoj Imperii, 1897. T. IV. Tetradi 1–3. Sankt-Peterburg. Vysotskij, S. S. 1948. Utrata srednego roda v govorax na zapad ot Moskvy. Doklady i soobšcˇenija Instituta russkogo jazyka AN SSSR 1: Moskva-Leningrad. Xejter, X. 1974. O razvitii tipov bezudarnogo vokalizma v russkix govorax Severnogo Pricˇudja ÈSSR. Dialektologicˇeskij sbornik 1974: 152–156. Xejter, X. 1968. Dialektnyj russko-èstonskij bilingvizm na territorii Èstonskoj SSR. Dialektologicˇeskij sbornik 1968: 14–23. Xejter, X. 1970. Osobennosti razvitija odnogo russkogo govora v inojazycˇnom okruženii. In Trudy 1970: 217–226. Xejter, X. 1971. Iz istorii izucˇenija russkix govorov severo-vostocˇnoj Èstonii. Trudy po russkoj i slavjanskoj filologii XX: 207–218. Tartu. Xejter, X. 1979. O nejtralizacii fonologicˇeskoj oppozicii po tverdosti-mjagkosti v odnom russkom govore kontaktnoj zony. In Pskovskie govory. Sbornik naucˇnyx trudov, 106–110. Leningrad. Zavarina, A. A. 1969. Iz istorii naselenija staroobrjadcev v Latgalii. In Ateizm i religija. Riga. Zavarina, A. A. 1977. Iz istorii formirovanija russkogo naselenija v Latvii vo vtoroj polovine XVIII-nacˇale XIX veka. Latvijas PSR Zina¯tn¸u Akade¯mijas Ve¯stis N2 (355): 70–84. Zavarina, A. A. 1986. Russkoe naselenie vostocˇnoj Latvii vo vtoroj polovine XIX-nacˇale XX veka. Riga, Zina¯tne. Zen’kovskij, Sergej. 1995. Russkoe staroobrjadcˇestvo. Duxovnye dviženija semnadcatogo veka. Moskva: Cerkov’. Zverkovskaja, N. P. 1985. Nekotorye cˇerty russkogo starožil’cˇeskogo govora Ludzenskogo rajona Latvijskoj SSR. In Dialektografija russkogo jazyka, 167–181. Moskva: Nauka.
AUTHOR "Anne-Charlotte Rendahl"
TITLE "Swedish dialects around the Baltic Sea"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Swedish dialects around the Baltic Sea* Anne-Charlotte Rendahl
1.
Introduction
This paper deals with that part of the Scandinavian dialect continuum which is confined within the borders of Sweden, Finland and Estonia, i.e. the part of the Baltic/Nordic area where “Swedish” is (or was) spoken. The aim is to give a brief overview of the dialect groups and linguistic features encountered, with some focus on the Baltic area. I use the expression “the Baltic area” rather loosely to refer to the coastal areas of the countries involved (the eastern part of Sweden, Swedish speaking Finland in its entirety and the once Swedish speaking parts of Estonia), thus excluding the western part of the Swedish mainland.1 The presentation is built on already existing dialect descriptions. The overwhelming majority of these descriptions saw the light of day in the heyday of Swedish dialectological research from 1880–1940 (cf. Delsing 1996). Noting the dates, it is hardly surprising that it is phonetics and phonology which receive almost exclusive attention in these works. The treatment of the phonology of a dialect is often followed by a short listing of morphological paradigms, but hardly ever is anything said of the use of the forms, and even rarer are mentions of syntax and semantics. To my knowledge, there are only two comprehensive treatments to date of the syntax of a Swedish dialect: Levander (1909) on Dalecarlian as spoken in Älvdalen and Lundström (1939) on the dialects of Nyland in southern Finland. As for the lexicon, there are quite a number of word geography studies focusing on the geographical extension of e.g. different forms of a word, often with emphasis on phonological differences. Needless to say, the nature of the sources has largely influenced the contents of this paper, partly by determining what it is possible to say anything about (i.e. primarily phonology), and partly by fixing the data sometimes 50 years, sometimes a hundred years back in time. The picture drawn on these pages, then, is rather of the linguistic practice of (half) a century ago, even though a number of the dialects are still alive and well, notwithstanding the increasing pressure from standard varieties. A strong point, however, is the relative simultaneity of the descriptions, the bulk of them having been produced in the first few decades of the 1900s.
138
Anne-Charlotte Rendahl
It should also be mentioned that, as may be evident from the previous paragraphs, it is the “genuine” dialect that is the primary focus of study rather than some local or regional variety of the standard language. The paper is divided into three parts. In the first part, I give an overview of the standard divisions made of Swedish dialects and give examples of the (types of) criteria which have commonly been used in the grouping process. In this connection, I propose, if not a new grouping, at least a different emphasis concerning part of the area. On the basis of a number of features it seems to make sense to make explicit the fact that the dialects along the Swedish east coast from Uppland in the north to northern Småland in the south (approximately) show a uniformity which sets them off from the more westerly dialects (traditionally referred to as götamål). It would thus make sense to talk about Eastern and Western mainland Swedish, with Eastern Swedish centred along the Baltic coast, a feature which is of special interest in the context of the present volume. This would give five major dialect groups: Southern, Eastern, Western and Northern Swedish on the Swedish mainland,2 and Trans-Baltic3 dialects covering the different varieties of Swedish of Finland and Estonia.4 The division into Eastern and Western Swedish made here is implicit (and sometimes explicit) in much of the dialect work done to date but it has not been apparent on the terminological level. Woven into the discussion on dialect divisions are sketches of the different areas in terms of characteristic linguistic features and, in appropriate cases, outlines of patterns of contact between different dialect (and language) areas. In the context of this volume, due to their geographical relevance for the Baltic area and their contacts with neighbouring Finno-Ugrian languages, the Trans-Baltic dialects are given a section of their own (part two of the paper). The third part of the paper is devoted to a review of characteristic linguistic features of the Swedish dialect area. The selection of features is primarily determined by what linguistic descriptions exist. In the area of morpho-syntax these are, as already mentioned, regrettably few, and so the review is something of a Swiss cheese. It does not give as full a picture as would have been possible had there been an abundance of literature on grammatical issues to consider. The features discussed in this part of the paper, however, represent what I have been able to squeeze out of the existing literature as I know it.
2.
Major dialect divisions
2.1 Dialect divisions primarily according to phonological criteria Several authors have discussed the division of Swedish (and Scandinavian) dialects and proposed groupings for larger or more restricted areas (e.g. Noreen 1903;
Swedish dialects around the Baltic Sea 139
Hesselman 1905; Wessén 1969). These accounts have all largely been based on phonological criteria. In this section, we will take as our point of departure the divisions made in Wessén (1969), a standard treatment of the Swedish dialects which by and large seems to articulate a consensus view of the major dialect divisions (from a primarily phonological perspective). Wessén makes a division into six major dialect groups.5 Starting from the south, these are (with Wessén’s labels in parentheses): Southern (sydsvenska), Western (götamål), Gotlandic (gotländska), Eastern (sveamål), Northern (norrländska) and Trans-Baltic (östsvenska) (see Map 1 for approximate geographical extensions). Certain features cluster together in the core areas of a certain dialect group and set the groups off from each other, even though the boundaries between the groups as well as between single dialects are, as a rule, fuzzy, with ample linguistic variation in transition areas. Examples of the distinguishing features used by Wessén and others are given here for each proposed major dialect grouping. Southern (sydsvenska). This area comprises the provinces of Skåne (Scania) Blekinge, and (partly) Halland, all under Danish rule until the 17th century, but extends into the province of Småland, traditionally belonging to Sweden (map 2 on p. 169 lists the provinces of the Swedish-speaking area). Features which characterize the dialects south of a line that runs across Småländska höglandet (in an approximate line from Varberg-Nässjö-Kalmar; Wessén p. 16), and thus set the Southern dialects off from the more northerly situated ones, are: (1) the occurrence of a dental lateral only, in contrast to more northern dialects which also sport the flapped “thick” (kakuminal) lateral; (2) a uvular r-sound which can manifest itself as a tremulant or a fricativized sound — north of the Southern dialects the r-sound is apical;6 (3) absence of retroflex consonants (supradentaler) which are found in all other dialect groups (but not in all individual dialects); (4) a conservation of the Old Swedish (fornsvenska) suffix -n in feminine definite forms (boken book-the ‘the book’, koen cow-the ‘the cow’) and plural neuter definite forms (husen house-the ‘the houses’, taken roof-the ‘the roofs’). A Southern feature with a geographically larger extension, primarily to the west and northwest (away from the Baltic, demonstrating affinities between the Southern and the Western dialects) are the so-called secondary diphthongs, that is, diphthongs which are relatively recent innovations, in contrast to those inherited from Old Swedish (sai ‘see’, gau ‘go’); they occur in the Southern area except the southern part of Scania (Skåne), but also in northern Halland and southern Västergötland. In approximately the same area, one finds number marking on the verb (number marking is, however, generally in regress).7 Concerning the delimitation of the Swedish dialects with respect to Danish, Wessén (1969) notes that the language on Bornholm is rather close to the dialects of southern Scania, whereas these in turn are close to the Danish dialects of southern Själland.
140 Anne-Charlotte Rendahl
i
am
Finnis h
No rt he rn
Sa
DALMÅL
Eastern
TransBaltic
Western Gotlandic
Southern
Map 1. Major dialect divisions of the Swedish-speaking area.
A feature that unites the dialects of Scania and Bornholm is the retention of the vowel -a in unstressed syllables (as opposed to the reduction of the vowel to -e or the loss of it which is found in Själland and Jylland: bide ‘to bite’ etc.; the -e seems to be spreading into the southwestern part of Scania, however (Wessén 1969: 21)). Western (götamål). This group of dialects has its centre in the western part of southern Sweden, more particularly in the province of Västergötland, traditionally thought of as the core area of the Götar. The area also includes the western part of Östergötland, long a centre of influence in Swedish history. Outside the present borders of Götaland there is the province of Värmland. The Western dialects share many (phonological) features with the Southern dialects, which set these two groups off from the Eastern (and even more from the northerly dialects),8 but they also share features with the Eastern dialects of central Sweden. Features shared by Western and Southern dialects are, for example, a
Swedish dialects around the Baltic Sea
number of systematic vowel differences as compared to the central Swedish dialects9 and vowel lengthening in originally short root-syllables (veka ‘week’, speken ‘(salt-)cured’ [Standard Swedish: spicken], vesen ‘faded’) where central Swedish instead has lengthening of the consonant. Other than phonology, common features mentioned by Wessén are, among others: – –
–
the demonstrative pronouns denna and detta where Eastern dialects have den här ‘this’ and den där ‘that’; the verb bli (bli, ble or blidde, blitt ‘to become’) is more frequently used than in the Eastern dialects where the verb varda (vart, vurti) is generally used (varda extending down through Södermanland and Östergötland into the northeastern tip of Småland; cf. Lindqvist 1947: map 501); another characteristic feature is the use of the auxiliary vara ‘to be’ in certain perfect constructions, namely in connection with verbs of motion and change: han är kommen he is come ‘he has come’, han är utgången he is outgone, han är gången ut he is gone out ‘he has left’, mor är gången till kyrkan mother is gone to church-the ‘mother has gone to church’, isen är smält ice-the is melted ‘the ice has melted’.
Eastern (sveamål). The Eastern area is divided into two subgroups: a northerly Core Eastern (uppsvensk), which is the core area centring on the province of Uppland, and Southeastern (traditionally referred to as Central Swedish, mellansvenska), which is an area of transition with a mix of linguistic characteristics of Core Eastern and Western dialects. The Southeastern area includes southern and western Södermanland and Närke, and extends along the coast of Östergötland, into northeastern Småland (north of Kalmar) and northern Öland (Wessén 1969: 30). A thorough discussion is found in Hesselman 1905).10 Although the area shares its core with the traditional Svea provinces, it thus extends beyond Svealand as it is usually understood. It is striking that the Eastern area by and large coincides with the extension of the Viking and Early Middle Ages military and taxation ledung organization. Wessén notes that the Core Eastern dialects (especially the sound systems; 1969: 36) are conservative in comparison to the Western. Wessén further notes in this connection that the innovations of the more southerly dialect groups (Southern and Western) often involve a more open pronunciation of vowels, and assimilation of consonant clusters. The majority of features which Wessén (1969) mentions as characteristic of the Core Eastern dialects are more or less archaic phonological features, conserved traits from the Old Swedish period which have been lost in neighbouring areas. For example: –
the Old Swedish clusters nd and mb are conserved (binda ‘to tie’, hundar ‘dogs’, lamb ‘lamb’, cf. the Western and Southeastern binna, hunna(r), låmm);
141
142 Anne-Charlotte Rendahl
– –
–
–
the cluster ngg, however, is only retained in some peripheral parts of the traditional Eastern area (Dalecarlia (Dalarna) and northern Roslagen); the stop–like quality of g is also retained in this area in a number of consonant combinations, e.g. varg ‘wolf ’, älg ‘moose’, nögd ‘content’, högd ‘height’ (cf. the Southeastern and Western varj, älj, nöjd, höjd); retention of the Old Swedish suffixal vowels, especially i but also u and o, where Western and Southeastern dialects have changed to e (bundi vs. bunnet ‘tied’, skurin vs. skuren or skåern ‘cut’, fjäril vs. fjärel ‘butterfly’); in western Uppland, there are traces of retained short root-syllables (kortstavighet) (smula ‘crumb’, furu ‘pine wood’, skada ‘wound’, vara ‘to be’ etc. with short root vowels) carrying the ancient so-called “balanced accent” (jämviktsaksent).
Non-archaic features which are typical of the Core Eastern dialects are for example, in phonology: – –
– –
loss of initial h (äst ‘horse’, alm ‘straw’) in Roslagen but also in Eckerö (western Åland), Södertörn and in Dalecarlia;11 loss of final voiceless stop t in unstressed syllables, in (a) neuter nouns (huse house-the ‘the house’, take roof-the ‘the roof ’) and (b) the supine (hon har måla ‘she has painted’; the loss of final t also includes most of the Southeastern (mellansvenska) area and Finland). In western Uppland and Södertörn (like in Stockholm) e and ä have “…merged in an in-between sound, more closed when long, more open when short” (p.35) which means that the following words have the same root vowel in this area: leka ‘to play’, leva ‘to live’, läsa ‘to read’, hem ‘home’, beck ‘pitch’, bäck ‘creek’.
Apart from phonology, a Core Eastern (non-archaic) trait is the generalization of the oblique form of masculine and feminine nouns with so-called weak inflexion in the Old Swedish (släda ‘sledge’, backa ‘hillside’, possa ‘bag’, and less frequently stugu ‘cottage’, kaku ‘cookie’, lädu ‘barn’ vs. the Western and Southeastern släe, backe, påse, ståva, kaka, laa)12 (cf. Schagerström 1882). The adjective suffix is -ug (lidug ‘free’, stadug ‘steady’, tokug ‘crazy’, kinkug ‘fretful’) as it is in South Eastern, Northern and (many) Trans-Baltic dialects while Western dialects have -ig.13 The elongated island of Öland off the Småland coast displays a mixture of Eastern and southwestern features, where the northern part tends to be more Eastern in character. A few examples are the thick or flapped (kakuminal) l which is found only in the northern part of Öland, and uvular r which is only found in the south. An archaic feature which remains only on north Öland are the conserved consonant clusters mb, nd and ld which are also found in other parts of the Baltic area such as the Core Eastern area, Gotland and the Trans-Baltic area (Hallberg 1996).
Swedish dialects around the Baltic Sea 143
Northern. North of the Eastern area along the coast of the Gulf of Bothnia we find the Northern dialects (norrländska, nordsvenska). Further inland, the dialects of Jämtland have many similarities to Norwegian dialects, and the dialects of Härjedalen, and Särna and Idre in northern Dalecarlia, are considered rather more Norwegian than Swedish (Wessén 1969: 39; cf. Levander 1925). The Northern as well as the Trans-Baltic dialects are generally considered (Hesselman 1905; Jansson 1936; Wessén 1969) to be intimately connected with Core Eastern dialects and can be seen as a branching-out (“utgrening”, Wessén p. 51) of the Core Eastern dialect area, the Northern and Trans-Baltic dialects having their “original” centre in Uppland.14 These two groups, the Northern and the Trans-Baltic, are usually seen as peripheral dialect groups due to their geographical location in relation to the southern Swedish perspective. Wessén points out that the dialects pertaining to these northern groupings, have, in some cases, not been reached by innovations spreading from the south, and have thus preserved several features pertaining to an older stage of the language (p. 51; cf. the archaic nature of Core Eastern (in relation to the Western and Southern dialects) mentioned earlier). Features which are used to delimit the Northern area are e.g. what is called vowel balance (vokalbalans; see below) and Norrlandic softening which means that g, k and sk become fricatives (or affricates) (j, tj and sj) not only initially but also medially before a front vowel (skojin forest-the ‘the forest’, slaji ‘(has) hit’, väddjin wall-the ‘the wall’, ryddjin back-the ‘the back’, drän(d)jin farmhand-the ‘the farmhand’, län(d)je ‘(for) long’, ditje ‘ditch’, spinnrottjen spinning wheel-the ‘the spinning wheel’, kvinnfoltje women folk-the ‘the women folk’, bäntjin bench-the ‘the bench’, fisjin fish-the ‘the fish’); this feature extends over a large northerly area including the northern parts of Värmland, Västmanland and Uppland as well as Österbotten and partly Nyland in Finland (also over into Norway).15 Apart from phonology, one finds, for example, loss of the present tense ending -er in strong verbs (han bit ‘he bites’, hon spring ‘she runs’, hon sov ‘she sleeps’) and a greater frequency of adjective incorporation (see below). The main stress on the last part of compound nouns is a salient trait (fönsterbrä’de ‘window sill’, näverta’k ‘roof made of bark’, kaffepan’na ‘coffee pot’) which is found in most of Norrland and also in Uppland and Södertörn as well as in parts of Finland but not in Dalecarlia. The dialects in the upper parts of the coastal area of Norrland, especially Norrbotten, have conserved many archaic features, for example dative and accusative case marking (Wessén 1969: 43–44). Trans-Baltic.16 This group includes the Swedish dialects of Finland and Estonia. There are many resemblances between this group and the Core Eastern dialects especially regarding the geographically closest dialects in Finland as those on Åland and the Åbo area, but there are also similarities between, for example, the dialects of Österbotten in Finland and the Swedish mainland dialects of Västerbotten and Norrbotten (cf. Jansson 1936; Ahlbäck 1971).
144 Anne-Charlotte Rendahl
The characteristics that Wessén (1969) mentions as features which the TransBaltic dialects have in common with the Core Eastern dialects are mostly archaic phonological features (see above under Core Eastern for examples of such features). Innovations which are common to Core Eastern and Trans-Baltic dialects include for example, loss of the final voiceless stop t in unstressed syllables (see above under Core Eastern), and the previously mentioned generalization of the oblique form of masculine and feminine nouns, as well as the adjective suffix -ug (again, cf. above). More genuinely Trans-Baltic features are mostly archaic ones like for example: – – –
–
the Old Swedish diphthongs are conserved except in (the mainland of) Åland (stäin ‘stone’, löus ‘loose’, öy ‘island’); a higher frequency of “breaking” compared to the standard language (bjära ‘to carry’, jäta ‘to eat’);17 g, k and sk are conserved before a front vowel in parts of Åboland and Nyland, and in Estonian Swedish (gäva ‘to give’, keldå ‘well/spring’, skinn ‘hide’, skäid ‘spoon’, skjut ‘to shoot’);18 short root syllables are “often” conserved, especially in two-syllable forms (liva ‘to live’, rutin ‘rotten’, buri ‘(has) carried’ (Nyland, Åboland, Österbotten), sonakonu ‘daughter-in-law’ (Estonia); (the examples above are from Wessén 1969: 46).
A noticeable feature of Trans-Baltic dialects is the general lack of a distinction between grave and acute accent (with the exception of western Nyland) which otherwise characterizes the Swedish dialects (more on this below). Gotlandic. The last group in Wessén’s classification comprises the dialects of the island of Gotland off the coast of southern Sweden; Gotlandic will be treated below in connection with the more detailed review of the Trans-Baltic dialects in Section 3. 2.2 Dialect groupings according to syntactic criteria In the one overview of syntactic phenomena in Swedish dialects that exists to date, Jörgensen (1970) sketches a broad three-fold division of Swedish dialects with respect to their syntactic features. The three zones (or areas) are a southern, a central and a northern zone, with fairly horizontal border lines, where the central zone is an area of overlap of a few features from the south and north and also displaying certain features which by-and-large are restricted to this central area. Jörgensen does not relate the Swedish dialects of Finland and Estonia to this grouping, but treats them as two distinct groups.19 Jörgensen bases his overview on extant dialect descriptions, and has been criticized by for example Delsing (1996: 26) for adopting an uncritical attitude to his sources, and for basing his general statements on a material which is in some cases
Swedish dialects around the Baltic Sea
too thin. For example, very few features are given from the province of Östergötland, in spite of this area’s crucial role for dialect grouping of the eastern central part of Sweden (see further below). Jörgensen himself, however, is careful to point out that his work does not encompass all dialect features which are “geographically restricted”, and that the exact extension of the features which he does go into is “very uncertain” (1970: 9). As a rule, Jörgensen (1970) provides rather fragmentary listings of features, often idiom-type expressions, as characteristic of a certain area. For the southern dialect group,20 examples of the features listed are: 1. the demonstrative pronouns denne ‘this (person)’ and detta ‘this’ trigger the definite suffix on the noun (Scania); 2. no dative case marking (Scania, Blekinge); 3. the adverb där ‘there’ functions as subject in existential sentences (där kommer ett tåg ‘there comes a train’) in Scania, Blekinge, Halland — in regional standard varieties also in Småland, as well as Finland; 4. omission of the relative pronoun in a way that differs from standard Swedish, namely primarily in the following two cases: when the main verb of the main clause is vara ‘to be’ (de e där ingen ved it is there nobody knows ‘that, there is nobody [who] knows it’ (Scania), vi va tre stycken påga lå i en soffa ‘we were three boys [who] lay on a couch’ (Scania)), or when the relative clause features any of the verbs bo ‘to live, stay’, stå ‘to stand’ or komma ‘to come’ (han har en dottår bor i samma hus ‘he has a daughter [who] lives in (the) same house’ (Scania), tok en spae sto där umme vägga ‘took a shovel [that] stood there against the wall (wall-the)’ (Ejskaer 1964; also on Bornholm); 5. verbs of motion and change take vara ‘to be’ as auxiliary (han är gången he is gone ‘he has gone’; occurs in, and south of, Bohuslän, Västergötland, Närke and Södermanland, “not common” in Östergötland, however). For the central zone,21 the dialect features which Jörgensen list as characteristic of “more than one province” number only three: 1. the form te (ti) is used to mark the infinitive in Södermanland, Uppland and Gotland (where there is variation with u, however) (dei jär int gutt ti vitä ‘it is not easy to know’);22 2. double negation can be heard in Uppland and Dalecarlia (Dalecarlia: die djaro ingum ingan skodo they do nobody no harm ‘they are not doing anyone any harm’. Another form of double negation featuring the suffix -e as a negative marker apart from the ‘usual’ negative adverb occurs in Sjuhundra härad in Uppland: ja ar-ntä sitt en-e I have-not seen him-neg ‘I haven’t seen him’ guvet om e mykky kjörsbär i år-e godknows if is much cherries in year-neg ‘God knows whether there are lots of cherries this year’ (Tiselius 1902– 03: 147–148);23
145
146 Anne-Charlotte Rendahl
3. first person possessive pronouns mina ‘my (pl)’ and våra ‘ours (pl)’ can be used on their own to express de mina (‘my relatives’) and de våra (‘our relatives’) in Uppland (in Dalecarlia, only våra can be used in this way). (A rather thin basis for calling the central zone anything else than a residual area, perhaps.) For the northern dialects,24 Jörgensen lists quite a number of features as occurring in “more than one province”, for example the following: extended use of the definite suffix (see below), the frequent use of compounding (adjective incorporation) instead of a noun phrase with an adjective attribute (nyrocken new-coat-the ‘the new coat’, storhästen big-horse-the ‘the big horse’, reingolve clean-floor-the ‘the clean floor’), dative case marking (Dalecarlia, Västerbotten and Norrbotten), etc. (more or less particularistic, i.e. not wide-spread, features). Note that Dalecarlia is part of the northern area in Jörgensen’s syntax grouping (considered as part of the Eastern dialect (sveamål) in the received classification), sharing with the (other) northern dialects among other features the use of dative case marking, extended use of the definite suffix, no adjective agreement in the plural in predicative position, adjective incorporation, idiomatic uses of the verbs fara ‘to go, travel’ and taga ‘to take’ and the adjectives fri ‘free’ and rädd ‘scared’ (and other adjectives) not followed by a preposition (not only archaic features, in other words), etc. How, then, does the three-fold syntactically based division correspond to the traditional division based primarily on phonological evidence? Well, of Wessén’s groupings, the Southern (sydsvenska) area is included in Jörgensen’s southern group, as is the Western (götamål) (except perhaps for some Western features which extend into Dalsland and southern Värmland). The rather comprehensive Eastern (sveamål) group also fits rather well with the syntactically based central group where southern and northern features to some extent overlap. The special place of the Core Eastern group in relation to the Southeastern group (mellansvenska) which is characterized as more mixed also fits with Jörgensen’s maps where Uppland is completely left out of the southern area. However, the southern border of the Southeastern area (mellansvenska) extends further south, and especially southeast, than Jörgensen’s middle area. In fact, Wessén stresses that the dividing lines for many features (again, supposedly primarily phonological), and thus for Western and Eastern, frequently go from the northwest in northern Värmland southeastwards through western Västmanland or northern Västergötland and down into northern or central or even southern Småland (e.g. Wessén 1969: 73). Among those features are, for example, voiced stops (b, d, g) between vowels and in final position following a vowel (kaga ‘cookie’, mad ‘food’) and the retention of a labial glide (w) following a consonant (twätta ‘wash’, kwärn ‘mill’, swin ‘swine’) (see Wessén 1969: 21 for more examples). Other scholars have offered evidence for a linguistically connected east coast area: Lindqvist (1947) demonstrates the existence of a northwest-southeast border
Swedish dialects around the Baltic Sea 147
in a wealth of plotted dialect maps; a more recent discussion is e.g. Fries (1994; and cf. Hesselman 1936). Danell (1912), in discussing whether the dividing line follows Lake Vättern or is centred more to the east through the western part of Östergötland, similarly offers evidence in favour of a more geographically vertical dialect division in this part of Sweden. Jörgensen’s division, where eastern parts like Östergötland are grouped together with Southern and Western dialects, does not appear to be well grounded. It is not supported by the syntactic data he presents: the only examples he gives from Östergötland are a few idiomatic expressions (e.g. ihop ‘together’ in the expression ta ihop med take together with ‘get started on’) which do not carry us very far. The Southeastern Swedish dialects on the whole, in fact, seem poorly charted when it comes to grammar (except for phonology and paradigm morphology). In Wessén’s (standard) dialect grouping, then, the Eastern group would seem more important to the Baltic area than the Southern, covering a longer stretch of the coast, well into northeastern Småland (the Western group instead pushing northwest-wards). In Jörgensen’s division, based on syntactic features, the northern border of most southern features is more or less horizontal, including all of Småland and Östergötland but excluding Närke, Västmanland, Dalsland and Värmland. There is consequently no Eastern area in Jörgensen’s division, which could possibly be due to the fact that he is not particularly interested in and does not specifically list other than, at best, features of idiomatic expressions from the lower east coast (Östergötland, (north)eastern Småland and Öland).
3.
Trans-Baltic
The Trans-Baltic dialects, i.e. the varieties of Swedish spoken in Finland and Estonia, have already been referred to in the previous sections on dialect divisions. This section contains a fuller description of this dialect area, of special interest due to its location and contact with the neighbouring Finno-Ugrian languages. I also include a subsection on Gotlandic which considering its geographical location could be said to be Inter-Baltic. 3.1 Swedish in Finland The fact that Swedish is spoken in Finland in the first place is the result of colonisation which probably started sometime in the early Middle Ages, i.e. from the 12th century onwards. The Swedish speaking population is concentrated to coastal areas, one of which is part of Österbotten to the north (from Sideby to Karleby) and the other comprises a more or less contiguous area in the south involving Åland and surrounding archipelago, Åboland (Åbo and its archipelago) and Nyland (from
148 Anne-Charlotte Rendahl
Hangö in the west to Pyttis in the east. The cities within these areas have mixed populations of primarily Finnish and Swedish speakers, and this is also the case in certain of the rural areas of primarily Åboland and Nyland. The peripheral geographical location of the Finnish Swedish dialects is consistent with the fact that the Finnish Swedish dialects have conserved a considerable number of archaic traits in both grammar and lexicon.25 Contacts with Sweden and Swedish dialects. The Swedish colonizing activity in Finland was quite substantial from the Middle Ages on. Ahlbäck believes that most of eastern Sweden (Östsverige) was involved in this colonizing venture (1971: 5). Contact has been maintained, between the Swedish speaking parts of Finland and Sweden, especially Stockholm and other ports in the Core Eastern area (Uppsverige). There has been close contact across the Gulf of Bothnia above all between Österbotten and northern Västerbotten on the Swedish side (in addition to the Core Eastern area (Uppsverige) and Stockholm). Åland in particular has been in very close contact with the eastern Uppland and Södermanland dialect area, and with varieties of the Swedish standard language (Ahlbäck 1971: 5). This pattern of contacts ties in with the fact that, linguistically, Åland seems to have something of a central role, sharing characteristics with Österbotten (and parts of Åboland) on the one hand, and with Åboland and Nyland to the east on the other (this is discussed in more detail below). Ahlbäck (1971:17) notes, however, that this does not mean that Åland has been the centre from which all innovations which have reached the Finnish mainland have spread. He suggests instead that the (linguistic) patterning is the result of contacts with different areas in Sweden, primarily the Core Eastern area (Uppsverige) and Norrland (the latter especially in the case of the northern part of Österbotten). In general, the farther east you go in Åboland and Nyland, the more archaic features are preserved in the dialects. (In central Nyland, however, there has been quite a strong influence from the city language (stadsspråket) of Helsinki from the beginning and middle of the 1800s.) According to Ahlbäck, the Finnish Swedish dialects which have been least influenced by the Swedish standard language in “recent times” are the ones of eastern Nyland, (large) parts of Åboland and all of Österbotten (1971: 6). The “genuine” dialects of Österbotten are well preserved and still serve as first language for a large part of the mostly rural population, whereas there are signs of dialect levelling in the southern Swedish speaking area (cf. e.g. Loman 1977; Ivars 1996 and 1988; Holm 1972). The internal relationships of the Finnish Swedish dialects were first extensively treated by O. F. Hultman (1894) in De Östsvenska dialekterna, and have later been discussed by e.g. Jansson (1939) and Ahlbäck (1971).26 Ahlbäck reproduces Hultman’s division of the “East Swedish dialects” (“Trans-Baltic” in our terminology) into three groups:
Swedish dialects around the Baltic Sea 149
1. northern Finnish (nordfinländska) comprising Österbotten, Åland and Houtskär, southern Korpo and parts of Hitis in Åboland 2. southern Finnish (sydfinländska) comprising the rest of Åboland and all of Nyland 3. the Baltic dialects (i.e. Estonian Swedish). The border between the northern and southern Finnish groups was drawn between Houtskär and Korpo in the Åboland archipelago. In many cases there are isoglosses here between more or less locally restricted (mostly phonological) features. In reality, though, according to Ahlbäck (1971: 15–16), there are very few characteristic features which distinguish the groups, and, furthermore, several of the ones that do exist are archaic.27 Ahlbäck contends, then, that Hultman’s grouping does not reflect linguistic reality as we know it, and goes on to sketch a relationship between the dialects which better fits the extensive review of dialectal features which he gives in his article. The features are from the level of phonology, and to a minor extent morphology. In Ahlbäck’s model, Åland is a zone of overlap, so to speak, both sharing features which do not occur in the rest of the area with (primarily southern) Österbotten, and, at the same time, sharing features which do not occur in Österbotten with Åboland and Nyland. This picture is a little too neat, of course, but it seems to hold true for a large number of features. According to Ahlbäck (1971: 16), Österbotten (especially the southern part), Åland and adjoining parts of Åboland have received a considerably greater number of linguistic changes from the west (i.e. from Sweden) than have the eastern dialects. On the other hand, there are several western features which are common to Åland, Åboland and Nyland but which are unknown in Österbotten. Below, I will exemplify the kinds of data Ahlbäck gives to support his division with Åland as a kind of pivotal centre. Features which are shared between Åland (and often neighbouring parts of Åboland’s archipelago) and southern Österbotten are, for instance (1971: 16f.): – – – – –
short æ before r and T has become e: be˜r ‘berry’, sje˜T ‘seal’. (This is similar in Uppland and Södermanland.); consistent vowel lengthening in old short monosyllables; differentiation into two distinct vowel qualities (hat-hatt ‘hatred-hat’); retroflex consonants (supradentals): fors ‘stream’, bort ‘away’, horn ‘horn’;28 the vowel of the definite suffix on strongly inflected masculine nouns (hästin horse-the ‘the horse’, vedin firewood-the ‘the firewood’) is retained (whereas it is lost in northern Österbotten, eastern and central Åboland and Nyland where, in accordance with Eastern Swedish, you get båtn boat-the ‘the boat’, vidn firewood-the ‘the firewood’, with a syllabified final -n).
Of the similarities which exist between the dialects of Åland and Åboland-Nyland, Ahlbäck mentions for example the following:
150 Anne-Charlotte Rendahl
–
–
–
–
–
the interdental voiced fricative is lost in words such as viðja ‘osier’, ryðja ‘clearing’, at the same time the vowel is lengthened (examples: vi˜o ‘osier’, ry˜a ‘clearing’). This is found in Åland and western Nyland, sporadically in Kimito (similar in Uppland and Gästrikland); the strongly inflected masculine nouns’ singular nominative suffix, -er, is retained in a number of abstract nouns, time nouns and “person nouns”: kvälder ‘evening’, freder ‘peace’, bytinger ‘urchin’. The -er suffix is also used, as in Standard Swedish, as a derogatory derivational suffix: bjäbber, blasker, dröler;29 verbs of the second conjugation have assumed the supine forms of the strong verbs, ending in -i: byggi ‘(has) built’, föri ‘(has) led’ (this corresponds to forms found in the Mälar region); regarding adjective inflection, the distinction between the masculine and the feminine has been neutralized, so that rasker and rask (‘quick, fast’, ‘capable’) for instance can be feminine as well as masculine: ein rask(er) kar ‘a fast man’, ein rask(er) kvinno ‘a fast woman’; in predicative position the same rule applies: karn/kvinnon ä rask(er) man-the/woman-the is fast ‘the man/woman is fast (capable)’. The only exception to this is eastern Åland where the gender distinction is usually upheld;30 the standard Swedish adjective suffix -ig (-ug in many dialects) corresponds to the suffix -in in Åland, (east) Åboland and Nyland; for example fatin ‘poor’ girin ‘greedy’ (-in is also found in the Core Eastern dialect area and to a certain extent in Österbotten (p. 17)).
When it comes to word geography, Åland in many cases has a similar intermediate position, sometimes leaning towards (southern) Österbotten, sometimes towards Åboland and Nyland. There is a noticeable, but not overwhelming (“ändå ganska obetydligt … på det hela taget” Ahlbäck 1971: 15), sprinkling of Finnish loanwords, varying in extent, however, in different areas. The greatest number of Finnish loan words are found in eastern Nyland, Nagu and Korpo in Åboland, and in the northernmost part of Österbotten. The fewest loan words are found in Åland — here there are more or less only Finnish loans which can also be found in Sweden (e.g. hötter ‘seaweed’, Finnish hautera, also known in Nyland, Åboland, Roslagen and Gotland). Common to all Swedish dialects in Finland are, for example, loss of final -t in unstressed position (taki, tatji etc. roof-the ‘the roof ’, kalla ‘(has) called’; also in Eastern and Northern dialects) and the generalization of different forms as the old case system is dissolving: the weakly inflected masculine nouns have generalized the old nominative form as the basic form for animals and persons, for other referents the old oblique form has been generalized (this according to the same rules as apply in Core Eastern dialects; Ahlbäck 1971: 12). Examples:
Swedish dialects around the Baltic Sea
Old nom Old obl
gubbe, gåbbe, gåbb bakka, bakk
Definite: Definite:
gubben, gåbbin bakkan
‘(the) old man’ ‘(the) hillside’
Archaic features in Finnish Swedish dialects are often shared with Core Eastern dialects. Some examples are: – – – –
– –
a is retained before original nd, mb and ld, partly also before ng; the above mentioned consonant clusters are retained unassimilated; initial h- is lost before a vowel (retained only to a limited extent); g retained as a stop in words of the type arg ‘angry’, älg ‘moose’, högd ‘height’ (except on mainland Åland where Southeastern/Western (mellansvenskt-götiskt) j prevails; in högd usually j even in eastern Åland and southern Österbotten); j conserved after k and g in weak feminines (änkjo, äntju ‘widow’, dyngjo, dynju ‘dung’) — usually lost in Åland, however; short CVC roots (kortstavighet) in bisyllables as in western Uppland (except for in Åland and Åboland’s southern archipelago).
The dialects of Åland are in many respects less archaic than other Finnish Swedish dialects, as evidenced from the listing above (and below). Certain Finnish Swedish archaisms which are lacking in the Core Eastern dialects in general can be found to a varying degree in other archaic dialects, e.g. on Gotland, in Upper Dalecarlia, in Västerbotten and Norrbotten (Ahlbäck 1971: 14). Among those are (the already mentioned): – –
Old Swedish diphthongs äi, öu and öy (not preserved in Åland, however); short CVC roots (kortstavighet) in monosyllables
Syntax. The only Finnish Swedish dialect area which has been subject to a comprehensive syntactic description is Nyland (Lundström 1939), and a few features found in that area will be listed here. For example, the dialects of Nyland characteristically have: –
–
a tendency to use double negative marking, that is, the negators inga ‘no (pl)’, it, int ‘not’ are used with negating indefinite pronouns (ingen, ingan, ingjen ‘nobody’, injender a ‘neither’, ingenting, injinting ‘nothing’) and negating adverbs (alder ‘never’, ingalunda ‘no way’, inginstans ‘nowhere’): nä˜r ja va i bastån i gå˜r, så va där int ingan u˜tom ja (when I was in sauna-the yesterday, then was there not nobody but me ‘when I was in the sauna yesterday, there was nobody there but me’); då så˜ han att on int hadd ingin skogga (then saw he that she not had no shadow ‘then he saw that she did not have any shadow’); the verb förbjud ‘prohibit’ is usually constructed with a negative (or two): han förbjö˜d, än hon int sku få˜ säj för injin (he forbade, that she not would may tell for nobody ‘he forbade her to tell anyone’) (Lundström 1939: 154–155); a tendency in western Nyland to use a preposed demonstrative pronoun (instead of the preposed definite article) when the noun is determined by a
151
152
Anne-Charlotte Rendahl
–
–
possessive and an adjective: den där svart hästn min to˜ å sje˜na (that there black horse-the mine took and bolt ‘that black horse of mine took to bolting’); den där ny˜ härrn hanses ä no˜ bra˜ (that there new master-the his is probably alright ‘that new master of his is probably alright’) (Lundström 1939: 20); the present participle is almost exclusively used as an adverbial, and almost always with verbs of motion, frequently the verb komma: så kommer Låvi˜sa skuttand (‘then comes Lovisa bouncing’); anan da˜in fo˜r dåm å˜ter u˜t ri˜dandist (other day-the went they again out riding ‘(the) next day they again went out riding’) (Lundström 1939: 108); some verbs have unusual meanings/uses; for example, the verbs fara ‘go, travel’,31 gå ‘go’ and ta ‘take’ can all be used to express “ingressive aspect” (roughly “to start doing something”, Lundström 1939: 132–143): Far no bara å lägg de! tykkt on (go now just and lay you! said she ‘you just go to bed now! she said’), åm nattn så komd tjy˜var, såm sko far själa (at night-the then came thieves, who would go steal ‘at night thieves came in order to steal’); å˜t åttati˜den gå˜r ja såva åm vintern (at around-eight go I sleep in winter-the ‘around eight o’clock I go to sleep in winter’) (Western Nyland); ska vi ga˜ såva nu, klokkån ä re˜ meir som ti˜ji (shall we go sleep now, clock-the is already more as ten ‘shall we sleep now, it is already past ten’, eastern Nyland); han to˜ å fo˜r te körkan (he took and went to church-the ‘he went to church’).
Outside of Nyland, a frequently mentioned feature of Österbotten dialects is the extended use of the definite suffix (see Section 4). 3.2 Swedish in Estonia The genuine Estonian Swedish dialects are hardly spoken in Estonia anymore — very few Swedish speakers remained behind in 1944 when the majority of the then remaining Estonian Swedish population transferred to Sweden and took up residence mostly in the Stockholm and Uppland area. Therefore, what I sketch here is an outline of a language that, in all probability, is no longer spoken. The existing sources of the language are based on data collected around the turn of the century and in the first decades of the 1900s. Existing descriptions of Swedish dialects which are still spoken, as we have seen, often date from the same period in which Estonian Swedish was most intensely described (and which was the period from which most dialect descriptions date anyway, 1880–1930 more or less), so the situation is the same in this respect for Estonian Swedish as for many other Swedish dialects which are still spoken. First a few words about geography and the (socio-) linguistic situation. Swedish in Estonia was spoken along the northwestern coast of the country and on the nearby islands, as well as on the island of Runö in the Gulf of Riga. The Swedish
Swedish dialects around the Baltic Sea
communities were the result of repeated waves of immigration (this choice of words makes it sound, perhaps, as if the number of people who came were larger than it really was) starting sometime in the late 13th century (H. Lagman 1971: 16–17) and, as in Finland, continuing to a varying extent through the centuries. The dialect areas which existed in the years immediately before the transfer of the population to Sweden in 1944 are given by H. Lagman as the following (1971: 20; Lagman cites Estlands svenskar 25 år i Sverige 1968 for the figures):32 1. 2. 3. 4. 5.
Nuckö (Rickul, Sutlep and Pasklep) 3335 Swedes. Ormsö 2203. The Rågö islands, Vippal and Korkis 970. Runö 311. Nargö 179.
The Nuckö area and Ormsö together are quite close linguistically and Lagman refers to them as core dialects as opposed to the remaining varieties which he calls peripheral. The latter are usually more influenced by Estonian due to more intense contact and a larger extent of bilingualism. Within the core dialects the contact situation also varies, however, so that Rickul on the west coast, the northernmost of the Nuckö communities, was apparently completely Swedish speaking, whereas the remaining communities Pasklep (on the Nuckö peninsula) and Sutlep (on the mainland just next to the peninsula) held mixed populations of Swedish and Estonian ethnicity. Lagman notes that, in the two mixed parishes (kommuner) of the Nuckö dialect area, Sutlep and Pasklep, the attitudes towards the Estonian language (and Estonians) were nevertheless very different: In Pasklep (on the peninsula) there was a fierce resistance to Estonian influence and conserving the Swedish was considered very important, whereas in Sutlep (on the mainland inside the peninsula) the Swedish language was rated very low by its own speakers who according to Lagman were “ashamed to speak Swedish in the presence of Estonians” (1971:21). According to Danell (1905–34), the Estonian linguistic influence in the Nuckö dialect area was strongest in Sutlep, particularly in the lexicon, but also in the sound system (p. 16). The Swedish spoken on the island of Dagö has, of course, its special history. In 1781, the majority of the population of this island migrated to Gammalsvenskby in the Ukraine where they and their descendants lived, continuing to speak Dagö Swedish in close contact with surrounding languages like Russian and German, until 1929, when most of them took up residence in Sweden. Below is a brief review of linguistic features which are found in Estonian Swedish. When nothing else is said, it is the Nuckö (-Ormsö) variety that is referred to, and the major source is Danell (1905–1934). Phonology. The Old Swedish diphthongs have been reduced to two, ai and au (Nuckö; other phonetic variants like äi and äu can be found in other varieties
153
154
Anne-Charlotte Rendahl
(Tiberg 1962: 32)) — stain (masc) ‘stone’, ai (fem) ‘island’, au (fem) ‘eye’. There are three laterals: the voiced dental lateral, a voiceless lateral (primarily in sl sequences) and the “thick”, flapped (kakuminal) l-sound. I have not found statements about the rules of distribution between the variants except for the voiceless l which corresponds to the Old Swedish sl, and tl after a short vowel (Danell). The consonants l, m and n can be syllabified (stavelsebildande), in suffixes where the vowel has been lost (båtn boat-the ‘the boat’, vattn ‘water’). Voiceless stops are unaspirated (Danell). (This would appear to be a candidate for influence from Estonian.) Short root syllables (kortstavighet) (short vowel+short consonant) are usually retained before a following vowel. The vowels i and u have usually here been conserved as closed vowels (gaTe (supine) ‘crowed’, vita ‘to know’, skuro (fem) ‘cut, slit’). The grave accent is lacking in Estonian Swedish, one finds only the acute accent (cf. below on pitch accents). Primary stress falls generally on the first syllable (cf. Danell p. 57, where a number of more or less cryptic statements can be found). The old consonant clusters mb, ld, nd as well as ngg, lg and rg have been retained (dromb ‘drum’, ‘bridge’; lamb ‘lamb’; lå˜mb ‘lame’; bTind ‘blind’; dingge ‘dung’; sårg ‘sorrow’). The stops g and k, as well as the cluster sk, have been retained before a front vowel in all positions (göra ‘to do, make’; skära ‘to cut’). (There is a certain variation, e.g. in Gutanäs k becomes t.) Danell’s summary of archaic features in phonology include the following: short root syllables (kortstavighet) (see above); diphthongs more or less conserved; one rising diphthong conserved (io˜, snio˜ ‘snow’, sio ‘lake’); g, k and sk conserved as stops; ld, mb, nd, ngg not assimilated; g conserved as a stop in the clusters rg and Tg. Among phonological features which have developed independently in Estonia are, for example, the loss of the grave accent (if one accepts this account, see below on pitch accents), truncation of an “unprotected” suffixal vowel after a long or unstressed syllable (apocope: he˜n (fem) ‘hen’, bakk (masc) ‘hillside’, kast ‘to throw’, arbet ‘to work’), and delabialization of the vowels y and ö and the diphthong öy, for example ni˜ ‘new’, be˜kor (fem.pl) ‘books’, ai (fem) ‘island’. Syntax and morphology. As in Northern dialects and the dialects of Österbotten, the definite form of nouns has a wider use than in central and southern Swedish dialects (see below on definite marking). There is one phenomenon which probably helps to enhance the frequency of the definite suffix in the dialect, and this is the special adjective incorporation construction “instead of” a noun phrase with an adjective attribute e.g. hvit-öken whitehorse-the ‘a white horse’, röd-köttet red-meat-the ‘red, lean meat’, also found in Northern and Österbotten dialects. This compound construction seems to require
Swedish dialects around the Baltic Sea
the definite suffix (even in indefinite contexts) (examples from Tiberg 1962: 98). Tiberg remarks that the genitive is “used more often” (than in standard Swedish, 1962: 98) and exemplifies with constructions such as ökens rumpan horsethe-gen tail-the ‘the horsetail’ which in this case, at least, is a part/whole relation and which in standard Swedish is usually expressed by means of a compound construction — nothing is said of the semantics or functions of these kinds of genitives (not found in the standard language), however, only that they occur more frequently and are different. The present participle is often used after te ‘to’ where the standard language has an infinitive: han e te itandor ‘he has sat down to eat’, han e te gråtand he is to crying ‘he is about to start crying’ (Danell p. 27). According to Danell, this construction conveys inchoative meaning, but in his examples one can see other uses which do not express inchoative meaning, such as tä håva inggetingg te itand they have nothing to eating ‘they have nothing to eat’. The infinitive is favoured in certain contexts. In connection with verbs of motion in idiom-like phrases, one can find the bare infinitive, for example kom ita come eat ‘come and eat’, gå tåva tuflana opp go take potatoes-the up ‘go and take up the potatoes’, conveying finality (“final betydelse”) (Jörgensen 1970: 38 citing E. Lagman 1958). Jörgensen (1970) citing E. Lagman also notes that the infinitive can occur as an object in phrases such as han ha fåi smörjas he has got grease-pass.inf ‘he has been greased’ (cf. below on the passive), or as Danell puts it “function as an easily accessible verbal noun” (p. 29f, 239): migganas tal o aikens ita women-the-gen speak and horse-the-gen eat ‘the women’s speaking and the horse’s eating’. Note also an example like: hon är opa molk she is on milk ‘she’s doing the milking’. The passive voice, as expressed by the suffixes -s (present) and -st (past), “appears to be considerably more frequent than in natural Swedish standard language”, says Danell (p. 250). Apart from the standard passive suffixes, a periphrastic passive has developed consisting either of the auxiliaries hava ‘to have’ or få ‘to get’ and the passive present infinitive (nothing is said about what governs the choice between hava and få). Examples: vil du håva bo‰a˘as häldor kambas? will you have brush-pass.inf or comb-pass.inf ‘do you want to have your hair brushed or combed?’, ia fikk bo‰a˘ as o kambas I got brush-pass.inf and comb-pass.inf ‘I was brushed and combed’, va˜gon ha få˜e smörias carriage-the has got grease-pass.inf ‘the carriage has been greased’, ia fikk kTämmas I got crush-pass.inf ‘I was crushed (in the crowd)’. Compare foAke kTä˜mst people-the crowded ‘people were jostling one another’ and ia ha bTue kTämdor ätost ti˜en I have become crushed/squeezed lately ‘I have been losing my breath lately’ (all from Danell p. 250). According to Holm (1952), the Estonian Swedish passive is unusual in many ways (pp. 400–401). Modal auxiliaries, for example the modal vill, take the s-passive ending and not the main verb, as in Holm’s slightly curious example: Människans fel villes inte nämna man-the-gen fault wanted-pass not mention ‘The man’s fault one
155
156 Anne-Charlotte Rendahl
did not want to mention’ (in standard Swedish: ‘Människans fel ville inte nämnas’). The corresponding sentence in Estonian has an impersonal passive construction involving the modal auxiliary: Inimese vigu ei tahetud nimetada (‘man-gen faultprtv.pl.neg wanted mention’). This Estonian construction is strikingly similar to the Estonian Swedish dialect usage (see Holm 1952: 61–62 for more examples). The general subordinating conjunction is, curiously, än (standard Swedish ‘att’). This morpheme with the same function is also found in eastern Nyland. In both places, however, än occurs in variation with the standard att, but, at least in Estonia, än is the more frequent choice (in Danell’s listing of subjunctions, the form att does not even appear, he only lists än). In the peripheral area of Vippal, the general subordinator än ‘att’ has been generalized to express final and causal meaning (H. Lagman 1971). Estonian Swedish shares the Northern feature of not letting the adjective agree with the noun when in predicative position (of the type dom e dum ‘they are stupid’; Danell, p. 264). Adjectives have preserved the old strong suffixes for nominative and accusative in the singular, but the function has changed: the former nominative endings are used in predicative position, whereas the former accusative forms are used when the adjective functions as an attribute. The paradigm looks like this (Danell):
PRED ATTR
masc
fem
neut
masc
fem
neut
starkor starkan
stark starka
starft starft
ippendor ippian
ippe ippia
ippet ippet
The weak inflection of the adjective is gone (as a result of apocope) and the bare stem functions as a weak form: stark ‘strong’, ippe ‘open’. The characteristic adjective suffix in Estonian Swedish is either -ot or -at (which has evolved out of the old -o˜tt), corresponding to standard Swedish -ig (bosot ‘chaffy’, brokat ‘motley’). Noun inflection is on the whole very reduced. The only marking a noun can get is number (i.e. plural) and the definite suffix (in addition to the infrequently occurring genitive -s mentioned above). There is, however, no plural marking on masculine and neuter (and in places feminine “på sina håll” Tiberg 1962) nouns when they are preceded by a cardinal numeral (the type ‘four bull’). Danell attributes this at least partly to influence from Estonian. The possessive pronoun can follow the noun, but I have not found any specifics about the rules for this in the descriptions (bo˜ne mitta child-the mine ‘my child’). The demonstrative pronouns (“in adjectival position”) tä or tai, tor and tom function as preposed definite article (Danell, p. 265).
Swedish dialects around the Baltic Sea
As in Nyland, double marking of negation is found quite often (“påfallande ofta”, Danell). A periphrastic construction with the verb vara ‘to be’ followed by the perfect participle is often used “instead of” ha ‘to have’ plus the supine:33 nu ä allon ite now is everyone eaten ‘now everyone has eaten’ han ä komendor OR han ha kome he is come OR he has come ‘he has come’ … om han här e vare if he here is been ‘… if he has been here’
Another unusual construction involving verbs is the habit of putting a negative marker (o-) on the (main) verb: han ä o-komendor he is un-come ‘he has not come’. (This is also found in Northern and Finnish Swedish.) Number-marking on the verb is well preserved, but person-marking on the verb is gone, all according to Danell. Tiberg (1962) notes that in the past tense, number (i.e. plural) marking on the verb is only found in the core dialects of Nuckö and Ormsö (and maybe there were tiny remnants of the system on Runö) “of the type jag var/vi voro [‘I was/we were’]” (it is however unclear what Tiberg’s “of the type” refers to). There does not seem to be anything explicitly written on sentence connection in any of my sources (except listings of connectors) but interesting things can be observed in the examples. For instance: han gik e lada ø kep grisar he went into market Ø buy pigs ‘he went to the market to buy pigs’ ia gick te skoen opa slå I went to forest-the on mow ‘I went to the forest in order to mow grass’
(and cf. the example from above hon är opa molk, she is on milk ‘she is doing the milking’). Contrasting the standard language with the Estonian Swedish dialects in these examples we see that the Swedish device för att ‘in order to’ corresponds to either zero or the preposition opa ‘on’ in Estonian Swedish. There is a closer correspondence, though, in the following example: … färe he än kri˜ana mä˜ soman skuld våra mä häilsa for it that creatures-the with summer-the would be with health ‘in order for the cattle to be healthy during summer’ (‘… för att kreaturen under sommaren skulle vara friska’).
157
158
Anne-Charlotte Rendahl
As far as the lexicon goes, there is great variation between the dialects, according to Tiberg, even in the basic vocabulary such as, for example, words for cattle etc. (Tiberg 1962). More interestingly, perhaps, Danell reports on a number of peculiar semantic developments, most notably in function words such as pronouns, conjunctions, adverbs and prepositions (including opa ‘on’, ite ‘in’) (cf. Danell, p. 28 who does not however go further into the subject than to list these (mostly function) words and their translations). 3.3 Gotlandic Geographically an inter-Baltic dialect, so to speak (or group of dialects), Gotlandic has its special history. The ancient language, Old Gotlandic, is preserved in runes and above all in the one handwritten manuscript that remains, the Gotlandslagen and Gotlandssagan (telling of the history of Gotland), dating from around 1350. Situated practically in the middle of the Baltic Sea, Gotland was an important trading centre during the Viking Age as well as throughout the Middle Ages when, for instance, the Hanseatic League was active, mediating linguistic contacts, primarily with Danish and German. Denmark had the greatest political and practical influence, and (the eastern part of) Sweden did not start playing a role in the development of Gotland until well after 1645 when Gotland was given to Sweden through the peace treaty at Brömsebro (cf. Sjöberg 1992). The Danish and German contacts have not entailed as much linguistic influence as one might perhaps assume. In fact, the relative number of German loan words, for instance, is lower in Gotlandic than in standard Swedish (Gustavson 1974). Overall, Gotlandic is characterized by a number of archaic features, especially on the phonological level, many of which are shared with dialects of the Core Eastern and Trans-Baltic area (as well as with other peripheral areas like Norrbotten and Dalecarlia). Features which Wessén (1969) notes as being shared by Gotlandic and Trans-Baltic dialects include: (1) the conservation of the Old Swedish diphthongs (bain ‘leg’, auge ‘eye’, hoyre ‘to hear’); (2) “breaking” (bjere ‘to carry’, jete ‘to eat’, sjegl ‘sail’; (3) retention of old consonant clusters (kväld ‘evening’, lamb ‘sheep’, langg ‘long’, ård ‘word’); (4) stop-quality in g, k and sk before an initial front vowel (gylte ‘sugg’, kälde ‘well, spring’, körke ‘church’, skäre ‘to cut’, skiaute ‘to shoot’); (5) stop-quality in the g of words like varg ‘wolf ’, talg ‘tallow’, slygd ‘handicraft’; and the non-archaic (6) loss of final unstressed-t (häuse house-the ‘the house’, bite ‘(has) bitten’) (as in Finland, the Core Eastern dialect area and parts of the Southeastern area; Wessén 1969: 49). In addition to the (mostly archaic) features shared with Trans-Baltic dialects (as well as, sometimes, isolated parts of the Swedish mainland), Wessén mentions a number of phonological characteristics, which mostly concern the vowels, as
Swedish dialects around the Baltic Sea 159
distinctive of Gotlandic dialects. Examples are (secondary) diphthongization of all long vowels except a (bäite ‘to bite’, bräun ‘brown’, nöy ‘new’, bröydar ‘brothers’, båuk ‘book’, såul ‘sun’; cf. the primary, inherited diphthongs aik ‘oak’, raud ‘red’, droyme ‘dream’) and the old iu which has turned into iau as in biaude ‘bid, offer’, diaup ‘deep’ (etc). A syntactic feature to add to the phonological abundance is the verb-final word order in short relative clauses: …vör sum gamblä jär we as old are ‘we who are old’ (Gustavson 1974: 26). (This is also an example of “vowel balance” (gamblä) which is found on Öland as well as Gotland.) Another feature pertaining to syntax is the scarcity of the passive -s ending: according to Gustavson (1974: 26) it is almost only used in the infinitive (kånnä ska snart skäras barley-the shall soon cut-pass.inf ‘the barley is soon to be cut’) — in other cases, the periphrastic bli passive construction is used (dei blai skur i går ‘it was cut yesterday’).
4. Overview of dialect features In section two of this paper I reviewed the criteria which form the basis for the distinction between Swedish dialects. In this section I will go deeper into these features, and also discuss some additional features, starting with morphology and syntax. Definite marking. Definiteness in Swedish dialects is expressed primarily by means of a suffix on the noun and, when the noun is preceded by attributes, a preposed article in addition to the definite suffix (both suffix and preposed article also mark gender). A feature frequently mentioned in the dialect literature is an extended use of the definite suffix in the dialects of Norrland and Finland in relation to the standard language and more southern dialects — the definite suffix in these dialects is used in contexts where the standard language (and more southern dialects) have a bare head noun (“indefinite” in the standard terminology). On the Swedish mainland this usage is reported from as far south as Uppland (Hesselman 1908; Tiselius 1902–03), Dalecarlia (Levander 1909) and Värmland (Jörgensen 1970). Hummelstedt (1934) reports this usage as characteristic of all dialects in Österbotten. In a recent study, Delsing (1996: 29–33) points out that many of the uses of the definite suffix in Northern and Finnish Swedish dialects can be characterized as “partitive” in analogy with the French partitive article: Hä finns vattne däri hinken it is water-the there-in bucket-the ‘There is water in the bucket’, Hä väks granen överallt it grows fir trees-the everywhere ‘There are fir trees growing everywhere’; he hadd vöre solskene å vakkervere it had been sunshine-the and good-weather-the ‘There had been sunshine and beautiful weather’, … som ha vure där å huggä vön [veden] … that have been there and cut wood-the ‘… who have been there and cut wood’.
160 Anne-Charlotte Rendahl
In a study of the definite suffix in the Närpes dialect of Österbotten, Nikula (1997) concludes that it is not, in fact, definiteness that is expressed by the definite suffix in the Närpes dialect, but referentiality. She formulates the following main rule for the use of the definite suffix: A noun that is not preceded by an attribute takes the definite suffix when the noun expresses a referential function. Otherwise there is no suffix (Nikula 1997: 205).34 A few of her examples illustrating the usage are: Tenn ha˜r di kväit-e there have they wheat-the ‘There they have (the) wheat’, Vi ska ha u˜t handlingg-an täi å we will have out documents-the them too ‘We will have (the) documents removed [uttagna] too’, Tu ha˜r håT-e åp böksår you have hole-the on trousers ‘You have (a) hole in your trousers’. A noun in a referring function also takes the definite suffix when preceded by a third person demonstrative pronoun even if there is an(other) attribute present that precedes the noun, e.g. His-in (sju˜k) mag-an männ kåmbär ti ta li˜ve oå˜v me, this (sick) stomach-the mine come to take life-the of me ‘This (sick) stomach of mine is going to kill me’ (‘Den här (sjuka) magen min kommer att ta livet av mig’).35 The Österbotten pattern of definite suffix marking also occurs in Northern Swedish dialects (cf. e.g. Larsson 1929: 124 for Västerbotten). For southern Finland (Nyland), Lundström (1939) reports the use of the definite suffix when the head noun is preceded by a genitive attribute or when the noun is followed by a possessive pronoun (when a possessive pronoun precedes the head noun there is no definite marking on the noun, however), which differs from standard Swedish (and southerly mainland dialects) where a bare noun occurs after the genitive attribute. (Example: nå va e Tåppasas katton probably was it Tåppas-gen cat-the ‘surely it was Tåppas’ cat’ (‘nog var det Toppas katt(en)’), Lundström 1939:16–17). The use of the definite in this way, after a genitive, also occurs in Österbotten and Northern Swedish dialects. The type of extended use of the definite suffix other than after a genitive that occurs in Northern dialects and in Österbotten, appears only to occur in Nyland when referring to (common) illnesses (boånen har jo çi˜kho˜stån children-the have as-you-know whooping cough-the ‘the children have the whooping cough’ ‘barnen har ju kikhosta(n)’, Lundström 1939: 12). Estonian Swedish also has definite marking on the head noun following a genitive attribute (prästens rocken priest-the-gen coat-the ‘the priest’s coat’, aikens airana horse-the-gen ears-the ‘the horse’s ears’) according to Danell (“as have, among others, Core Eastern dialects”, 1905–34: 227). In the strict Nuckö dialect, however, the definite suffix is not used when the noun is preceded by a possessive pronoun (here we have hans bok ‘his book’ and not *hans boken his book-the ‘his book’; just like in Nyland, incidentally). The extended use of the definite suffix occurs in roughly the same areas36 as show a strong tendency to use adjective incorporation instead of noun phrases with an adjective attribute (which would require the preposed definite article den (uter) or det (neuter)), yielding expressions such as rödköttet (red-meat-the) for the
Swedish dialects around the Baltic Sea
standard det röda köttet ‘the read meat’, hvitöken (white-horse-the) for en vit häst ‘a white horse’ (Tiberg 1962: 98).37 Case. Only traces remain of the old case system distinguishing between the nominative, genitive, accusative and dative case. The dative (and the accusative) is still found in active use in parts of Norrbotten and Västerbotten in the very north, in Jämtland and Härjedalen, and in the dialects of Upper Dalecarlia (Reinhammar 1973: 16).38 It is completely gone from the Trans-Baltic dialects of Finland and Estonia. Where it is still used, dative case marking can be assigned to nouns, adjectives, cardinal numerals, the indefinite article and to pronouns (p. 28). The marking on nouns is generally retained in the definite form only and with place names. (Examples from Dalecarlian as spoken in Älvdalen: ig far upi Övdaln I go up-in Övdal-acc ‘I’m going to Älvdalen’, ig bydd‰er i Övdalim I live in Övdal-dat ‘I live in Älvdalen’.) Reinhammar mentions that there is variation between the dative and the accusative “even in more stable dative dialects” (1973: 229), and the accusative is more common in the outskirts of dative areas than in core areas (this would perhaps suggest that dative forms are being replaced by accusative forms). Pamp (1978: 30) notes that in the “archaic” dialect of Attundaland in Uppland there is a distinction between the nominative and the accusative case (but he gives no reference to any description of this dialect). Typical for the Baltic area, then, is that even in the peripheral/archaic Swedish dialect areas (Gotland, Estonia and Finland), the dative and accusative cases have disappeared. The genitive case, as expressed by the suffix -s, is in general use. However, from many areas it is reported to occur rather sparingly and with a restricted functional use — this seems particularly true of the northern part of Sweden (actually including Bohuslän, Värmland, Dalecarlia and Uppland) and of the dialects of Finland and Estonia39 (Jörgensen (1970: 40–41).40 Where the descriptions make note of which functions the genitive is used in, it is quite common that the genitive -s suffix is in general use with proper names and expressions referring to persons but unusual with count nouns. This is reported from Bohuslän, Värmland, Dalecarlia, Uppland, Hälsingland, Medelpad and Ångermanland (Jörgensen 1970: 41). Instead of a genitive suffix construction, a construction involving a preposition or compounding is most often used (bröste på a breast-the on him ‘his breast’ (Jämtland), mössa åt pöjken hat for boy-the ‘the boy’s hat’ (Dalecarlia)). In Västerbotten, the dative case can also be used: boka farfarum book-the grandfather-dat ‘grandfather’s book’, as well as the nominative when placed after the head noun (which in turn is marked by the definite suffix: höje hästen hay-the horse-the ‘the horse’s hay’ (Jörgensen 1970: 43; I have not found any information on restrictions on its use in these contexts).
161
162 Anne-Charlotte Rendahl
Gender. Swedish dialects generally have a three gender system — feminine, masculine and neuter — expressed in the choice of pronoun (hon, han, det ‘she, he, it’) and in noun inflexion in dialects where this is retained (as well as for example in adjective agreement).41 The Swedish standard language differs in this respect from most dialects in having a two way gender distinction, uter (den) and neuter (det) (apart from the pronouns hon and han for referring to animates). There is one dialect which has simplified the system so that there is only one gender left, so to speak, for referring to inanimates. This is the variety spoken in Karleby and Nedervetil in northern Österbotten in Finland, where all nouns take the indefinite article en and the definite preposed hä ‘the’ or häde ‘that’ (nouns or proper names referring to animates trigger the masculine and feminine definite forms han (masc), hon (fem), hande (masc), honde (fem)) (Huldén 1972: 47). Huldén argues that the simplification of the system is the result of large-scale change in noun and adjective inflexion (mostly due to sound change like apocope and loss of final postvocalic -t and -n) which made the gender assignment of nouns less transparent in the inflexional morphology (1972: 78).42 Adjective agreement. An adjective generally agrees with the noun, whether it occurs as an attribute or in predicate position. In Northern dialects, however, as well as in Estonian Swedish, the adjective (as well as the perfect participle) in predicate position does not agree with the noun if this is in the plural (dom ä dum ‘they are stupid’ ‘dom är dumma’). According to Bergman (1952: 155, cited in Jörgensen 1970), this feature is found as far south as Dalecarlia and Uppland. It also occurs in Värmland (Jörgensen 1970: 25), for example: han hadd pänningar gömd ‘he had money hidden’ (standard Swedish: ‘han hade pengar gömda’). In Finland, the lack of agreement in the plural is found in Österbotten but not in Nyland (Lundström 1939). The adjective suffix -ug is found in Eastern, Northern and many Trans-Baltic dialects (a suffix -in (fatin ‘poor’, girin ‘greedy’) is found in Åland, eastern Åboland, Nyland, and also in Core Eastern and Österbotten dialects to a certain extent). Western dialects have -ig (which is also the standard variant). Verb morphology. Number and person marking on the verb (where person marking occurs it is only in the plural) is still found in a few limited areas, namely northeastern Scania, and in the adjoining Lister peninsula in Blekinge, Swenning 1917–1937) and in an area in the west encompassing Halland, southern Västergötland and western Småland (Pamp 1978), and also including Fårö, north of Gotland, which has plural-marking (Wessén 1969; Gustavson 1974).43 A distinction of person in the plural is also common in Upper Dalecarlia and in the dialects of Västerbotten and Norrbotten. In Estonian Swedish, person marking on the verb is gone but number is generally distinguished in the present tense. In the past tense, plural is only marked in the core dialects of Nuckö and Ormsö, but apparently
Swedish dialects around the Baltic Sea 163
sparingly (Tiberg 1962: 100).44 In the Finnish Swedish dialects, the verb seems to have lost number and person marking completely (gone, like the accusative and dative case marking on nouns; a residue is found, however, in the imperative mood (Ahlbäck 1971)). The infinitive marker is ti or te in most Eastern dialects, in Finnish and Estonian Swedish and in Gotlandic. Auxiliary constructions. Periphrastic perfect constructions generally involve variants of the verb hava ‘to have’ (plus the main verb in the supine). In some areas in southern and western Sweden and in the core dialects of Estonian Swedish, the auxiliary vara ‘to be’ followed by the past participle is favoured in certain contexts. In the Nuckö dialect of Estonia, the vara-construction is the most frequent one in contexts which express a change of state or place (H. Lagman (1971: 227). Some examples: han är gåle opp-stive he is already up-stood ‘he has already risen’ (‘han har redan stigit upp’), hå˜nde är opp-sullna hand-the is up-swollen ‘the hand is swollen’ (‘handen har svullnat upp’), and involving the negator o, as in: ve ära o˜ite we are un-eaten ‘we haven’t eaten’. A negative marker on the main verb (han är o-kommen he is un-come ‘he hasn’t come’) is found in Northern dialects as well as in Finland and Estonia. As a southwesterly dialect feature, Wessén (1969: 29) describes the practice of using the auxiliary vara ‘to be’ in certain perfect constructions, namely in connection with verbs of motion and change of state: han är kommen he is come ‘he has come’, han är utgången he is out-gone, han är gången ut he is gone out ‘he has gone out’, mor är gången till kyrkan mother is gone to church-the ‘mother has gone to church’, isen är smält ice-the is melted ‘the ice has melted’. Double negation is common in (at least) Nyland, Dalecarlia and Estonia. (A special kind of double negative marking is found in some dialects of the Core Eastern area, see above.) Word order features. Only a few features will be mentioned here. The first concerns the possessive pronoun which can follow the head noun (huset mitt house-the my ‘my house’) — this happens most frequently in Northern and Trans-Baltic dialects (Jörgensen 1970; Delsing 1996). The type huset mitt disappears completely south of Uppland, Västergötland and Bohuslän (Jörgensen 1970: 45), while the type involving a kinship term (mor min mother my ‘my mother’) extends down to (but not including) the Southern dialect area (Delsing 1996: 42). Stress is quite often a factor in the relative position of the possessive pronoun: in many dialects, the pronoun precedes the noun when stressed, but follows it when unstressed. This is the case in the following areas: Västergötland, Värmland, Dalecarlia, Medelpad, Västerbotten, Norrbotten, Nyland och Österbotten (Jörgensen 1970: 45). The possessive pronoun always follows the head in Ångermanland and Västerbotten (Jörgensen, p. 45).
164 Anne-Charlotte Rendahl
Word order in main and embedded clauses in standard Swedish differs in how the sentence adverbial occurs in relation to the finite verb: in main clauses the adverb follows the finite verb whereas in embedded clauses, it precedes the finite verb. In Norrland, however, as far as embedded clauses with the general subordinator att ‘that’ are concerned, at least, main clause word order is also embedded clause word order: så fådräs e jo att ve skull ätt båra så sera then is-demanded it asyou-know that we would not only sow grain-the ‘then it is demanded, that we not only should sow the grain’ (‘så fordras det ju, att vi inte bara skulle så säden’, Dahlstedt-Ågren (1980: 279, cited in Jörgensen 1970: 27). This practice is also found in colloquial standard Swedish. The word order of adverbials can differ between the dialects (in addition to the Northern att-clause practice just mentioned). In southern Sweden, stressed verb particles often occur after the object in examples such as jag har brutit den av ‘I have broken it off’ (‘jag har brutit av den’, Lombard cited in Jörgensen 1970: 14) and vör skall släpp kräki äut mårgu ‘we will let the cattle out tomorrow’ (‘vi skall släppa ut kreaturen i morgon’, the last example from Gotland from Gustavson 1974: 25, cited in Jörgensen, p. 14). Common in Scania is “circumposition” of adverbs (split adverbs), in the dialects as well as in the local form of standard language (Jörgensen 1970: 15, example from Bergman) här är varmt inne ‘here is warm inside’ (standard Swedish: ‘det är varmt härinne’). It may be noted in passing that the adverb där ‘there’ often functions as formal subject in southern Sweden (see Section 2). Geocentric space. In the many examples in Danell (1905–34) there is some evidence of a so-called geocentric spatial system being used (cf. Levinson 1992) in Estonian Swedish alongside a more relative one. There seems to be frequent use of the geographically fixed cardinal directions (north, east, west, south) besides the relative “right” and “left” and other spatial terms in order to locate objects and people relative to a place. In the example (already used above) from northern Rickul, gå ut åt no˜da-si˜a fosta bak-före kåstains-fo˜tn go out towards north-side porch behind chimney-the ‘go out towards the north side of the porch behind the chimney’ (standard Swedish: ‘gå ut åt nord-sidan om förstun bakom skorstensfoten’), a place next to the porch is located by means of the cardinal direction no˜da-si˜a ‘north side’ and in the same example another spatial term, bak-före ‘behind’, is used to complete the spatial description.45 The temporal connectors då and när show a variational pattern (studied by Wallmark 1936) which can be summarized in the following way: när is the most frequently used connector in southern and central (“mellersta”) Sweden, on Åland and Nyland, while då is used in northern Sweden, Österbotten and Estonian Swedish (Jörgensen 1970: 50). (There is also variation between the forms within the areas, i.e. in many cases both forms occur to some extent.)
Swedish dialects around the Baltic Sea 165
A feature concerning preposition constructions which ties the Northern and the Finnish Swedish dialects together is that a preposition is omitted after the adjectives fri (‘free’) and rädd (‘scared’) (and maybe one or two other adjectives in Northern dialects). In Finland the preposition is also omitted after certain verbs (Jörgensen 1970: 48). As for phonology, a number of features have already been mentioned as characteristic of certain areas in the previous sections of this paper. Below, I will go somewhat deeper into a few of these. Diphthongs. Primary diphthongs. The three diphthongs of Old Swedish (au, öy and ei) have all been monophthongized (to variants of long ö, ö and e respectively) in all but a few geographically peripheral areas, namely dialects of Norrbotten, Västerbotten and Jämtland in northern Sweden (conserved to different extents, DahlstedtÅgren 1980: Map 5), the “Norwegian dialects” of northern Dalecarlia (Särna and Idre), Gotland, Estonia, and Österbotten in Finland. The southern Swedishspeaking part of Finland (Åboland and Nyland) now shows tendencies towards monophthongization. Secondary diphthongs. Diphthongs which do not stem from Old Swedish but which are later innovations are usually referred to as secondary diphthongs. These occur mainly in a contiguous area in southern Sweden. Southern Scania is exempted from this area but it includes the northern part of Scania, Blekinge, southwestern Småland, Halland and the very southernmost part of Västergötland (Wessén 1969: 20). The island of Gotland also has secondary diphthongs which makes Gotland excessively rich in this particular respect since Gotland also has retained diphthongs from Old Swedish (Gustavson 1974). Other areas with various kinds of non-inherited diphthongs outside the southern dialect area are e.g. Södermanland, Österbotten and Upper Dalecarlia (particularly Älvdalen). Syllable length. A few features relating to syllable length in Old Swedish which have had implications for the dialects will be mentioned here. There was supposedly a four-way quantity contrast in root syllables in Old Swedish: V:C, VC:, V:C: and VC (Pamp 1978; Wessén 1958). The last of these, the short root syllables (VC), have either been subject to vowel or consonant lengthening, or have retained their short character: – Retained short root syllables (kortstavighet) are found in many of the peripheral/archaic areas: Nyland, eastern Åboland and Österbotten in Finland, Estonia, part of the Northern dialect area (Norrbotten, parts of Jämtland and Ångermanland), Upper Dalecarlia and in the western part of Uppland (e.g. Uppland: smula ‘crumb’, furu ‘went’, vara ‘to be’, spada ‘shovel’, carrying the “balanced accent” (jämviktsaccent); Wessén 1969).46
166 Anne-Charlotte Rendahl
– Syllable lengthening: vowels vs. consonants. Vowel lengthening occurred mainly in the south: in Southern and Western dialects, as well as in the dialects of Gotland and parts of Roslagen and Södermanland (Ahlbäck 1971:10). North of this, consonant lengthening is the general rule except for the areas where short root syllables are mostly retained (the borders are tricky, cf. e.g. Pamp 1978). Grouped together in this respect are Åland, the archipelago of Åboland and the Core Eastern area where sometimes the vowel, sometimes the consonant is lengthened (la˜kan or lakkan ‘sheet’, sella ‘harness’, krå˜sa ‘to crush’; Ahlbäck 1971: 10). A northern Swedish feature is the so-called “vowel balance” (vokalbalans) which is related to the long syllables of Old Swedish — in such words the suffixal vowel is weakened (its quality changes) and yields pronunciations such as stintä (stinte) ‘girl’, kastä (kaste) ‘to throw’, bitä (bite) ‘to bite’. In Jämtland and Västerbotten, the last vowel is usually lost altogether (stint ‘girl’, kast ‘to throw’, bit ‘to bite’, gubb ‘old man’).47 As already noted, this is a northern feature — the southerly border of the vowel balance area along the coast is between the dialects of Medelpad and Hälsingland (Hälsingland has -a throughout: kasta ‘to throw’, springa ‘to run’). Apart from the Northern dialects, vowel balance is found in Trans-Baltic dialects (in Finland primarily in Nyland and Åboland (Ahlbäck 1971)) and not least in Dalecarlia (Wessén 1969: 52); it is also found, in fact, as far south(east) as the islands of Öland and Gotland. Pitch accent. Swedish has two accents in two-syllable words, acute and grave (or Accent 1 and Accent 2), which are distinguished primarily by their different tonal patterns (pitch). This contrast is found in all dialects except the Swedish dialects of Finland,48 Estonia and Överkalix in Norrbotten (Gårding 1974: 56–57, Figures 13, 14; cf. Gårding 1977). In these areas, there is only the acute accent. The lack of the grave accent is usually ascribed to contact with the surrounding Finno-Ugrian majority languages, which do not have a tone accent (this goes also for Överkalix, where the Finnish (and Sami) linguistic presence is considerable). However, an alternative explanation is that the Trans-Baltic dialects were already removed from the centre of innovation (presumably the core Eastern area) by the time the accent distinction was established on the Swedish mainland (Ann-Marie Ivars, lecture notes). The pitch accent distinction seems to be operative only in words where the root syllable has at least one long segment. Words with short root syllables, in those dialects that preserve them, take the so-called “balanced accent” (jämviktsaccent). The dialects differ in the exact realisation of the accents.49 As Elert (1995) puts it: “Den genomgående skillnaden mellan akut och grav accent i olika delar av Sverige är att det finns en fördröjning av tonhöjdsfallet i ord med grav accent i förhållande till det som sker i ord med akut accent” ‘The consistent distinction between acute and grave accent in different parts of Sweden is that there is a delay of the pitch fall in words with the grave accent relative to that which is found in
Swedish dialects around the Baltic Sea 167
words with the acute accent’ (p. 131), delay of the pitch fall in words with the grave accent being the operative words here. Based on an extensive survey of pitch patterns in Swedish dialects, Gårding (1974) makes the generalisations that (1) the acute accent never has more than one pitch peak, and (2) if in a dialect the acute and grave accents both have only one pitch peak, then the peak of the grave accent is always later than that of the acute accent (Gårding 1974: 58). On the basis of the temporal pitch patterns, then, of the acute and grave accents, Gårding makes a broad grouping of the dialects into four areas (the division is not absolute, rather a tendency towards a certain accent pattern in an area). Below is an approximate reproduction of her summarizing table: Acute accent
Grave accent
Area
Type 1 Type 1a Type 1b
one peak early in the first syllable late in the first syllable
Type 2
one peak
Type 2a
late in the first syllable
one peak late in the first syllable Southern Sweden early in the second sylla- Gotland and Dalecarlia ble two peaks Eastern area & most of Norrland one peak in each syllable Western
Type 2b
in the second syllable
one peak in each syllable
The areas which are dominated by Type 1 accent patterns, that is those where the grave accent has only one pitch peak, are southern Sweden (roughly Scania, Blekinge and southern Halland) and Gotland and Dalecarlia.50 The patterns of these areas are distinguished by the temporal characteristics of the pitch peaks: in southern Sweden, the acute accent is characterized by having a peak early in the first syllable whereas the grave accent has the peak late in the (likewise) first syllable (type 1a); while in Gotland and Dalecarlia, the peak of the acute accent comes late in the first syllable whereas for the grave accent it comes early in the second syllable (type 1b). The Type 2b pattern, where the acute accent has its peak in the second syllable and the grave accent has two peaks, is primarily found in the Western dialect area, i.e. this is the dominating pattern in an area comprising Västergötland, most of Östergötland, northwestern Småland, Bohuslän, Dalsland and Värmland. The pattern that is characteristic for central Sweden (Mellansverige) (2a), where the acute accent has a peak late in the first syllable and the grave accent has two peaks (one in each syllable), extends down along the Baltic coast to southern Småland and Öland and is yet another indication of an eastern dialect area along the Swedish Baltic coast in contrast to Western dialects.51 The core area of this pattern is Södermanland (and Närke), Uppland and Gästrikland. Most of the
168 Anne-Charlotte Rendahl
dialects in Norrland also have the tonal accent pattern that is characteristic for central eastern Sweden. Despite the differences in the patterns of phonetic realisation of the accents, the assignment of a certain accent to a certain word is generally the same across the dialects. The differences which exist usually concern proper names and loan words. In central Sweden (Mellansverige), for example, disyllabic family names such as Bergström and Björklund are pronounced using the grave accent while in Norrland the acute accent is generally used with names of this type (Elert 1995: 131). Regional differences in the distribution of accents across the lexicon beyond the realm of proper names and loan words exist; for example, southern Uppland and northern Södermanland in fact generally have the grave accent instead of the acute (Ericsson 1914; Kruuse 1908; Nyström 1997). Another prosodic feature is voice quality. Elert (1995) gives a brief orientation in regional differences in voice quality. Among the things he mentions are the tense voice, often combined with a creaky voice quality, which characterizes many speakers in Småland, and a tendency towards breathy voice in the inner parts of northern Norrland. The average pitch level also tends to vary regionally. For example, it is relatively low in Norrland, tends to be higher the further south you go culminating in northern Götaland; to the south of Götaland, again, the average pitch level is generally lower, especially in Scania in the very south (Elert 1995: 133). Finally, the geographical extension of some archaic phonological features: – –
–
the consonant clusters nd, mb etc. (lamb ‘lamb’) are still found in the Core Eastern area, Upper Dalecarlia, Gotland, north Öland, Finland and Estonia; the sequences -rg, -lg, -gd (varg ‘wolf ’, älg ‘moose’, högd ‘height’) are still pronounced with a voiced velar stop in the Core Eastern, Gotland, Finland (except Åland) and Estonia. Southeastern and Western dialects have a palatal fricative in these positions (-rj, -lj, jd); initial g, k and sk are conserved as stops before a front vowel (gikk ‘went’ etc) in parts of Åboland and Nyland, Estonia and Gotland, as well as in some Core Eastern dialects.
A non-archaic phonological feature is –
the loss of final -t in neuters and the supine which occurs in the Eastern dialects, in Gotland and all of Swedish Finland.
On the border of phonology and morphology is the syllabified final -n (hästn horsethe ‘the horse’, båtn boat-the ‘the boat’, vattn ‘water’ instead of e.g. hästin etc.) which is found in north Österbotten, eastern and central Åboland, Nyland, the Eastern dialect area and Estonia.
Swedish dialects around the Baltic Sea 169
Norrbotten
Lappland
Västerbotten Jämtland Härjedalen
Ångermanland Medelpad Hälsingland
Österbotten
GästrikDalarna Åboland land VästÅland Värm- manland Uppland Nyland land Söder BoDals- Närke manland ESTONIA hus- land Österlän Västergötland Gotland götland SmåHalland Öland land Blekinge Skåne
Map 2. Historical provinces of the Swedish-speaking area.
5.
Concluding remarks
In this paper, I have reviewed and discussed Swedish dialect divisions. I have also given an overview of the varieties of Swedish found in Finland and Estonia, referred to here as the Trans-Baltic Swedish dialects. Finally, I have given an overview and discussion of dialect features from different linguistic levels. In section one, I proposed three new labels for dialect areas: Eastern (for the traditional sveamål), Western (for the traditional götamål) and Trans-Baltic (for the traditional östsvenska mål). Of the traditional terminology, the labels götamål and sveamål, are something of misnomers, emphasizing as they do a horizontal division of the southern third/half of Sweden (excepting of course the Southern dialect area) and thus obscuring the southeastern push of many Core Eastern characteristics, and
170 Anne-Charlotte Rendahl
the relative linguistic uniformity of a large part of the Baltic coastal area. In my view, then, the labels Western and Eastern Swedish would conform better with linguistic reality, emphasizing the fundamentally vertical, or rather diagonal, linguistic split of this part of Sweden. From the historical point of view, the Swedish dialectal map reflects the political and demographical situation of the Viking and Early Middle Ages quite strikingly. As already stated in the introduction, and reflected throughout this paper, the work done so far on Swedish dialects is abundant on the levels of phonetics and phonology and certain aspects of the lexicon, less so on the level of morphology, and very sparse in the area of syntax. Many interesting phenomena in the morphology and syntax of Swedish dialects are waiting to be explored, but they will not wait indefinitely — the pressure from (varieties of) standard Swedish on the dialects is quite intense, and many dialects are in a process of rapid change and/or threatened with extinction.
Notes * I would like to thank Östen Dahl for discussing and commenting on earlier versions of this paper, and a special thanks for producing the maps. I also thank Kari Fraurud for reading and commenting on an earlier version of the paper. 1. The area around the Gulf of Bothnia is included in this use of “the Baltic area” due to among other things the intimate relations, linguistic and contact-wise, between Österbotten and more southern parts bordering on the Baltic Sea proper, especially the Åland-Åboland area and Uppland-Södermanland. 2. With the status of Dalecarlia slightly altered with respect to the traditional view: as far as syntactic criteria go (which have been very sparsely used in standard work on dialect grouping) the dialects of Dalecarlia in many ways seem to have more in common with Northern dialects than with Eastern Swedish dialects with which they are traditionally grouped. I therefore choose not to include Dalecarlia in any of these groupings, but view it, for now, as forming a group of its own. 3. This from an ethno-centric Swedish mainland perspective. 4. I owe the terminology to suggestions from Östen Dahl, who also helped me see and make explicit the fundamentally vertical dialect division of middle/southern Sweden (East-West) in opposition to the previously accepted more horizontal picture patterning with the old Svealand– Götaland borders (cf. below). Notice that the dialects which are traditionally — and rather confusingly — called “Eastern Swedish” (östsvenska) have been renamed “Trans-Baltic”, “Eastern” being reserved for what is traditionally called sveamål. 5. Wessén (1969: 12–13) is careful to point out that there are no sharp boundaries between dialects; instead there is a continual change, where boundaries are mostly quite arbitrary and it is the core areas of each group that can be thoroughly distinguished from each other. 6. There is, however, quite an extensive area of transition (northern Småland, Västergötland, southeastern Värmland and Östergötland) where both the uvular and the apical are used in complementary distribution: the uvular is used word-initially and in geminates, the apical in all other cases (Wessén 1969: 16).
Swedish dialects around the Baltic Sea
7. But quoting G. Sjöstedt (no date given), Wessén notes that the plurals are best conserved (“most common in genuine speech”) in southern and middle Halland, southwestern Småland, northeastern Scania and Lister (1969: 20, ftn. 1). 8. Cf. Wessén’s northwest-southeast border line for many “important” features (Wessén 1969:72–73); an important collection of language maps also emphasizing the northwest-southeast linguistic division of (the southern third of) Sweden is Lindqvist 1947; a more recent appreciation of the same theme is Fries 1994. 9. Examples are: e and ö instead of i and y (fesk ‘fish’, töst ‘quiet’) in a westerly area covering large parts of Småland (in the southwest) as well as Scania and Blekinge and the western part of Sweden northward up through Värmland (the innovation does not apply across the board: you find i/y preceding the consonant clusters nd, ng and nk, and in certain areas also preceding ll (or its variants ld and lt) etc. (cf. Wessén, p. 23); open u or å (båske ‘bush’, flåtet ‘flown’); a closed o preceding gg, kk, ng and nk (hogga ‘cut, chop’, tong ‘heavy’, bonke ‘bowl’). 10. The province of Dalecarlia is traditionally included in the Eastern group (cf. ftn. 2). Even Wessén, however, points out that the dialects in upper Dalecarlia are quite separate from the rest of the Eastern dialects “with their very elderly character and great diversity” [“med sin mycket ålderdomliga prägel och sin starka splittring i underarter”] (1969: 30). As a peripheral/archaic dialect group in the Swedish context, upper Dalecarlia shares a number of features with other peripheral dialects like Estonian Swedish and dialects of Finland, as well as with dialects of Norrbotten and Västerbotten in Sweden. 11. Wessén notes that this feature was also found “earlier” in the archipelago of Östergötland “… and is thus one of the many Core Eastern features which have spread southwards along the east coast…” (1969: 31) (“och är sålunda ett av de många uppsvenska drag, som har spritts söderut utefter östkusten.”) 12. In the transcription of Western dialects, the letter å stands for ô. 13. Wessén notes (1969: 37) that German influence is behind the use of the adjective suffix -ig in the standard (written) language. 14. Cf. also the spread of features from the Core Eastern area southwards along the east coast (Wessén 1969: 31). 15. According to, among others, Hesselman (1905) (cited by Wessén, p. 43, ftn. 1) the “Norrlandic softening” has had a much larger geographical extension including parts of Södermanland and the archipelago of Östergötland. 16. As mentioned in the introduction, the Trans-Baltic dialects are treated in more depth in Section 3 of this paper. 17. This is a feature also common to Gotlandic and dialects of Dalecarlia and Roslagen (cf. själ ‘seal dog’ around the Baltic, säl on the Swedish west coast (Wessén 1969: 46). 18. Traces of this are also found in the Core Eastern area (middle eastern Uppland, southern Södertörn) and in Gotland. 19. This choice is not motivated, but merely stated: “The Finnish Swedish and Estonian Swedish dialects have been treated/considered as groups in their own right” (“Som grupper för sig har de finlandssvenska och de estlandssvenska målen betraktats” Jörgensen 1970: 9). 20. The definition of this dialect area, Jörgensen phrases in the following way: “Med syntaktiska drag i sydsvenska mål avses sådana företeelser i syntax, som huvudsakligen förekommer i och söder om Bohuslän, Västergötland och Östergötland samt på Gotland. Vissa av dessa drag kan dock förekomma så långt norrut som Södermanland, Närke och Värmland.” (1970: 10)
171
172 Anne-Charlotte Rendahl
21. The area referred to as containing Central Swedish (mellansvenska) dialects, Jörgensen delimits in the following way: “Med syntaktiska drag i mellansvenska mål avses sådana företeelser inom syntaxen som förekommer norr om Bohuslän, Västergötland och Östergötland men söder om Gästrikland, Hälsingland och Härjedalen. En företeelse har även noterats från Gotland.” (1970:18). 22. This feature extends over a larger area than the middle zone, however. 23. Tiselius (1902–03) describes the use of the negative suffix -e in the following way: “[It] occurs very often at the end of negated sentences ante pausam and is added to words, which end in a vowel as well as to those which end in a consonant. Often there is a[nother] negation in the same sentence, but sometimes the negation lies in the construction as such as in the third example [see the second example above in the main text]” (p. 147, my translation). “[Den] förekommer synnerligen ofta i slutet av nekande meningar ante pausam ock tillägges ord, som sluta på vokal såväl som sådana på konsonant. Ofta står i samma mening en negation utsatt, stundom ligger den endast i konstruktionen såsom i det tredje av de nedan anförda exemplen (see above in the main text)” (p. 147). 24. Jörgensen delimits the area referred to as containing Northern (nordsvenska) dialects in the following way: “Med syntaktiska drag i nordsvenska mål avses sådana företeelser i syntax, som förekommer i och norr om Gästrikland, Hälsingland och Härjedalen. Vissa av dessa drag kan förekomma så långt söderut som i Värmland och Södermanland.” (1970: 23). 25. We are concerned here with the Finnish Swedish dialects. The Finnish Swedish standard language has its roots in the mainland Swedish standard language under heavy influence from the city-language of Åbo (which was the administrative center during the Swedish hegemony) from the middle ages and onwards, and perhaps to a lesser extent from the Finnish Swedish “genuine” dialects (Ahlbäck 1971: 20). Somewhat simplified, it is standard (mainland) Swedish with a fair amount of influence from Finnish (this is very noticeable especially on the phonological level). A source to consult is e.g. Bergroth (1917). 26. The primary source for the section on Finnish Swedish dialects is Ahlbäck (1971). 27. Cf. Ahlbäck’s comment (1971:16) to the effect that archaisms rather have a negative value as a dialect criterion — they tend to show the lack of contact with the centre(s) of spread (innovations). 28. In Åboland and Nyland there is an alternation between completely unretroflexed pronunciations like fors and bort and pronunciations where there is a retroflex but preceded by the r (Ahlbäck 1971: 12). 29. Österbotten has -il for the corresponding function: tosil, väsil etc. — both these cases have counterparts in the Core Eastern — and -il also occurs in southern Finland as well as in Sweden, at least in the Mälar region (Ahlbäck 1971: 16). 30. In Österbotten, there is a completely different system. In predicate position, there is no suffix on the adjective. When the adjective occurs as an attribute, a special inflection has evolved which has its point of departure in the old masculine accusative form ending in -an (storan etc. ‘big’) (in central Österbotten there is no suffix on the adjective however); example: ein grannan kar ‘a goodlooking man’ vs. karin je grann man-the is goodlooking ‘the man is goodlooking’. Feminine nouns trigger the following marking on an adjective functioning as an attribute: grannun in the south, granni(n) in the north (of Österbotten) (Ahlbäck 1971: 17). 31. There are similar uses of this verb in Northern dialects and in upper Dalecarlia (Älvdalen). 32. There were also about a thousand “Swedes” living in the cities/towns, of which 798 in Tallinn (where there was also a Swedish primary school (teaching no doubt in some standard variety of Swedish)).
Swedish dialects around the Baltic Sea
33. I have not found any information on whether the vara-construction is the preferred pattern, nor in what contexts/with what main verbs it is used. 34. “Referential function” is explained in the following way: “Referential function means that a relation to a referent is established” (Nikula 1997: 205). 35. For further details of the analysis, the reader is referred to Nikula (1997). Another function of the definite suffix which we have not gone into here is to mark gender. Nikula takes this to be one of two primary functions of the definite suffix (the reference-establishing one being the other; 1997: 212). 36. That is Norrland, the Core Eastern area, Österbotten and Estonia. Cf. Wessén (1969: 43) who states that adjective incorporation is common in the Core Eastern area as well as in Finnish Swedish dialects. 37. This practice of compounding would apparently “lead to” a decrease in the use of the preposed definite article, in relation to the standard language and more southerly dialects. 38. Among young speakers of these dialects the dative usage pattern is changing, however, and the dative case seems to be disappearing from active use. 39. Cf. a typical comment by Danell (1905–34: 20): “Of case inflection there remains but a fading genitive (always in -s)” “Av kasusflexion återstår endast en föga livskraftig genitivbildning (alltid på -s).” Danell goes on: “As casus rectus the old accusative has in general replaced the nominative” “Som casus rectus har i allmänhet den gamla ackusativen undanträngt nominativus.” (p. 20). 40. For Estonia, however, cf. Tiberg who, in opposition to Danell, states that the genitive case is more frequently used (“brukas oftare”) in Estonian Swedish than in the standard language, and “not only of the type personal owner but also of the type ökens rumpan horse-the-gen tail-the ‘the horse’s tail’” (Tiberg 1962: 98). 41. However, as is the case in standard Swedish, there seems to be a tendency for the uter (nonneuter) pronoun den ‘it’ to replace the feminine and masculine personal pronouns hon and han (given here in their standard forms) thus replacing the threeway system (hon, han, det) with a two way- one (den (uter), det (neuter)) which does not distinguish between masculine and feminine (except for referring to animates). Such a process is for example documented for many dialects of Nyland (Sandström 1997). This may be true for various middle variants which are sometimes called regional standard varieties or regional dialect etc. (to which Sandström appears to be referring in the case of Nyland) but it may not be true at all in “genuine” dialects: cf. Davidson (1990: 151), citing Thelander (1975) on the Burträsk dialect of Västerbotten where the older system is alive and well and even seems to be gaining terrain incorporating new words into the old system (mostly, it seems, assigning them masculine gender). 42. A contributing factor which Huldén mentions is the fact that the dialect (in the period of change) probably was spoken by a considerable number of speakers who had Finnish as their stronger language. Gender assignment, not being a rule governed process, is a likely area for hesitance in speakers with a weaker proficiency in a language (cf. Andersson 1992) and an uncertain and variable use in a rather large group of speakers could well rub off on the dialect of the whole speech community (cf. Hulden 1972: 78–79). 43. For example in Lister, there is a distinction in number between singular and plural, and a distinction in person within the plural: 1st and 2nd plural both end in -n while 3rd plural ends in -a (past tense -e) (Swenning 1917–1937: 536). 44. It is difficult to interpret Tiberg on this: “… only of the type jag var/vi voro” — does this mean that it is only the copula that distinguishes number in the past tense? According to Danell, verbs
173
174 Anne-Charlotte Rendahl
of certain classes (“i vissa konjugationer”) are marked for number in the past tense, and number is also distinguished in the imperative (Danell 1905–34: 25, Tiberg 1962: 100). 45. Use of the cardinal directions to describe spatial location also occurs in Dalecarlian as spoken in Älvdalen (along with the more well-known paraphernalia of spatial description, as well as locally determined terms which refer to up- and down-river). Geocentric spatial systems are found in for instance some Australian and Mesoamerican languages (see e.g. Haviland & Levinson 1994). Geocentric space has not received attention from dialectologists but from more or less anecdotal evidence one can infer that it is perhaps (or has been) fairly widespread in Swedish dialects, not only in upper Dalecarlia and Estonia. According to Dagens Nyheter (August 9, 1996), cardinal directions are used to locate entities in space on the tiny Vinön in Lake Hjälmaren in the Southeastern Swedish dialect area, and personal communication from Erik Andersson (via Östen Dahl) suggests that this is a widespread practice in Åland. 46. Polysyllabicity is a factor in the (non-)conservation of short root syllables: Estonian Swedish, for example, has short bisyllables but no short monosyllabic words. 47. Here, the root syllable usually gets a two peak accent (cirkumflex) “as a replacement” (Wessén 1969: 40). 48. Åland has both accents, however, and there are “traces” of grave accent in western Nyland (Elert 1995: 131). 49. The realizations of the accents in general tend to carry over to the respective regional varieties of Standard Swedish. 50. Upper Dalecarlia, however, shows a mixture of tone accent patterns (Gårding 1974). The most conservative dialects tend to have type 2a, suggesting that this pattern which is also found in most of the Eastern and Northern areas, is the oldest one. 51. Cf. Wessén (1969)’s (and others’) proposed division in northwesterly-southeasterly direction across the southern third of Sweden, and the discussion above in Section 2.2.
References For the interested reader I include a rather extensive bibliography. Some of the items that appear on the list are not referred to in the text. Ahlbäck, Olav. 1971. Svenskan i Finland. Skrifter utgivna av Nämnden för svensk språkvård 15. 2nd edition. Stockholm: Läromedelsförlagen. Ahlbäck, Olav. 1983. Österbottniska dialekter. In: Svenska Österbottens historia IV. Svenska Österbottens landskapsförbund. Vasa. Alving, Hjalmar. 1917. Kalmarmålets ljudsystem. Svenska landsmål och svenskt folkliv 1917. Stockholm. Andersson, Anders-Börje. 1992. Second Language Learners’ Acquisition of Grammatical Gender in Swedish. Gothenburg Monographs in Linguistics 10. Department of Linguistics, University of Gothenburg. Åström, P. 1893. Degerforsmålets formlära jämte exkurser till ljudläran. Svenska landsmål och svenskt folkliv 13. Stockholm. Bergman, Gustaf. 1893. Alundamålets formlära. Svenska landsmål och svenskt folkliv 12. Stockholm. Bergman, Gösta. 1952. Nordliga provinsialismer i det bildade svenska talspråket. Lund.
Swedish dialects around the Baltic Sea
Bergroth, Hugo. 1917. Finlandssvenska. Handledning till undvikande av provinsialismer i tal och skrift. Helsinki: Holger Schildts förlag. Bucht, Torsten. 1962. Språket i Härnösand. Stockholm: Svenska bokförlaget. Dahlstedt, Karl-Hampus. 1971. Norrländska och rikssvenska. I Norrländska och nusvenska Tre studier i nutida svenska. Lund: Studentlitteratur. Dahlstedt, Karl-Hampus. 1972. Den nordiska språkgemenskapen. Folkmålsstudier 12: 9–25. Helsinki. Dahlstedt, Karl-Hampus and Ågren, Per-Uno. 1980. Övre norrlands bygdemål. Skr. utg. av Johan Nordlander-sällskapet 2. Umeå: Cewe. Danell, Gideon. 1905–1934. Nuckömålet. Svenska landsmål och svenskt folkliv B 33. Stockholm. Danell, Gideon. 1912. Några anmärkningar om Vättern som språkgräns. In: Xenia Lideniana. Stockholm: P. A. Norstedt & Söner. Davidson, Herbert. 1990. Han hon den. Genusutvecklingen i svenskan under nysvensk tid. Lund: Lund University Press. Delsing, Lars-Olof. 1996. Nominalfrassyntax i skandinaviska dialekter. I Syntaktisk variasjon i nordisk målføre, Ø. A. Vangsnes (ed.), Nordica Bergensia 9, 1996, Nordisk institutt, Universitetet i Bergen. Edlund, Lars-Erik (ed.). 1994. Kulturgränser — myt eller verklighet? Diabas 4. Umeå: institutionen för nordiska språk, Umeå universitet. Ejskær, Inger. 1964. Nogle ejendommeligheder ved relativkonstruktionen i dansk ømål og i skånsk. Acta philologica scandinavica 26, København. Elert, Claes-Christian. 1994. Indelning och gränser inom området för den nu talade svenskan — en aktuell dialektografi. In: Edlund (ed.). Elert, Claes-Christian. 1995. Allmän och svensk fonetik. 7th edition. Stockholm: Norstedts förlag. Ericsson, Torsten. 1914. Grundlinjer till undersökningen av Södermanlands folkmål. Svenska landsmål och svenskt folkliv B 8. Stockholm. Freudenthal, O. A. 1878. Über den Närpesdialecht. Helsinki: J. C. Frenckell & Sohn. Fries, Sigurd. 1994. Dialektgränser och kulturgränser. In: Edlund (ed.). Gårding, Eva. 1974. Kontrastiv prosodi. Lund: LiberLäromedel. Gårding, Eva. 1977. The Scandinavian Word Accents. Lund: CWK Gleerup. Grip, Elias. 1901. Skuttungemålets ljudlära. Svenska landsmål och svenskt folkliv 18. Stockholm. Gustavson, Herbert. 1974. Inledning till gutamålets studium. Gotlandica nr 5. Visby. Hagfors, K. J. 1891. Gamlakarlebymålet. Ljud- och formlära samt språkprov. Svenska Landsmål 12. Stockholm. Hallberg, Göran. 1996. Öland: Dialekter. In Nationalencyklopedin. Höganäs: Bra Böckers Förlag. Haviland, J., Levinson, S. (eds.). 1994. Linguistics 32–4/5, special issue on Spatial conceptualization in Mayan languages. Hesselman, Bengt. 1905. Sveamålen och de svenska dialekternas indelning. Uppsala. Hesselman, Bengt. 1908. Uppländskan som skriftspråk. In Uppland, vol. 2. Stockholm: Wahlström & Widstrand. Hesselman, Bengt. 1936. Ordgeografi och språkhistoria. Stockholm: Hugo Gebers förlag. Holm, Gösta. 1952. Om s-passivum i svenskan. Lund: CWK Gleerup. Holm, Gösta. 1972. Nylandssvenskars och finnars svenska. Folkmålsstudier 12: 27–43. Helsinki. Holm, Gösta. 1980. Monoftongeringens kronologi och de finlandssvenska folkmålen. Studier i Nordisk Filologi 62: 101–107. Helsinki. Huldén, Lars. 1972. Genussystemet i Karleby och Nedervetil. Folkmålsstudier 12: 45–82. Helsinki. Hummelstedt, Eskil. 1934. Om användningen av substantivets bestämda form i österbottniskt folkspråk. Folkmålsstudier 2: 134–36. Helsinki.
175
176 Anne-Charlotte Rendahl
Hummelstedt, Eskil. 1939. Östsvenska verbstudier. Inkoativa verb på -na och verb med k-, l-, r-, s- eller t-suffix i Närpes-målet. Helsinki. Isaacsson, August. 1923. Om södra Fjärdhundralands folkmål. Svenska landsmål och svenskt folkliv B 21. Stockholm. Ivars, Ann-Marie. 1988. Närpesdialekten på 1980-talet. Studier i Nordisk Filologi 70. Helsinki. Ivars, Ann-Marie. 1996. Stad och bygd: Finlandssvenska stadsmål i ett regionalt och socialt perspektiv. Folkmålsstudier 37. Helsinki. Jansson, Valter. 1939. Om de östsvenska dialekternas ställning. Svenska landsmål och svenskt folkliv. 1939. Uppsala. Jörgensen, Nils. 1970. Syntaktiska drag i svenska dialekter. En bibliografisk översikt. Lund: Studentlitteratur. Källskog, Margareta. 1993. Uppländska. Språkprov med kommentar. Dialekt- och Folkminnesarkivet, Uppsala. Karsten, Anton. 1892. Kökarsmålets ljud- och formlära. Svenska landsmål och svenskt folkliv 12. Stockholm. Klockars, J. 1933. Malaksmålets ljud- och formlära. Folkmålsstudier 1: 17–62. Helsinki. Kruuse, E. 1908. De lefvande folkmålen. In Uppland, vol. 2. Stockholm: Wahlström & Widstrand. Lagman, Edvin. 1958. Studier över Nuckömålets nominala ordbildning. Stockholm: Almqvist & Wiksell. Lagman, Herbert. 1971. Svensk-Estnisk språkkontakt. Studier över estniskans inflytande på de estlandssvenska dialekterna. Stockholm: Almqvist & Wiksell. Larsson, Seth. 1929. Substantivböjningen i Västerbottens folkmål jämte en exkurs till ljudläran. Uppsala. Levander, Lars. 1909. Älvdalsmålet i Dalarna. Ordböjning och syntax. Stockholm: P. A. Norstedt & Söner. Levander, Lars. 1925. Dalmålet: beskrivning och historia. Part 1. Uppsala. Levinson, Stephen. 1992. Primer for the field investigation of spatial description and conception. Pragmatics 2–1, 5–47. Lindkvist, Einar. 1942. Om Gästriklands folkmål. In Från Gästrikland 1942. Gästriklands Kulturhistoriska Förenings meddelanden. Gävle. Lindqvist, Natan. 1947. Sydväst-Sverige i språkgeografisk belysning. Skrifter utgivna genom landsmålsarkivet i Lund 2. Lund. Loman, Bengt. 1972. Om tvåspråkigheten i östra Nyland. Folkmålsstudier 12: 127–143. Helsinki. Loman, Bengt. 1977. Språk och skola i Kronoby, Österbotten. FUMS 52. Uppsala universitet. Uppsala. Lundström, Gudrun. 1939. Studier i nyländsk syntax. Stockholm: Norstedt & Söner. Nikula, Kristina. 1997. Species i finlandssvensk dialekt. In: Reinhammar, M. (ed.), 203–214. Noreen, Adolf. 1903. Vårt språk I. Lund: Gleerup. Nyström, Staffan. 1997. Grav accent i östra Svealands folkmål. In: Reinhammar, M. (ed.), 215–222. Pamp, Bengt. 1978. Svenska dialekter. Stockholm: Natur & Kultur. Reinhammar, Maj. 1973. Om dativ i svenska och norska dialekter. Del 1. Dativ vid verb. Uppsala: Almqvist & Wiksell. Reinhammar, Maj (ed.). 1997. Nordiska dialektstudier. Föredrag vid Femte nordiska dialektologkonferensen Sigtuna 17–21 augusti 1994. Skrifter utgivna av språk- och folkminnesinstitutet genom dialektenheten i Uppsala. Ser. A: 27. Rutberg, Hulda. 1924. Folkmålet i Nederkalix och Töre socknar. Svenska landsmål och svenskt folkliv B. 28, H.2. Stockholm.
Swedish dialects around the Baltic Sea 177
Sandström, Caroline. 1997. Grammatiskt genus i nyländska dialekter. In: Reinhammar, M. (ed.), 263–272. Schagerström, August. 1882. Upplysningar om Vätömålet i Roslagen. Svenska landsmål och svenskt folkliv 2. Stockholm. Schagerström, August. 1949. Grammatik över Gräsö-målet i Uppland. Svenska landsmål och svenskt folkliv B. 51. Stockholm. Sjöberg, Åke G. 1992. Gotland: Historia. In Nationalencyklopedin. Höganäs: Bra Böckers Förlag. Sjöstedt, Gösta. 1944. Det nordöstskånska dialektområdet. Arkiv för Nordisk Filologi 59: 73–96. Lund. Stenbom, Eric. 1915. Njurundamålets formlära (Medelpad). Svenska landsmål och svenskt folkliv 1915. Stockholm. Swenning, Julius. 1917–1937. Folkmålet i Listers härad i Blekinge. Ljudlära och formlära. Svenska landsmål och svenskt folkliv B. 36. Stockholm. Thelander, Mats. 1975. Kodväxling eller kodblandning? ett försök att beskriva språksociologisk variation i en norrländsk talgemenskap. FUMS 40. Uppsala. Tiberg, Nils. 1933. Estlandssvenska och estniska. Svenska landsmål och svenskt folkliv 1933. Uppsala. Tiberg, Nils. 1962. Estlandssvenska språkdrag. Estlandssvenskarnas folkliga kultur 6. Acta Academiae regiae Gustavi Adolphi XXXVIII. Lund. Tiselius, G. A. 1902–1903. Ljud- och formlära för Fasternamålet i Roslagen. Svenska landsmål och svenskt folkliv 18. Stockholm. Tjäder, Börje. 1988. Mellansvensk diftongering i ny belysning. Svenska landsmål och svenskt folkliv 1988. Uppsala. Vallmark, Sven. 1936. Fördelningen av de temporala konjunktionerna då och när i svenskt talspråk. Svenska landsmål och svenskt folkliv 1936. Uppsala. Vendell, Herman. 1882. Runömålet. Ljud- och formlära samt ordbok. Svenska landsmål och svenskt folkliv 2. Stockholm. Vestlund, Alfred. 1923. Medelpads folkmål. Svenska landsmål och svenskt folkliv B. 48, H.4. Stockholm. Västerlund, Rune. 1988. Alundamål. Utskrift av en uppländsk inspelning. Svenska landsmål och svenskt folkliv 1988. Uppsala. Wessén, Elias. 1958. Svensk språkhistoria: 1. Ljudlära och ordböjningslära. Stockholm: Almqvist & Wiksell. Wessén, Elias. 1965. De nordiska språken. Stockholm: Almqvist & Wiksell. Wessén, Elias. 1969. Våra folkmål. Nionde upplagan. Lund: Fritzes. Wessman, V. E. V. 1936. Bidrag till kännedomen om Österbottens svenska folkmål. Folkmålsstudier 4. Helsinki. Zilliacus, Kurt. 1992. Åboländska. Forskningscentralen för de inhemska språken. Helsinki.
AUTHOR "Johanna Laakso"
TITLE "The Finnic languages"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
The Finnic languages Johanna Laakso
1.
Introduction
The aim of this survey is to introduce the Finnic languages for linguists not acquainted with the Finno-Ugrian or Uralic languages. (The terms “Finno-Ugrian” and “Uralic” are here used as synonyms, since there may not be such a clear dichotomy between the Finno-Ugrian and Samoyed branches of the Uralic language family as assumed in earlier research, cf. Salminen [ed.] 1993.) It is intended to introduce or supplement — rather than replace — the previous surveys of Finnic, especially the indispensable monograph by Laanest (1982) and articles included in Sinor [ed.] (1988) and Abondolo [ed.] (1988). The bibliography and references to sources are by no means extensive; I have tried to concentrate on research published after the previously mentioned books. Sources for recent research can also be found e.g. in Suhonen (1995); for those able to read Finnish, Häkkinen (1996) is a lucid survey of research into the roots of Finnish (and Finnic). The origin of the Finnic ethnonyms, together with some central questions of Finnic ethnic history, are closely examined in Grünthal (1997). This survey is based on existing sources; no independent basic research has been done. Unfortunately, this means that the smaller Finnic languages receive much less attention than Finnish and Estonian. Even contrastive research tends to concentrate on these two languages (for interesting comparisons between the development of Finnish and Estonian see e.g. Metslang 1997). For Karelian, Veps, Votian, Ingrian and Livonian, there is less material available; their (historical) phonology and morphology are fairly well known, but their syntax is particularly badly in need of (more modern) research. The existing views on their (historical) morphosyntax could perhaps also be revised. For this reason, Finnic syntax is more or less neglected in this paper. In order to give due attention to questions connected with language contacts and areal linguistics, I have attempted to concentrate on the position of the Finnic languages among other branches of Uralic, the relationships between them and the history of language contacts. The Uralic prehistory of Finnic must be referred to, in order to show which features of Finnic are of ancient Uralic heritage and which can
180 Johanna Laakso
be explained as more recent areal or contact phenomena. The Uralic background of Finnic phonology, word structure and morphology is concisely presented in Janhunen (1982) and Sammallahti (1988). 1.1 General characterization, speaker statistics
and geographical distribution The Finnic (or Fennic) languages are a distinct branch of the Uralic (Finno-Ugrian) language family. Sometimes they are, in a somewhat confusing way, called “Baltic Finnic” while in older practice the term “Finnic” or even “Finnish” was applied to almost all Finno-Ugrian languages! The Finnic languages are fairly closely related; some of them (like Finnish and northern Karelian) can almost be considered mutually intelligible. In terms of the traditional family-tree description, they are descendants of a common protolanguage (Proto-Finnic) probably spoken until about 2000 years ago. Even after that, innovations could spread between Finnic dialects. The closest relatives and neighbours of Proto-Finnic were Proto-Sami in the north and, possibly, the more or less hypothetical, extinct Uralic languages (Merya, Muroma, Meshchera?) in the south-east. These languages (or dialects) are mentioned in old Russian chronicles but not recorded; they are sometimes hypothesized as the “missing link” between Proto-Finnic and the ancestors of the present-day Volgaic languages, Mordvin (= Moksha and Erzya) and Mari (“Cheremis”). Contacts between Proto-Finnic and its southeastern relatives, whatever they were, probably continued until the expansion of Slavic tribes to what is now Northwestern Russia (towards the end of the first millenium AD) and, between some Finnic and Sami languages, up to the present day. The Finnic languages are so closely related that it is meaningful to describe their common features on the basis of a reconstructible proto-language. This reconstruction, of course, shares the general shortcomings of such reconstructions (cf. Korhonen 1978) and should be considered an instrument in describing the relationship between present-day Finnic languages rather than a completely realistic description. The following are the languages of the Finnic group: –
Finnish: roughly 5 million speakers in Finland and northernmost Sweden (where the Torne River dialect of the Swedish side is now claiming independent status as a separate language, meän kieli ‘our language’). Finnish is also spoken by old emigrant groups in Norway (the so-called kven language) and Ingria. (Ingria is “Ingermanland”, the vicinity of St. Petersburg, where many Finns emigrated to, especially in the 17th century, when Ingria was annexed to Sweden. Due to massive repression in the Soviet era, war-time evacuation and post-Soviet emigration, many Ingrian Finns now live in Russia, Estonia, the
The Finnic languages
–
–
Republic of Karelia, Finland or Sweden.) Other important groups of emigrant Finns live in Russian Karelia, Russia, America, Australia and (especially since the 1960s) Sweden. Estonian: roughly 1 million speakers in Estonia; spoken also by emigrant groups in Russia, Western Europe (especially Sweden), America, Australia etc. (considerable numbers of Estonians fled to the West in 1944). The southeastern dialect, which is markedly different from Standard Estonian, is now supported by a group of activists claiming a status as a separate language, võro kiil. Karelian (Karelian-Olonetsian): around 70,000 speakers in Russia (the Republic of Karelia; due to massive emigration in the 17th century: the vicinity of Tver’ in Russia proper), emigrant groups in Finland (usually only the oldest generation). (The Karelian language should not be confused with the “Karelian” dialects of Finnish spoken in easternmost Finland and before World War II on the Karelian Isthmus, i.a. the city of Viipuri/Viborg/Vyborg.)
Karelian Finnish
Vepsian
Estonian
Livonian
Map 1.
Ingrian Votian
Karelian proper (North & South) Olonetsian Ludian Vepsian
181
182 Johanna Laakso
– – –
–
–
Ludian (sometimes considered a dialect of Karelian): around 5000 speakers in the Republic of Karelia. Veps: around 6000 speakers in the Republic of Karelia and in the neighbouring areas of the districts of Vologda and St. Petersburg. Ingrian (Izhorian): around 300 speakers in Ingria (the vicinity of St. Petersburg); not to be confused with Ingrian dialects of Finnish spoken by Finnish (Lutheran) immigrants who came to Ingria mainly in the 17th century. Votian: around 60 speakers in Ingria (numbers estimated vary according to the criteria set for native informants; all speakers are old and their language is heavily mixed with elements from neighbouring Finnic languages and Russian). Livonian: a few dozen speakers in Latvia (the pre-war Livonian-speaking area at the tip of the Courland peninsula, also in Riga and other towns and cities of Latvia).
For the geographical distribution see Map 1. 1.2 General characteristics of the Finnic languages 1.2.1The sound system The Finnic languages, especially Finnish, are often called phonologically conservative. This concerns the vowels in particular and the stability of original unstressed and suffixal syllables: e.g. kala ‘fish’ and pesä ‘nest’ look identical (at least graphemically) in Finnish and in most Proto-Uralic (PU) reconstructions. Mainstream reconstructions of PU vocalism, word structure and phonotactics (e.g. Janhunen 1982; Sammallahti 1988; UEW) mostly rely on Finnic on the one hand and reconstructed Proto-Samoyedic on the other. The reconstructed changes that have led to the Finnic vowel system have mostly meant enrichment: the introduction of new vowels, long vowel phonemes and diphthongs. On the other hand, the original PU (or: FU) consonant system was probably heavily reduced when Proto-Finnic came into being: distinctive palatalization (PreFinnic *s´, *c´, *d’, *n´, probably also *l’) and the phonemes *š, *cˇ and *] disappeared altogether. The hitherto unchallenged explanation by Lauri Posti (1954) attributes these phenomena to early Baltic and Germanic influence: the phonemes not present in the Baltic and Germanic systems would simply have been filtered out. 1.2.1.1 The vowels. The Proto-Finnic vowel system (in the first syllable) was probably the following: i, ii e, ee ä, ää
ü, üü (·yÒ, ·yyÒ in Standard Finnish orthography) (ö, öö)
u, uu o, oo a, aa
The Finnic languages
The quantity correlation probably did not widen to cover all vowels before the Late Proto-Finnic period. The vowel ö was added to the paradigm relatively late; it appears in descriptive words (e.g. Fi. Est. köhi- ‘to cough’), loanwords (e.g. Est. köök ‘kitchen’) and cases of assimilation (e.g. Fi. löytä- ‘to find’, cf. Est. leida < ProtoFinnic *leütä-). The diphthongs originally only included those ending in i (< PU j, ]¢ ) and u/ü (< PU w, ]). It is usually assumed that only high and low non-rounded vowels (i, ?çi , a, ä) were allowed in PU non-initial syllables. Older reconstructions use e for i in noninitial syllables; this is typologically less credible but more in accordance with the evidence of present-day Finnic languages, where these words still show e in non-final positions (underived native words with a non-low vowel in the 2nd syllable, e.g. Fi. lapsi ‘child’: stem lapse-, as opposed to loanwords and derivatives in i, where the stem vowel is always i, e.g. lasi ‘glass’: lasi-, kasvi ‘plant’ [from kasva- ‘to grow’]: kasvi-). The Finnic rounded vowels in non-initial syllables go back to sequences of nonrounded vowels and labial consonants (e.g. Fi. nato, Est. nadu ‘sister-in-law’ < PU *natçi w), or they are a result of later developments. By the time of Proto-Finnic, rounded vowels had started to appear in non-initial syllables, too; the last innovation was suffixal ö, needed as a counterpart for o to complete the vowel harmony. Votian, which otherwise has retained vowel harmony, still shows archaic cases like tüttärikko ‘girl’ with the suffix -kko (instead of *-kkö). Most Uralists consider vowel harmony to be an ancient Uralic feature (an alternative explanation has been presented by Korhonen 1988). Originally, it meant that initial-syllable i, e, ä and ü could only be combined with i (or e) and ä in the non-initial syllables, and initial-syllable u, o, and a only with i (or *çi ?) and a. By the time of Late Proto-Finnic, there were also combinations of the type e–a and i–a (especially in loan words). This means that suffixes with non-neutral vowels (a, ä, newer suffixes with o, ö, u, ü) had two variants, e.g. *ta and *tä for the partitive ending (henceforth the usual archiphonemic notation is used: tA). In non-initial syllables, only short vowels were allowed. The changes in the vowel system after the Proto-Finnic stage include the following: –
The mid-vowel çe (õ in Standard Estonian orthography) appears in southern Finnic languages (Votian, Estonian, Livonian) mainly as a more or less systematic counterpart for northern Finnic e in back-vowel words (cf. Est. võlg ‘debt’, lõug ‘chin’ — Fi. velka, leuka). Interestingly enough, this new phoneme, obviously brought forth by regressive assimilation, appears (although not exclusively) in those languages where the original vowel harmony has disappeared (i.e. Estonian and Livonian). This is contrary to the progressive assimilation manifest in the Finnic vowel harmony, i.e. the fact that first-syllable or stem vowels determine the vowels of the following syllables.
183
184 Johanna Laakso
–
– –
–
–
–
Diphthongization of long vowels. The long vowels ee, oo and öö have given ie, uo and üö in Finnish, Karelian and Livonian and dialectally in Estonian and Votian. Terho Itkonen (1983) has considered this to be an originally Western Finnic innovation; when spreading eastwards, it has widened to encompass aa and ää (> oa or ua, eä or iä in East Finnish dialects and Karelian). New diphthongs have been introduced by loanwords and changes deleting original consonants, e.g. Est. näo ‘face-gen.sg’ < *nägo-n, toa ‘room-gen.sg’ < *tuba-n. In Finnish, Karelian, Ingrian and Votian, suffixation and deletion of suffixal consonants have created long vowels and diphthongs in non-first syllables, e.g. Fi. kalaa ‘fish-prtv’ < *kala-da < Pre-Finnic *kala-ta, Fi. lampaita ‘sheeppl-prtv’ < *lampah-i-ta (cf. Veps. kalad, lambhid). In Estonian and Livonian, centralizing sound changes have reduced the vowel repertoire in non-first (unstressed) syllables and deleted the vowel harmony (cf. e.g. Fi. sysi ‘charcoal’, kysy- ‘to ask’, kala ‘fish’, kylä ‘village’, nato ‘sister-in-law’, karhu ‘bear’ — Est. süsi, küsi-, kala, küla, nadu, karu). In dialects of Karelian, Veps, Ingrian, Votian and Estonian, a higher mid vowel (ı˘) appears. At least in the easternmost Finnic languages it has probably been introduced or at least enforced by the influence of Russian loanwords. In Veps, the long vowels have (generally) been shortened.
1.2.1.2 The consonants. After the previously mentioned changes that impoverished the original Pre-Finnic (more or less identical to PU) consonant system, the consonant phonemes in Proto-Finnic included three stops (p, t, k), one sibilant (s) and one affricate (c, only in word-internal position), two nasals (m, n), two liquids (l, r) and three spirants (h, v, j). Of these, h was a new phoneme, replacing old *š or representing old s in some positions (producing alternations like Fi. mies ‘man’: gsg. miehe-n). All of these, except the spirants, could also be geminated. This simple system is complicated by gradation (Germ. Stufenwechsel), a consonant alternation which appears in all Finnic languages except Veps and Livonian; more or less similar phenomena are found in most Sami languages (except South Sami) and some languages of the Samoyedic group. It has been considered either a relatively young innovation, developed separately in these branches of Uralic, or ancient Uralic heritage preserved only in these most conservative branches (Helimski 1995). Two explanations have been suggested: it can be considered a natural result of Uralic word structure and phonotactics (e.g. Janhunen 1982), but Posti (1954) tries to connect the Finnic gradation with “Verner’s Law” in Germanic. The relationship between Finnic and Sami gradation is specially problematic. Its conditions are basically the same in both groups, but its distribution causes problems for the relative chronology of these proto-languages. In Veps and Livonian, there is no indisputable evidence for original gradation; many scholars assume, like Laanest (1982: 106–112), that if there ever was gradation, it
The Finnic languages
was not phonemicized before it was lost in these languages. The gradation means that original Finnic stops alternated with voiced spirants (*b, *d, *g, according to the traditional reconstruction), while original geminate stops alternated with shorter geminates. Traditionally, radical (= stem-internal) gradation is distinguished from suffixal gradation. In word stems subject to radical gradation, the weak grade (spirants or shorter geminates) always appeared at the beginning of a closed syllable, e.g.: sika ‘pig’: gsg. *siga-n, kota ‘hut’: gsg. *koda-n, katto ‘roof ’: gsg. *katto-n. The radical gradation concerned stem-internal consonants (between the first and the second syllable), while suffixal stops were subject to suffixal gradation: strong grade after a stressed (= odd-numbered) syllable, weak grade after an unstressed (= even-numbered) syllable, as can still be seen in cases like Fi. maa-ta ‘earth-prtv’, kalaa (< *kala-da) ‘fish-prtv’, [dial. or arch.] korkeata (< *korkeda-ta) ‘high-prtv’. The original suffixal gradation has often been obscured by analogical tendencies towards the formal unity of suffixes in all positions. In almost all Finnic languages, the supposed voiced spirants have been replaced by other consonants (liquids or v, e.g. Fi. puvun ‘dress-gen’ < *pugu-n, luvan ‘permission-gen’ < *luba-n, dial. kären, kälen < *käde-n ‘hand-gen’; stops, e.g. Est. heleda ‘bright-gen’ < *heledä-n) or they have disappeared completely (Fi. dial. käen, Est. käe ‘hand-gen’, Fi. sian, Est. sea ‘pig-gen’ < *siga-n). In Finnish, (northern) Karelian and Ingrian, the supposed shorter geminates have merged with original single stops. These changes — together with others obscuring the original openness or closedness of syllables and thus the conditions for the gradation — have contributed to the morpho(phono)logization of the gradation. Other post-Proto-Finnic changes in the consonant system include the following: –
–
–
Karelian (except the northernmost dialects), Veps, Votian and Livonian now have voiced stops (and sibilants). This is probably due to both foreign (Russian or Latvian) influences and internal developments (cf. Laanest 1982: 122–124). In Estonian and Ingrian, original stops between vowels and/or voiced consonants appear as half-voiced b, d and g (in Standard Estonian orthography: ·b, d, gÒ). Standard Finnish d (the weak-grade counterpart of t), which is the only voiced stop in native Finnish words, is a result of the literate reanalysis of ·dÒ; the letter ·dÒ originally stood for d, which still appears in some southwestern Finnish dialects. Karelian, Veps, Votian and Livonian also use š (in loanwords and in descriptive words) and word-initial affricates. Affricates also appear in southern Estonian dialects (e.g. tsiga ‘pig’). Palatalization has been reintroduced to the system of eastern Finnish dialects, Karelian, Veps, Estonian and Livonian. Its forms and distribution vary; most often it appears in consonants preceding (original) i or j, and sometimes the deletion of original i has given the palatalization a distinctive function. At least
185
186 Johanna Laakso
–
–
–
in Karelian and Veps it is probably also influenced by Russian. ] (originally an allophone of n before k or its weak-grade counterpart) is an independent phoneme at least in Finnish (and some other dialects), where geminate ] functions as the weak-grade counterpart of nk. In some dialects of Finnish and southern Estonian, word-final *k has changed to a glottal stop or a weaker glottal pressure; in some other dialects of Finnish, it survives as a sandhi phenomenon. In Veps (dialectally), the geminate stops have been shortened.
1.2.2Word structure and prosodic features It is generally assumed that the original structure of Uralic underived content (nondeictic) words was disyllabic, of the type (C)V(C)CV (or, maybe, (C)V(C)(C)CV). Word-initial or syllable-initial consonant clusters were not allowed. After the word stem there could be one or more suffixes representing the structure types C, CV, CCV or even longer (complex?) structure types (Janhunen 1982). Compound words consisting of two nouns may also be an old Uralic phenomenon, although, in more distantly related Uralic languages, they are usually analyzed and written as two-word constructions with a nominative attribute, e.g. Komi lym mort, cf. Fi. lumiukko (lumi+ukko) ‘snowman’. The Finnic languages have generally preserved the supposed Uralic word structure well. The tendency towards the original disyllabic word type is still visible in e.g. Finnish word formation (cf. Laalo 1983). The greatest changes in this respect are found in Estonian (and some Finnish, especially southwestern Finnish dialects), Livonian and also Veps, where unstressed vowels and syllables have often been deleted in regular apocope (e.g. Est. mets ‘forest’, cf. Fi. metsä, Veps. ak ‘woman, wife’, cf. Fi. akka) and syncope (e.g. Est. tütred ‘daughters’, cf. Fi. tyttäret). These deletion tendencies, — which often affect the suffixes, — have given rise to morphophonological phenomena and thus contributed to typological changes, especially in Estonian and Livonian. The deletion of some Proto-Finnic suffixal consonants, especially intervocalic spirants, has created long vowels and diphthongs in non-initial syllables, especially in Finnish, Karelian, Ingrian and Votian. These long vowels have probably also existed in Estonian and Livonian, as an intermediate stage preceding the present short vowels (e.g. Est. kala ‘fish-prtv’, Liv. ka’ll6 < *kalaa < *kala-da, cf. Fi. kalaa). In Veps, these suffixal consonants (original stops and h) have often survived (cf. Veps. otmha ‘take-inf-ill’ < *otta-ma-han ~ Fi. ottamaan, Est. võtma; Veps. kalad ‘fish-prtv’ < *kala-ta). In all Finnic languages, certain groups of nouns and verbs have “consonant stems”, i.e. the original stem-final vowel — usually e — is deleted before a CV suffix, e.g. Fi. and Est. pese-n ‘wash-1sg’: Fi. pes-tä, Est. pes-ta ‘wash-inf’. On the basis of some rudimentary lexicalizations (e.g. Fi., Kar., Ingr. and Vot. tuttava, Veps.
The Finnic languages 187
tutab, Est. tuttav ‘acquaintance’ < *tunt-ta-pa from tunte- ‘know’) it can be deduced that this phenomenon originally had a wider distribution. As a rule, the main stress in Finnic (as in Proto-Uralic as reconstructed by most linguists) is on the first syllable. The place of the secondary stress may vary due to word-internal quantity relationships, morphological structures and apocope. Some languages also allow the main stress to fall on non-initial syllables in loanwords. This is probably due to very strong foreign influence and bilingualism (e.g. Kar. davai < Russ. davaj) or, also, internal factors like centralization; e.g. Est. kultuur ‘culture’ — in Estonian, an unstressed long vowel would be impossible, cf. Fi. kulttuuri. In most Finnic languages, quantity is phonemic: most vowels and consonants can be either short or long (the length is shown in most orthographies by doubling the letter). E.g. Fi. tule ‘come! [sg]’, tulee ‘(s/he) comes’, tullee ‘(s/he) will probably come’ [potential], tuulee ‘(the wind) blows’. In Estonian, the deletion of unstressed vowels, syllables and even whole suffixes has given rise to a new kind of quantity alternation (in many cases not shown in the standard orthography). Long vowels, diphthongs and syllable-final consonants (or, to put it in another way, long syllables) can, according to the traditional view, be either “half-long” or “over-long”; the latter, so-called 3rd quantity grade, typically appears in word forms where the number of the syllables has been reduced. Thus, the first syllable in linna “town-gen” (< *litna-n, 2nd grade) is shorter than in linna “town-prtv” or “town-ill” (< *litna-da or *litna-hen, 3rd grade), while both are longer than in lina ‘linen’. Some linguists have even wanted to distinguish a fourth phonological quantity grade; they claim that there is a distinction between e.g. the partitive and illative forms in the above example and that the illative form also has a different tone contour. However, there seems to be no indisputable empirical evidence for this. (For references and discussion about the quantity grades and the supposed status of Estonian as a tone language see especially Laanest 1982: 93–95; Hint 1983; Lehiste 1998.) A somewhat similar morphophonological quantity alternation of diphthongs and consonant clusters has also been observed in Livonian, e.g. jaalga ‘foot’ — part.sg jallg6. Livonian also has a Latvian-like system of accents or tones. 1.2.3Morphology and morphosyntax The Finnic languages have a rich morphology, which employs mainly suffixes. Prefixes are used only in a few marginal cases (e.g. Fi. epä-, Est. (< Fi.) eba-, a negative prefix for some adjectives and nouns), with the exception of Livonian, where Latvian verb prefixes (expressing aspect, Aktionsart etc.) are widely used. Most of these suffixes are based on ancient Uralic elements, although a substantial reorganisation of old suffixal elements has also taken place in Pre- and Proto-Finnic. In many instances, Finnic seems to have preserved the original agglutinative character of Uralic morphology. This may, however, be partly a methodological
188 Johanna Laakso
illusion, considering the generally agglutinative tendency of reconstructed protolanguages and some counter-evidence e.g. the supposedly old age of consonant stems and — maybe — gradation. Estonian and Livonian are often mentioned as counter-examples of languages shifting towards the flective type. Phonological developments such as the deletion of unstressed syllables, final consonants and even whole suffixes in certain stem types (Est. kala ‘fish’: kala ‘fish-gen’: kala ‘fish-prtv’ — cf. Fi. kala: kala-n: kala-a), morphologization of gradation in Estonian or umlaut phenomena in Livonian have resulted in morphological stem alternation, e.g. Est. siga ‘pig’: gen sea, Liv. tämm ‘oak’: gen tamm (cf. Fi. sika: sia-n, tammi: tamme-n). However, there are also opposite tendencies recreating agglutinative constructions: e.g. in Estonian, the part.pl in -sid (instead of the old suffixes for plural partitives that had developed into various different forms). Finnic, like Uralic languages in general, has no grammatical gender or other systems of noun classes. The word classes can only be distinguished on the basis of lexico-semantic criteria or (derivational) morphology; in principle, all underived word stems in all word classes are subject to the same phonotactic constraints. 1.2.3.1 Number. Of the three Uralic grammatical numbers, only two, singular and plural, survive in Finnic. The dual, which is reconstructible on the basis of Sami, Ob-Ugrian and Samoyedic, had disappeared by the Proto-Finnic stage. Before Proto-Finnic, the plural markers were probably originally used in predicate verbs (and nouns) and subject nouns (attribute concord is a Finnic innovation); possessive suffixes probably also had variants expressing dual or plural possessum. Plural marking can be considered obligatory in Finnic. It is often assumed that the frequent (although not exclusive) usage of singular instead of dual or plural in some cases, for e.g. paired body parts or garments (‘eyes’, ‘trousers’ etc.), represents an ancient usage. The singular forms are unmarked everywhere. For nouns, the plural suffixes are t and i. Because the other Uralic languages (outside the Finnic-Sami-Mordvin group) use many kinds of plural suffixes, some of them relatively young, there is no clear consensus about the age and original function of the Finnic suffixes. However, Honti (1996) considers both t (an old determinative suffix?) and i (an old collective suffix?) to be ancient plural markers. The suffix t appears in the nom.pl in all Finnic languages (except the southern Estonian dialects, where the plural suffix is [zero] or glottal stop), and also in the 3pl suffix for verbs (the 3rd person forms are, originally, deverbal nouns). It is also historically included in both types of gen.pl in Finnic: the so-called 2nd genitive or “western type” goes back to the nom.pl and the genitive n — e.g. Fi. (dial., arch.) kalain, Est. kalade < *kaladen < *kalat+(e+)n ‘fish-gen-pl’ —, while the “eastern type” consists of the same ending and the old plural stem in *i: Fi. kalojen, Kar. kaloin, Veps. kaloid’en, Vot. kalojee < (*kaloiden <) *kaloi+ten. In Estonian, the
The Finnic languages 189
gen.pl in de has been reanalyzed as a new plural stem for all other cases except nom and part (e.g. kala-de-st ‘fish-pl-elat’); alongside that, archaic “plurals in i” also appear in some instances. In Livonian, the same element appears in Dat. and Transl.-Comit. (e.g. kala-d-6n ‘fish-pl-dat’). With the exception of secondary forms like those mentioned above, i (< *j) is the plural marker for non-final positions. It appears thus in all plural case forms except the nom.pl, the “western type” gen.pl and the forms secondarily based on it (the Estonian de plural and its counterpart in Livonian). Together with the original stem vowels, i has participated in many sound changes, e.g. -a+i > oi: Fi. kaloja, Est. kalu < *kaloida < *kala-j-ta ‘fish-pl-prtv’. In many cases, the deletion of i in unstressed diphthongs or its merger with the stem vowel i has created uniformity between singular and plural and thus given rise to the use of new plural markers, e.g. de and the part.pl in -sid in Estonian, lOi (e.g. pappi-loi-lla ‘priest-pl-adess’) in dialects of Finnish, Karelian and Ingrian, a in Votian and Estonian (e.g. Est. dial. võrg-a-l ‘net-pl-adess’ instead of *võrgul, homophonous with the singular form). 1.2.3.2 Person. Like all Uralic languages, the Finnic languages use three grammatical persons in all (both) numbers. Person markers include person endings on verbs and possessive suffixes indicating the possessor. Person marking on verbs is obligatory (except in the indirect mood in Estonian and Livonian and the conditional mood in present-day Estonian), but there are slight differences in the form and use of some endings. In Livonian, the 1sg endings have been replaced with 3sg endings, which has been attributed to analogy with the imperfect forms (where 1sg and 3sg have merged). In Karelian, Veps, Votian and Ingrian, the “passive” (impersonal) form is (more or less exclusively) used instead of 3pl. This is probably due to Russian influence, because in Russian the 3pl form is also used for the generic, “impersonal” functions usually covered by the Finnic “passive”, e.g. Kar. (hüö) ote-ta-h ‘(they) take-“pass”’. In Finnish dialects and in colloquial Finnish, the same impersonal form in (T)TA is used instead of 1pl, e.g. me ote-ta-an ‘we take’ instead of Standard Finnish me ota-mme. The possessive suffixes have been lost in Estonian and Livonian and tend to be lost, at least dialectally, in (colloquial) Finnish. Their retention or deletion seems to depend on their syntactic position and function (see e.g. Paunonen 1995). Thus it is, in principle, possible that the possessive marking on nouns has never been obligatory in all instances (the Standard Finnish system would thus represent a literary overgeneralization and simplification). In some cases, it could have been used to express more subtle shades of e.g. definiteness (as in some related languages). The person markers seem to have been developed from personal pronouns (at some Pre-Uralic stage?), as indicated by the same consonant elements: *mV for 1st,
190 Johanna Laakso
*tV (ti regularly > si in Pre-Finnic) for 2nd and *sV(n) for 3rd person. In the Finnic system of possessive suffixes, these elements have been mixed with older (ProtoUralic?) elements indicating duality or plurality of the possessum (*n, *k?): sg 1. ni 2. si 3. nsA
pl mme, mmA nne, nnA nsA
By the Proto-Finnic stage, the different suffix series for different numbers of the possessum would have merged, which probably explains the n in 1sg (instead of **mi), the geminate mm and nn in 1pl and 2pl and the n in 3sg nsa/nsä. The person endings on verbs are somewhat similar to the possessive suffixes: sg 1. n 2. t 3. (pA)
pl mme, mmA tte, ttA (bAt)
The 3rd person ending originally represents a suffix for deverbal nouns. The same suffix also appears in participles and some derivatives, as in Fi. syö-pi (dial.) ‘eat-3sg’, syö-vä ‘eat-pres.part’ [the weak-grade alternative, *bA, has been generalized for participles], syö-pä ‘cancer’ (“devouring (disease)”, lexicalized derivative). Especially in Finnish and dialects of Estonian, the 3sg ending has nearly or (in some cases) completely worn off, thus leaving the 3sg form (almost) unmarked. Other Finnic languages still have a certain (generalized) marker for 3sg e.g. Standard Estonian -b. In other moods and tenses than the indicative and the present tense, there is no special 3rd person marker. The only exception is Finnish (Standard Finnish together with most dialects), which has reanalysed-vAt (< *pA + pl. t) as a 3pl marker and extended its use also to the conditional mood and to the preterite (imperfect tense): näk-i-vät ‘(they) saw’ [see-pret-3pl], näk-isi-vät ‘(they) would see’ [see-cond-3pl] — cf. Est. näg-i-d, näe-ksi-d. 1.2.3.3 Tense. The Finnic languages have four tenses: two old ones (the present tense and the imperfect (preterite) tense) and two Finnic innovations (the compound tenses, perfect and pluperfect). The present tense has no synchronically overt marking; however, the person endings for 1pl and 2pl (mmA ~ mme, ttA ~ tte) probably show traces of an original k (thus: < *k + mV, *k + tV). This k also appears in the present tense “passive” (“impersonal”) suffix *(ttA-)k-sen (best preserved in Estonian and replaced with analogical k-less constructions in Finnish, Karelian and Ingrian), e.g. Est. ela-ta-kse ‘they [generic] live’. Since this k in the present tense “passive” forms corresponds to i in the imperfect tense (e.g. Est. ela-ti, Fi. ele-ttiin < *elV-tt-i-hen < *elä+ttA+j+sen ‘they [generic] lived’), it can be (re)interpreted as a present tense marker.
The Finnic languages
The imperfect (preterite) tense marker is i (< *j). Like the plural marker i, it has often caused or participated in sound changes with the stem vowel. In the southern Finnic languages, Votian, Estonian and Livonian, this has given rise to a new imperfect tense marker in s (originally reanalyzed from words where the regular Proto-Finnic sound change ti > si has changed the stem-final t (or its weak-grade counterpart) to s in the imperfect form, e.g. Est. maga-si-n ‘I slept’ < *makat-i-n). Thus, Fi. lu-i-n ‘I read [pret]’ (< *lug-i-n) corresponds to Est. luge-si-n (but the southern Estonian dialects have s-less forms like loi — an exact counterpart of the Finnish form). The compound tenses, the perfect and the pluperfect, consist of the verb ‘to be’ in the present or the imperfect tense, respectively, and the perfect participle: e.g. Fi. ole-n syö-nyt ‘I have eaten’ (be-1sg eat-pf.part), Est. ol-i-n söö-nud ‘I had eaten’ (be-pret-1sg eat-pf.part). These tenses have quite obvious models in the neighbouring languages, although more or less similar compound tenses are also found in more distantly related Uralic languages. In the eastern Finnic languages, at least in Veps, the compound tenses are seldom used, which could be either an archaic feature or a result of Russian influence. In a more general framework, the development and grammaticalization of the compound tenses is also linked to the birth of copular constructions (which, in itself, could be a result of foreign, especially Germanic influence). The use of copula was extended from prototypical predicative clauses to participle constructions (perhaps used in a perfect or anterior function similar to that of the present-day perfect tense). (Cf. Korhonen 1981: 310; Trosterud 1994.) 1.2.3.4 Modality. The grammatical moods in the Finnic languages are the indicative (unmarked), imperative and conditional moods. In addition to these, there is a dialectally restricted potential, and an indirect or relative mood (modus obliquus) which only appears in Estonian and Livonian. The imperative forms employ suffixes containing the element k. The simplest form is 2sg in *-k (in most Finnic languages > ø), with no person marker. 2pl is formed of *kA (or *kO) and (mostly) the 2pl suffix *tA or *te, e.g. Fi. luke-kaa ‘read [pl]!’ < *luke-ka-da, Est. luge-ge < *luke-ka-de, Liv. lu’ggigid < ?*luke-ko-it(V). The 1pl form contains the same imperative suffix and the person ending; in Ingrian and Votian and also in colloquial Finnish and Estonian it is usually replaced with an indicative form (Est. loe-me! ‘read-1pl’, Fi. lue-taan! ‘read-pass’ for ‘let’s read!’ — the “passive” form is used also instead of indicative 1pl in colloquial Finnish), Livonian uses an analytic construction with las ‘let’. The 3rd person forms show *kA or *kO and the suffix *hen (< *sen), e.g. Fi. luke-koon ‘may s/he read!’ (< *luke-ko-hen), luke-koot ‘may they read!’ (an analogical formation with the plural marker t), Est. luge-gu ‘may s/he/they read!’, Veps. lugegaha ‘may s/he read!’ (< *luke-ka-hen).
191
192 Johanna Laakso
On the basis of the distribution of *kA and *kO, especially in Finnish (where kO also appears in a special “optative” in 2sg — mainly in (folk) poetry, e.g. ottaos ‘take!’ < *otta-go-s(i?)), it has been assumed that there has originally been a difference in the function of these two suffixes. The eastern Finnic languages (Novgorod and Tver Karelian, Veps, Ingrian, Votian) show an imperative suffix -ka, -ko etc., obviously borrowed from the Russian imperative clitic -ka; this clitic could, in turn, be a loan from some Uralic language, as Décsy (1967: 155–156) presumes. Analytic constructions, usually employing the verbs anta- or laske- ‘let’, are used in imperative or “optative” (non-prototypical, e.g. 1st or 3rd person) functions in many Finnic languages, e.g. Votian lakka pajatan ‘let me speak’, Est. las (ta) tuleb ‘let him/her come’ (with a finite verb form), Liv. las lu’gg6g ‘may he read!’ (with an imperative form). The conditional marker in Southern Finnic (Estonian, Livonian and partly Votian) is *-ksi-, in other Finnic languages *-isi-, which is historically connected with the diminutive suffix (Lehtinen 1983). The person endings are similar to those of the imperfect (preterite) tense. Another structural similarity is the conditional past tense, which is formed — analogically to the perfect tense — using the conditional form of the verb ‘to be’ and the perfect participle, e.g. Fi. ol-isi-n luke-nut, Est. oleksi-n luge-nud ‘I would have read’ (be-cond-1sg read-pf.part). Veps and the Veps-based dialects of Karelian (Ludian, Olonetsian) use a synthetic conditional past tense, e.g. Veps. luge-nu-iži-n´ (read-pf(.part)-cond-1sg); a similar conditional past tense was also propagated by some Estonian linguists in the time of the Estonian “Language Renewal” (e.g. lugenuksin), but without much success. The primary function of the conditional mood is probably to express the unreality of the proposition; it is also used to express wishes (Est. Oleks mul (ainult) raha! ‘If (only) I had money!’), requests (Fi. Avaisit nyt ikkunan! ‘(I wish) you would open the window now!’) or politeness (Fi. Tässä olisi rouvalle ‘Here’s [lit.: here would be] (something) for the lady’). It is also used in a “conjunctive” function in subordinate clauses — Fi. Hän pyysi, että tulisin ‘s/he asked that I would come’ — and also to express future in subordinate clauses: Fi. Hän tiesi, että tulisin ‘s/he knew that I would come’. The syntactic functions of the conditional thus seem to come close to e.g. Swedish and German models. (Cf. Kauppinen 1998.) The potential marker is ne, related to the conditional markers in other Uralic languages (e.g. Hungarian). The use of the potential seems to be receding in most Finnic languages: Estonian (Standard Estonian and northern dialects) and Livonian have lost it completely (in many parts of Estonia, ne only survives in folk poetry as a semantically empty syllable to fill the metre). In Votian and Ingrian, the use of the potential is obviously dwindling. This might also be one of the general symptoms of approaching language death, although the complete paradigms collected in the 19th century (cf. Laanest 1982: 238–239) are probably not very trustworthy as indicators
The Finnic languages 193
of actual use. The use of the potential mood is strongest in Finnish (especially the eastern dialects) and Karelian. According to usual handbook definitions, the potential mood expresses uncertainty or probability: e.g. Fi. tul-le-e ‘(s/he) will probably come’ (comepot-3sg), southern Estonian saa-nõ-s ‘er [/sie] dürfte bekommen’. The eastern Finnish dialects typically use the potential as a discourse strategy, and in Karelian the use of the potential seems to be connected with syntactic factors like subordination (Forsberg 1998). The indirect mood is used in Estonian and Livonian to express that the speaker has not him/herself seen or heard what s/he reports. Its markers are manifold. Standard Estonian vat (ta/nad tule-vat [s/he/they come-indir] ‘they say that s/he will come’, ‘I’ve heard that they will come’ etc.) can be traced to the present participle suffix va and the partitive ending t. In dialects of Estonian, the present participle in -v or infinitives in TA or ma are also used. Livonian employs -ji (< *-jA, a deverbal noun suffix which forms agent nouns and present participles). The variety of these markers, together with the lack of person endings, shows that the indirect mood is a result of relatively recent grammaticalization. The indirect mood in Estonian and Livonian has been connected with the socalled relative mood in Latvian and Lithuanian (Comrie 1981: 154), although it is difficult to prove any direct influence in either direction. 1.2.3.5 Negation. Most Finnic languages, like most Uralic languages in general, use a special negative verb e- (with more or less regular person endings), while the main verb is in a special form ending in *-k (in most Finnic dialects > ø, in Standard Finnish surviving in some sandhi phenomena), e.g. Fi. e-mme ota-ø ‘we don’t take’ [neg-1pl take-(conneg)]. Exceptions are Estonian, where the 3sg form ei (< *e-pi) has been generalized to all persons, and Livonian, where the plural negative forms also inflect the main verb. The negative verb usually precedes the main verb, the most notable exception being Southern Estonian (Savijärvi 1981). The negative verb, however, is defective: in most Finnic languages it is inflected only in the present tense and the indicative mood. The imperative forms have a different stem, äl(A)- ~ el(A)- ~ al(A)- ~ är(A)-, and the main verb is often in a different form peculiar to negative imperatives, e.g. Fi. äl-kää teh-kö ‘don’t [pl.] do!’ (neg.imp-imp.2pl do-imp.conneg), or else in a form identical with the corresponding non-negated form, e.g. Est. är-ge teh-ke (neg.imp-imp.2pl do-imp.2pl). In almost all Finnic languages the negative verb has lost its original past tense inflection. In Southern Estonian, the past tense forms still survive in the form es for all persons (originally 3sg), and even complete past tense paradigms have been collected — although only in one dialect, Kodavere in Eastern Estonia. The loss of the past tense (and, analogically, perhaps the conditional) forms of the negative verbs is connected with the development of copular sentences and the perfect and
194 Johanna Laakso
pluperfect tenses (Korhonen 1981: 310; Trosterud 1994). Now the preterite forms usually consist of the inflected negative verb (in the present tense) and the main verb as a perfect participle, e.g. Fi. ei teh-nyt, Est. ei tei-nud ‘(s/he) didn’t do’ (neg.3sg do-pf.part); the perfect and pluperfect forms similarly negate the auxiliary ‘to be’, e.g. Fi. e-n ole-ø teh-nyt ‘I haven’t done’ (neg-1sg be-conneg dopf.part), e-n ol-lut teh-nyt ‘I hadn’t done’ (neg-1sg be-pf.part do-pf.part). In non-finite constructions, negation can be expressed with adnominal negative (“caritive”, “abessive”) suffixes, e.g. Fi. hän lähti syö-mä-ttä “s/he left without having eaten” (s/he left eat-inf-abess). One of the very few Finnic prefixes also serves lexical negation: Fi. epä-, Est. eba- (originally, the 3sg — or, formally, the present participle — of the negative verb), e.g. Fi. epätavallinen ‘unusual’, Est. ebaõnn ‘misfortune’. 1.2.3.6 Passive and related phenomena. Finnish grammatical tradition uses the term “passive” for a form category that is not fully compatible with the Indo-European passives: the agent is left out but the object remains syntactically an object (its case, similarly to objects of imperative verbs, is either nominative(-accusative) or partitive), e.g. Fi. Omena-t syö-tiin ‘(all) the apples were eaten’ (apple-nom.pl eat“pass”.pret), Est. Kiriku-t ehita-takse ‘a/the church is being built’ (church-prtv build-“pass”.pres). Agents, if allowed at all (constructions with agents are often judged to be ungrammatical or questionable, clumsy or foreign-sounding, although Estonian seems to allow more and more agent constructions — cf. Tommola 1993), are usually expressed with constructions employing postpositions like Fi. toimesta (“through the action of”) or Est. poolt (“from the side of”). Finnic “passives” are also frequently formed from intransitive verbs, e.g. Fi. Yläkerrassa nukutaan [upstairs-iness sleep-“pass”.pres] ‘Upstairs there are people sleeping’. Because of these differences between the Finnic “passive” and the passives in other languages, alternative terms have also been suggested. Estonian grammarians use “impersonal voice”, while, in Finland, linguists have proposed terms like “indefinite”, or — because the use of the “passive” implies that the hidden agent is a human being — “ambipersonal” or “4th person” (Shore 1986; Tommola 1993). The morphological markers of the Finnic “passive” are TTA or TA, originally a causative verb suffix (Lehtinen 1984), followed by the tense marker (*k for present tense, except in Finnish and Karelian, where analogical k-less constructions are used; *i for past tense) and the suffix *sen ~ *hen (historically related to the 3rd person pronoun hän). Examples of this are Fi. syö-tiin < *söö-t-i-hen (eat-“pass”pret-“3sg”) and Est. süü-a-kse < *söö-dä-k-sen (eat-“pass”-pres-“3sg”). The perfect and pluperfect forms consist of the auxiliary ‘to be’ in 3sg (or in the “passive”) and the passive perfect participle, e.g. Fi. on syöty (~ coll. ollaan syöty), Est. on söödud ‘(it) has been eaten, somebody has/they have eaten’ (be-3sg.pres eat-pass.pf.part).
The Finnic languages 195
It is often assumed that Uralic originally had a personal passive (like that still found in the Ob-Ugrian languages, cf. Kulonen 1989) in W. This suffix (> Finnic u, ü) or complex suffixes with W are now used in lexical verb derivation to express reflexive, passive or related functions (e.g. Fi. alentu- ‘to debase or humiliate oneself ’ alenta- ‘to lower’, Kar. n´ägü- ‘to be seen, to be visible’ n´äge- ‘to see’, Est. sulgu- ‘to be closed’ sulge- ‘to close’). (The Estonian derivatives in u are often relatively young; Hint (e.g. 1991: 143) suspects Russian influence in their increasing popularity, but it may be due to general tendencies towards more synthetic structures — these were favoured by language planners, cf. Raag 1998 — and even Finnish models.) Another suffix used in reflexive, passive or related functions is TE (TTE). In Western Finnic languages, it only appears in a few lexicalized, obscure derivatives (with no known base), e.g. Est. lange-, Fi. dial. lankee- ‘to fall’ < *la]ke-de-. In eastern Finnish dialects, Karelian, Veps, Ingrian and partly Votian, the same suffix is also used more extensively to form reflexives from transitive verbs, e.g. Fi. dial. muuttain (< *muutta-de-n) ‘I change myself (into another shape)’, Kar. keändäi (< *kääntä-de-k) ‘turn (yourself) around!’. The geographical distribution of this “reflexive conjugation” shows that Russian influence has at least partially contributed to its development. (Cf. especially Koivisto 1995.) 1.2.3.7 Case. Although many Uralic languages are now famous for their elaborate case systems, the abundance of noun cases need not be an ancient feature. Many of the cases in Finnic and e.g. Hungarian, are relatively young, and some Uralic languages, notably the Ob-Ugrian languages, have very few cases. The Uralic languages extended the original case system by developing new series of local cases on the basis of already existing suffixes. A typically Uralic feature is the exploitation of the tripartite system of “to”-, “in”- and “from”-cases, but there are also many examples of new cases brought forth by agglutination of postpositions (see especially Korhonen 1996 [1992]). In a wider framework, the tendency to develop elaborate case systems is linked to the fact that, as Eeva Kangasmaa-Minn (1993) points out, the Uralic languages typically use noun morphology to express meanings that many other languages encode in verb morphology and syntax, most notably as aspectual shades. 1.2.3.7.1 Grammatical cases. The nominative is always unmarked. As well as the subject, predicatives and objects of imperative and “passive” verb forms can also appear in the nominative. The genitive is used to designate the possessor in a wide sense. It is also used in attributes of nominalized verb forms, for both the object and the subject of the underlying sentence, which makes constructions like Fi. sotilaiden ampuminen ‘the shooting of/by the soldiers’ ambiguous. In compound words, Estonian in particular
196 Johanna Laakso
favours genitive-initial constructions even in cases where the denotee of the first part is not strictly a possessor, e.g. Est. lume+memm [snow-gen-old.woman] ‘snowman’ — cf. Fi. lumi+ukko [snow.nom-old.man] ‘id.’. The gen.sg suffix n has regularly disappeared (together with other word-final n’s) in Estonian, Livonian and some dialects of other Finnic languages. For gen.pl there are two competing suffixes, which is one of the main criteria in dividing the Finnic languages into the Western and Eastern groups. The “western type” consists of plural t, a fill-in vowel and the genitive n, e.g. Fi. dial./arch. kalain, Est. kalade < *kala-den < *kala-ten. In the “eastern type” the same complex suffix has been attached to a plural stem in i (e.g. kaloi < *kala-j), e.g. Fi. kalojen < *kaloi-den. The old accusative in *-m has — due to a regular Finnic sound change: wordfinal -m > -n — coincided with the genitive (except in the personal pronouns, where there are different forms), and thus some modern grammars do not distinguish a special accusative case. It is used in total (resultative) objects (except in “passive” or imperative sentences, where the corresponding objects are in the nominative case). The partitive suffix tA was originally a local (“from”) case ending, as some lexicalized adverbs still show. Its development to a case for partial or non-resultative objects obviously began in the Pre-Finnic stage, since Mordvin also has some examples of partial objects in the “from” case. In present-day Finnic, the partitive is used to indicate indefiniteness, partiality (in the case of uncountable or plural objects) or non-resultativeness — e.g. Fi. syö-n leipä-ä, Est. söö-n leib-a (“eat-1sg bread-prtv”) ‘I eat bread (i.e. some bread, not the whole loaf)’, Fi. ammu-n lintu-a (“shoot-1sg bird-prtv”) ‘I fire at the bird’ (cf. ammun linnu-n [bird-acc] ‘I shoot the bird (dead)’). It is also used for objects of negated verbs. The use of the partitive vs. the accusative thus serves to express aspectual or even temporal shades (the accusative together with the present tense verb form often actually implies that the action will be fulfilled in the future). The partitive seems to be developing into a general low transitivity marker (cf. Helasvuo 1996), especially in the north. Finnish also uses the partitive for “partial” or “generic” predicatives with plural or uncountable subjects (puuro on kuuma-a ‘the porridge is hot’ [hot-prtv], norsut ovat viisa-i-ta ‘elephants [in general] are wise [wise-pl-prtv]’) and indefinite (plural) subjects (poik-i-a juoksee pihalla ‘there are boys running on the yard’ [boy-pl-prtv run-3sg.pres yard-adess]). The corresponding words in Estonian would be in the nominative case. 1.2.3.7.2 Local cases. The Finnic languages have two tripartite — “to”, “in”, “from” — series of local cases: “internal” (“inside”) and “external” (“outside”, “upon”, also used in more “grammaticalized” functions, to express possessor, beneficiary, instrument etc.). The case suffixes consist of reconstructible Proto-Finno-Ugrian
The Finnic languages 197
local case endings: locative nA, separative tA, lative *n, *s or *k — (still surviving in lexicalized adverbs), preceded by s (internal) or l (external): Proto-Finnic inessive ssA (< *s + nA) elative stA illative sVn
This case system as a whole is probably a Proto-Finnic innovation, although the s-cases have counterparts in Sami and Volgaic, and the l-cases in the Permian branch of Uralic. A noteworthy exception to the system reconstructed above is in Votian, where the inessive ending is -zA (< *sA). In Estonian, phonological developments have resulted in a split of the original illative into two cases: the old illative type, where the ending has (almost) completely worn off, has been replaced by a new illative with an analogical ending -sse. It does survive, in some cases, as a special “aditive”, expressing concrete locality (e.g. tuppa ‘into the room’ — toa-sse in structures requiring the illative like mis puutub toasse ‘concerning the room…’). 1.2.3.7.3 Other cases. The all-Finnic cases also include the translative (in *-ksi, *-kse-), which expresses “becoming, turning into something”, e.g. Kar. jäi leššekši ‘became a widow’ (in Livonian it coincides with the comitative, also in -ks), the essive (formally: old locative in -nA), which expresses “being, acting as”, e.g. Votian lahsõna ‘as a child’, and the abessive in *-ttA(k), e.g. Fi. leivättä, Veps. leibätä ‘without bread’. Post-Proto-Finnic developments have given rise to new cases in some Finnic languages. Many of these are originally agglutinated postpositions, like Estonian and Votian comitative (dialectally in other Finnic languages, too), e.g. isa-ga ‘with the father’, cf. e.g. Fi. isän kanssa; the Estonian comitative is also used in an instrumental function. Before the original postposition, the original case ending (genitive, partitive, local case) is often visible. Most new cases of this type appear in Veps and dialects of Karelian, e.g. Veps. stola-n-no ‘to the table’, mama-lon ‘at mother’s place’, ma-d-med ‘along the earth’, kormana-s-pai ‘from the pocket’, tatake ‘with the father’, eht-ha-sai ‘until evening’ (cf. Tikka 1992). In addition to agglutinated postpositions, there are also some (more or less marginal) case endings which have been formed analogically, e.g. extending the existing model of local cases (thus the “exessive” in Fi. dial. koto-nta ‘from home’, Vot. karjuši-nt ‘from the state of being a shepherd’ is based on the essive in -nA and the other “from”-cases in -tA). Some case endings obviously consist of old suffixal elements but have developed to case status only in some Finnic languages, like the Estonian terminative in -ni (e.g. põlvi-ni ‘up to the knees’; cf. adverbs like Fi. ku-ni ‘until’); some suffixes are still (semi-)productive adverb suffixes rather than case endings, e.g. the all-Finnic instructive (e.g. Fi. sormi-n, Vot. sõrmi-i ‘with fingers’, Fi.
198 Johanna Laakso
jala-n ‘on foot’) and the so-called prolative in e.g. Fi. puhelimi-tse ‘by phone’, Ingrian meri-tse(g) ‘by sea’. As for the origins of the instructive, it has also been associated with the genitive and also the dative-genitive present in Livonian (and, in the syntactic use of genitives, perhaps in (western) Finnish). The function of the Livonian dative has probably been influenced by Latvian (Halling 1997). 1.2.3.8 Comparison. All Finnic languages have a comparative suffix *-mpA (e.g. Fi. suure-mpi ‘greater’, gen. suure-mma-n); in Livonian comparison is mostly expressed with analytical constructions. The suffix has cognates also in more distant FinnoUgrian languages, but because their function (except the Hungarian comparative suffix -bb!) is not completely identical with the Finnic-Sami suffix and because other Finno-Ugrian languages use other suffixes or constructions to express comparison, mainstream Finno-Ugrists have considered the comparative in its present form a (Pre-)Finnic innovation. The superlative in *-in (< *mA): *-impA- (analogous with the comparative) only appears in Finnish and Karelian (and, as a loan from Finnish, in Estonian: -im); other Finnic languages use analytic constructions. 1.2.3.9 Word formation. The Finnic languages have a rich system of suffixal word formation. It is obviously an ancient inheritance. Practically all derivational suffixes are indigenous and have cognates in other Uralic languages; among the allFinnic suffixes there is only one indisputable loan, the noun suffix Fi. -niekka — e.g. parta-niekka ‘bearded (man)’ —, Kar. -n´i(e)kka, Est. -nik etc. < Russ. -nik (although more recent derivational suffixes in individual languages like Finnish include some loans, cf. e.g. Hahmo 1997). Foreign influences can be suspected in the use of some derivational suffixes and subsystems, notably the reflexive verbs in eastern Finnic languages. The derivation system in the Finnic languages is more or less uniform in central respects (for all-Finnic comparisons see e.g. Markianova 1987; Laakso 1989). In Estonian and Livonian the general suffix-deleting sound changes have “impoverished” or reshaped the derivation system, as can be seen from e.g. comparative studies of Estonian and Finnish derivation (Kasik 1989; Ojutkangas 1997); in a wider framework, this could be interpreted as areal tendencies towards greater derivational diversity in the east and more compact systems in the west (i.e. also in Western Finnish dialects). 1.2.3.9.1 Noun formation (general). Laanest (1982) lists 26 “primary” (non-complex) and 13 complex all-Finnic noun suffixes. For some of these, the suffix status is synchronically very questionable, because they only appear in lexicalized formations
The Finnic languages 199
with no known synchronic base. Some suffixes form both verb-based and nounbased nouns; many suffixes form both nouns and adjectives. 1.2.3.9.2 Deverbal noun formation. The abundance of non-finite verb forms can be seen as part of the previously-mentioned Uralic tendency to encode syntactic and semantic features especially in noun morphology, which makes the verb morphology more labile and prone to various developments. All Finnic languages have two series of infinitive forms: forms in *-tA- and *-mA-. The all-Finnic tA-infinitives include the usual “dictionary infinitive” in *-tA-k (originally: lative) and gerundive (inessive or instructive) forms, e.g. Fi. hän kulki laula-e-n ‘s/he wandered, singing’ (sing-inf-inst), Est. joos-te-s ‘running’ (run-inf-iness). The suffix -mA- is used together with local and abessive case endings, e.g. Fi. tuli anta-ma-an ‘(s/he) came (in order) to give’ (give-inf-ill), Kar. magoa-ma-šta ‘from sleeping/lying’ (sleep/lie-inf-elat), Ingr. kündä-mä-däg ‘without ploughing’ (plough-inf-abess). The all-Finnic non-finite verb forms also include present participles in *-pA (active; identical with the pres.3sg ending) and *-tta-pA (“passive”) and past participles in *-nUt (active) and *-ttU (“passive”). Other suffixes for verb nominalization include -mA, -ntA, -Us and -minen (nomina actionis, -minen is, in some languages, also used in modal constructions), *-ek (nomina acti/instrumenti), -jA (nomina agent) etc. 1.2.3.9.3 Verb formation. The denominal verb suffixes sometimes have clearly definable semantic functions (e.g. -ne- forms verbs denoting change towards a state denoted by the stem). Sometimes their only analyzable function is to form verbs out of nouns while other features of the derivative are only conditioned by the stem and other (e.g. contextual) factors (Kytömäki 1992; Laakso 1997). The syntactic and semantic functions of the deverbal verb suffixes are difficult to analyze, because they are intertwined with various semantic factors; these suffixes are sometimes used to express phenomena related to aspect and Aktionsart, but often not in a systematic or clearly definable way. The “modifying” verb suffixes are usually interpreted as expressing momentaneity (e.g. Fi. hyp-ähtä- ‘jump (once, quickly)’), frequentativity (hypp-ele- ‘jump (many times, lightly)’) or continuativity (hypp-i- ‘jump (for a longer time)’). In eastern Finnic languages, Russian influence has probably encouraged or facilitated the usage of inchoative verb suffixes (Lehtinen 1985), and Russian influence can also be seen in the wide usage of frequentative or momentane verb suffixes in expressing aspect phenomena (Lehtinen 1990). Among the verb-based verb derivatives, the most clear-cut subgroups are the causative/factitive/curative verbs (TTA and related complex suffixes) and the reflexive verbs (e.g.*U, *TE and related complex suffixes). For both groups, the area
200 Johanna Laakso
of strongest and widest usage seems to be in the easternmost Finnic languages (Eastern Finnish dialects, Karelian, Veps, Ingrian). In these languages, reflexive derivation has extended its scope (relatively recently, as shown by the multitude of suffixes) and even formed a distinct “reflexive conjugation” (see especially Koivisto 1995); Russian influence has probably contributed to this development. 1.2.4The lexicon 1.2.4.1 Parts of speech. Older handbooks regard certain phenomena as evidence for an original lack of a word-class (or a noun/verb distinction) in Proto-Uralic, for example: – –
–
the existence of certain ambiguous “nomen-verbum” word stems, (e.g. Fi. tuule- ‘wind; (of the wind) to blow’); some instances of noun comparation (e.g. Veps. mamš ižandemb mužikad ‘the woman is “the man in the house” more than her husband’ [“woman mastercomp man-prtv”]) and the fact that many derivational suffixes can be attached to both nouns and verbs.
This view is obsolete and unfounded, also in the light of Finnic data. Although there are no phonotactic criteria for distinguishing e.g. nouns from verbs, word class is lexically determined and the cases of real conversion are few and marginal — except, perhaps, in Estonian, where changes deleting original suffixes have created many instances of more or less genuine-looking noun-to-verb (or verb-to-noun) conversion (cf. Vare 1993; Laakso 1997: 276–278). Nouns and verbs as word classes in Finnic are thus an ancient inheritance, while the adjectives are more difficult to distinguish from nouns (adjectivity is often indicated by the means of derivational morphology, and there seems to be at least one all-Finnic derivational suffix, *-edA — e.g. Fi. pimeä ‘dark’ — with no function other than acting as an adjective marker). The adpositions and adverbs are mostly based on (native) nouns, often inflected in local cases (original or present-day), e.g. Fi. alla ‘under’ < *al-na (lower side-ess). Like other Finno-Ugrian languages with the exception of the Ugric branch (cf. Seilenthal 1981), Finnic languages can use identical word forms as both adpositions and adverbs (e.g. Fi. taka-na ‘at the back; behind’). While other Finno-Ugrian languages use postpositions, Finnic languages also have some prepositions (the same word can often be used in both positions). The head word of the adpositions is often in the genitive or partitive. Conjunctions as a word class seem to be a Pre- or Proto-Finnic innovation. It is usually assumed that Proto-Uralic had no conjunctions but used e.g. asyndetic structures or left-branching nominalizations instead of present-day dependent clauses. The Finnic conjunctions have no cognates in other branches of Uralic: they are either more or less transparent formations, based on native (pronoun) stems (e.g. Fi. e-ttä, jo-tta ‘that’ ¨ pronoun stems e-, jo-), or loanwords: Germanic (e.g.
The Finnic languages 201
all-Finnic ja ‘and’), Russian (in eastern Finnic languages, e.g. Kar. Veps. Ingr. Vot. da(i) ‘and’) or Latvian (in Livonian, e.g. bet ‘but’). 1.2.4.2 The etymological origins of the Finnic vocabulary. Finnic has preserved a few hundred (around 300?) word stems of clearly Proto-Uralic or Proto-Finno-Ugrian origin. In addition to these, the all-Finnic lexicon includes several loanword strata and many all-Finnic derivatives (e.g. emäntä ‘hostess, mistress’ ¨ emä ‘mother’, a Uralic word). Much of the all-Finnic word stock has no etymology as yet. In principle, words could be loans from an unknown, extinct “Proto-European” language (as Ariste 1981 presumes), or Indo-European. New Indo-European etymologies have been found (especially by Jorma Koivulehto) for many words previously included in this group. Many hitherto unetymologized words could also be explained as obscure or irregular derivatives (cf. especially Koponen 1998; Nikkilä 1998). The oldest loanword strata, according to most recent research, are the ancient loanwords from a language form very close to (or identical with) Proto-IndoEuropean. The contacts with Indo-European languages form a continuum up to the present day. Even the earliest contacts seem to have taken place near the Baltic Sea, because many early IE loanwords only appear in the Finnic(-Sami-Mordvin) languages. (Koivulehto 1997, 1999) While some (Pre-)Baltic or (Pre-)Germanic loanwords also appear in other branches of Uralic, the specifically Finnic loanword strata were traditionally thought to begin with the bulk of Baltic and Germanic loanwords. In the light of recent research, the Germanic contacts seem to have begun already around 1500 BC, in the Nordic Bronze Age (Koivulehto 1997 – but cf. Fromm 1997), and, — at least for western Finnic languages, — continued, with later Germanic languages, up to our days. In addition to technical (e.g. [Fi.] rauta ‘iron’ < Germanic) and sociocultural (e.g. [Fi.] kuningas ‘king’ < Germanic) terminology, words for body parts (e.g. hammas ‘tooth’ < Baltic) and kinship terms (e.g. morsian ‘bride’ < Baltic) were also acquired. The latest all-Finnic loanword stratum is Slavic (i.e. (mostly?) Old Russian) and includes e.g. the culturally interesting words pappi ‘priest’ and risti ‘cross’. — Considering the existence of ancient, even PIE loanwords, the Germanic, Baltic and Slavic contacts must be seen as part of a long continuum of IE contacts. 1.2.4.3 Post-proto-Finnic developments in the vocabulary of the Finnic languages. From the early Middle Ages on, the Finnic languages can be roughly divided into two groups: eastern and western. In the west, Estonian and Finnish were mainly influenced by the superstratum languages German and Swedish, respectively. Apart from continuing the influx of Germanic loanwords into Finnic, these languages also mediated most “international” words (i.e. also words of e.g. Latin or French origin) into Finnish and Estonian. For Livonian, Latvian — after German — has probably
202 Johanna Laakso
had a similar role. In the east, the Karelians, Vepss, Votes and Ingrians belonged to the Russian (Novgorod) cultural sphere and the Orthodox Church and received most of their new loanwords from that direction. To this simplified picture we must add the complicated internal relations between Finnic languages and dialects. Eastern Finnish dialects, themselves descendants of the same proto-dialect as Karelian, have had lexical contacts with (Northern) Karelian; the southern Karelian dialects bordering to Finnish have also been influenced by Finnish. In Votian, phonological criteria prove a great part (maybe even one third, cf. Itkonen 1983) of the vocabulary to be loans from Ingrian or Finnish. Ingria, in general, is an area of messy interrelations, bilingualism and Sprachbund phenomena. The northeastern coastal dialect of Estonian shows many Finnish loanwords (Söderman 1996), and Livonian has loanwords from Estonian. In the north, speakers of the easternmost Finnic languages (Veps, Karelian) or their predecessors had contacts with the Komi, as shown by some Finnic loans in Komi (mostly the northwestern dialects) and, for a couple of words, even Udmurt (Hausenberg 1985). There are very few, if any, Komi loans in Karelian and Veps, but more could perhaps be found with systematic investigation. A somewhat more important loanword stratum in Finnish (especially but not exclusively the northern dialects) and Karelian consists of Sami loanwords. In the south, Estonian has some Latvian loanwords. In modern times, the rise of literary standard languages has brought forth many innovations in the vocabulary, especially of Finnish and Estonian. The most important ways of creating new words have been derivation, compounding and adopting dialect words; Standard Finnish has also taken up Karelian words (from folk poetry, especially the national epic Kalevala, which is perceived to belong to the Finnish language). The planners of Modern Standard Estonian borrowed many words from Finnish, and later contacts — e.g. tourism and Finnish mass media as the Estonians’ “window to the west” during the Soviet era — brought Finnish loanwords especially to colloquial Estonian. Finnish is an obvious model or source for neologisms for the new-born (or reborn) Karelian and Veps literary languages, too. Estonian also has many completely artificial word stems. Most of them were formed during the Language Renewal movement at the beginning of this century (especially by Johannes Aavik), in order to create shorter (non-compound) and aesthetically more acceptable expressions for important concepts, e.g. relv ‘weapon’ (instead of older sõjariist ‘war-tool’), küülik ‘rabbit’ (instead of kodujänes ‘home-hare’).
The Finnic languages 203
2.
History and development of the Finnic languages
2.1 The Uralic roots and the relationship to (Proto-)Sami While Finnic clearly belongs to the Uralic language family as an independent branch, its relationship with other branches of Uralic is not completely clear. Its close relationship with Sami and Mordvin does not necessarily imply a common intermediate protolanguage (“Proto-Volgaic-Finnic”), as assumed traditionally; at least some lexical and other features shared by the three or four (north)western branches of Uralic — Sami, Finnic, Mordvin and, to a lesser extent, Mari — could perhaps also be explained as contact phenomena, as Terho Itkonen (1997) has recently suggested. This problem is closely connected with questions concerning the Uralic Urheimat. On the basis of early IE loanwords in Uralic it could be placed somewhere near the middle part of the river Volga (one of the traditional candidates for the Uralic ancestral home); because some early IE loanwords only appear in Finnic(-Mordvin-Sami) and there is no archaeological evidence for a substantial migration from the east to the present-day Finnic areas, the Uralic Urheimat can perhaps be localized as a large area stretching from the Volga to the Baltic Sea (Koivulehto, forthcoming). This would make Finnic or its Uralic protoforms the first known language spoken in these areas. In any case, the possible Uralic (cf. Sammallahti 1995) or non-Uralic precedents of Pre-Finnic, if any, must have disappeared completely, so that hypotheses like Ariste’s (1981) assumptions about a “Proto-European” lexical substratum in Finnic can never be verified. The relationship between Finnic and Sami is likewise problematic. Though the Sami people are genetically different from the Finns and Karelians, their language can, without great difficulties, be traced back to a reconstructed Sami-Finnic protolanguage, also called (Early) Pre-Finnic (Fi. varhaiskantasuomi). Considering the fact that some Finnic structural innovations and many IE (Germanic or Baltic) loanwords have no direct counterpart in Sami, it remains to be shown whether –
–
the deviant features in Sami (e.g. differences in the basic vocabulary) are retentions from an original (Uralic or non-Uralic) substratum language which the Sami later gave up to adopt Pre-Finnic (or to accept a slow Finnicization of their Uralic idiom), or whether the Finnic peoples are actually “Indo-Europeanized Sami”.
In any case, the Sami people, the original inhabitants of (most of) Finland and Karelia, have continuously been receding or subject to Finnicization, and their assimilation has probably left its traces in Finnish and Karelian — these traces still need a thorough investigation.
204 Johanna Laakso
2.2 Early contacts: Finnic as “Indo-Europeanized Uralic”? Many of the features that distinguish Finnic from other branches of Uralic can be ascribed to IE — (Pre-)Germanic or (Pre-)Baltic — influence. In addition to a more or less steady influx of IE loanwords from Proto-IE up to the present, there are (morpho)syntactic features: attribute concord and the use of copula, SVO word order (while other Uralic languages typically have SOV) and the existence of some prepositions alongside the Uralic-type postpositions. The phonological changes leading to Proto-Finnic (simplification of the consonant system, possibly the gradation) have earlier been pointed out as IE influence by Posti (1954). Together with the “western” genetic features of Finnic speakers, this implies that the IE-Finnic contacts must have been quite deep, including bilingualism, mixed marriages etc. As there are no traces of corresponding Finnic influence in (Pre-)Baltic or Germanic, it is assumed that the IE population(s) in question lived among the Pre-Finnic speakers and were later linguistically assimilated. The earliest such populations could be the Battle-Axe culture (2500–2000 B.C.). 2.3 Proto-Finnic dialects and their descendants The present-day Finnic languages cannot be divided into groups each descending from only one Proto-Finnic dialect. Instead, the Finnic language area is split at least by two different isoglosses. Features distinguishing southern Finnic languages from northern ones include e.g. the mid-vowel i present only in Estonian, Livonian and Votian and the use of the 3sg pronoun tämä (Est. tema) instead of northern hän. On the other hand, Estonian, Livonian and Western Finnish dialects can be distinguished from other Finnic languages on the basis of e.g. the genitive plural ending or the reflexive verbs: derivatives in U and related suffixes in the west vs. reflexive conjugation in the east. Both these divisions can also be supported by lexical criteria. These two different and obviously old groupings — north-south vs. west-east — can be combined by operating with three (or more) Proto-Finnic dialects: southern, eastern and northern Finnic (Itkonen 1983). The eastern Proto-Finnic dialect would have no immediate descendants but its influence would explain the common features in Eastern Finnic languages — Veps and Proto-Karelian (the protolanguage of Karelian, Ingrian and Eastern Finnish dialects) and partly Votian. Another tripartite division is presented by Viitso (1998), who considers Livonian and southern Estonian to be radically different from other Finnic languages (these, in turn, are genetically divided into a southern and northern group and so on). Problematic points in all these groupings are the position of hypothetical dialects (e.g. the language probably spoken east of Lake Peipsi, before the Russification of that area, and its relationship with Votian and eastern dialects of Estonian) and the
The Finnic languages 205
W Finnish
E Finnish
Karelian Ludian Vepsian Ingrian
N Estonian
Livonian
Votian
S Estonian
Figure 1.
interrelations of the languages of Ingria, where the two borders cross and the contacts have been constant and deep. In any case, Proto-Finnic cannot have been a homogenous language form; its different dialects were spoken in a large area, isolated but also connected by e.g. the Gulf of Finland, so that innovations can have spread between dialects already clearly differentiated. This means that the development leading from Proto-Finnic to present-day Finnic languages is too complicated to be explained in terms of a traditional family-tree description. The problems with all binary (or even tripartite) genetic taxonomies of Finnic languages are clearly summarized by Salminen (1998): given the possibility of innovations spreading, it is unrealistic to posit dialect or language splits on the basis of one or two sound changes. Many Finnic languages can be seen as results of dialect admixture, as descendants of more than one Proto-Finnic dialect. Thus Southwestern Finnish dialects show influences from Estonian; Eastern Finnish dialects are, essentially, Fennicized (Proto-)Karelian, and Votian has its roots both in the east and in the west. In the east, Ludian and Olonets Karelian can be characterized as “Karelianized Veps”. The most recent examples of admixture are, perhaps, the Estonian and Finnish standard languages. Standard Finnish was created in the 19th century and can be seen as a partly artificial amalgam of materials from different dialects (cf. Paunonen 1993). The creators of modern Standard Estonian, besides adding Southern Estonian words to its Northern basis, also resorted to Finnish models when deGermanizing their language (Finnish, as phonologically more archaic, was perhaps felt to be “purer”). They borrowed, (or tried to borrow) not only words but also elements (like the superlative, cf. Raag 1998), syntactic constructions and word
206 Johanna Laakso
order patterns (SVO order in dependent clauses in contrast to the Germansounding verb-final order). 2.4 Present-day Finnic languages and their dialects: The scattered links
of the broken chain The border between Eastern and Western Finnic also divides Finnish into its two primary dialect groups: Eastern and Western. Western Finnish, especially the southwestern dialects, formed the basis of the oldest written Finnish, but the rise of Finnish to official status in the 19th century, together with the so-called “battle of dialects”, resulted in a new standard language containing features from both western and eastern dialects. (The spoken form of this standard language is now giving way to colloquial variants — so-called “local colloquial languages”, based on but different from regional dialects in the traditional sense.) In the east, the differences between Finnish and North Karelian dialects are very small (considering that Eastern Finnish and Karelian are probably descendants of the same protoFinnic dialect, one could speak of a dialect continuum), while Finnish and southern Karelian cannot be considered mutually intelligible. In Karelian, the differences between the main dialects — northern or White Sea Karelian, southern Karelian proper (and Tver’ Karelian), Olonetsian (Aunus) and Ludian — have been considered great enough to justify the creation of at least three literary languages in the last few years. These lexical and phonological differences (voiceless stops and sibilants in North Karelian vs. voiced ones in the other dialects, word-final A > U in certain cases in Olonetsian, differences in gradation etc.) make the main dialects mutually almost unintelligible. As mentioned earlier, Ludian can also be regarded as a separate language, as a transition dialect between Karelian and Veps. For a long time, Ingria, the area south of the eastern end of the Gulf of Finland, was an important contact area where influences coming from different Finnic languages met. The Russian expansion, the founding of St. Petersburg in the beginning of the 18th century, and, finally, the catastrophes of the 20th century (mass deportations and other repressive measures by the Soviet authorities, World War II with the siege of Leningrad and the evacuation of the remaining Finnic population from Ingria) definitely broke the chain of Finnic dialects around the Gulf of Finland. Now the Finnic languages of Ingria — Ingrian and, especially, Votian — are on the verge of extinction. Two dialects of Votian deserve special attention: the dialect of Kukkuzi with its many Ingrian-like features (e.g. lack of the mid-vowel), and the so-called Kreevin dialect spoken near Bauska in Southern Latvia until the 19th century. The Krevins were descendants of people brought from Ingria as prisoners-of-war in the 15th century (for a thorough analysis and interpretation of the few remaining texts see Winkler 1997).
The Finnic languages 207
Estonian dialects can be divided into two main groups: northern and southern. Due to both linguistic and political factors, both dialects had their own literary languages (the so-called languages of Tallinn and Tartu), until Modern Standard Estonian started to emerge, based on the northern language, in the 19th century. The differences between northern and southern Estonian are great and salient (including central morphological features like the endings of the inessive or the plural). There is a movement claiming separate language status for the most markedly different, southeastern Viru dialect (võro kiil) and attempting to create a unified literary language for it. In the southern dialect group, the northernmost Tartu dialect and the westernmost Mulgi (or Viljandimaa) dialect (cf. Pajusalu 1996) show more “northern” features. — Many Estonian linguists have distinguished the northeastern coastal dialect as the third main dialect of Estonian. In their view, its Finnish-like features (e.g. lack of apocope and syncope) would be ancient heritage rather than Finnish (or Ingrian) influence; however, it seems that at least its lexical characteristics are largely explicable by contacts with Finnish (Söderman 1996). Livonian, originally spoken in large areas of present-day Latvia, survived to the 20th century only on a narrow strip of coast at the tip of the Courland peninsula; the dialect spoken in Livonia proper, east of the Riga Bay, died out in the 19th century. Although isolated by land, the Courland Livonians had regular contacts with the Estonians of the Saaremaa island by sea. The remaining speakers of Courland Livonian now live scattered throughout Latvia. The Finnic area can thus be regarded as a dialect chain broken in many places by political borders, the expansion of neighbouring languages (Russian, Latvian) and even massive deportations or post-war migration. Because of the close relatedness of the Finnic languages, it is often difficult to distinguish contact phenomena from common heritage or convergent developments, or to tell whether external influences (e.g. loanwords) were acquired at the time of the reconstructible protolanguage or later, either separately or through other Finnic languages. Areallike bundles of features can, in fact, consist of phenomena of very different ages and backgrounds. 2.5 Present status, use, and the future Of the Finnic languages, Finnish and Estonian now have a secure status: they are official languages spoken by the majority of the population. In Finland the former superstratum language, Swedish, does not threaten Finnish, neither do the other old minority languages (Sami, Romani) or the languages of the still relatively small immigrant groups (Russian, Estonian and others). Rather, some Finnish linguists have expressed their concern that globalization and international contacts might make Finns themselves dangerously neglect their native language and constrict its use.
208 Johanna Laakso
In Estonia, the Russian minority (mostly post-war immigrants) is large and its rights are still a delicate question. However, Russian does not constitute any direct threat to Estonian, which survived through the Soviet era as a language used in and fit for practically all areas of both public and private life. The smaller Finnic languages can all be considered endangered. Their history — with a few exceptions for Livonian — tells of constant assimilation to Russian (Latvian in the case of Livonian), almost always the language of education, religion and official use. Although many small literary languages were created at the beginning of the 1930’s (e.g. for Veps, Ingrian and Tver Karelian), the terror at the end of the 1930’s put an end to their cultivation. After World War II, a national revival has only been thinkable in the few years since the collapse of the Soviet Union. In Karelia, the contacts with Russian are old and have resulted in Sprachbundlike phenomena between Karelian and local dialects of Russian (Sarhimaa 1995). In the 20th century, both immigration from other parts of Russia and the Russification of Karelians have tipped the balance so that Karelian is spoken by a small minority and severely endangered. During the Soviet era, the Republic of Karelia suffered from inconsistent language policies: in 1937–1940, the authorities tried to promote a quickly-generated and short-lived Karelian literary language. Before and after that Finnish — spoken by many Finnish immigrants in Soviet Karelia and connected with political goals, but unintelligible for southern Karelians — had an official status. The attempts to save Karelian and create a powerful standard language have to cope with the problem of great differences between dialects. It may not be possible to create a unified literary language for all Karelians, at least not in the foreseeable future. The situation is made even more difficult by the economical problems of present-day Russia. Most fluent speakers of Karelian, as well as Veps, belong to the older generations, and the language of younger speakers is often mixed with Russian elements and direct Russian influences (cf. e.g. Pyöli 1996). Apart from Russian, the almost extinct languages of Ingria, Ingrian and Votian, are also heavily mixed with elements from other Finnic languages of Ingria or, for some of those Ingrian evacuees who finally settled down in Estonia, also Estonian. While Vepss, like Karelians, are currently trying to establish a written language and revive the language for the youngest generation, Ingrian and Votian are already on the verge of extinction. The Votians never had a written language; for Ingrian (like Veps), a short-lived literary language was created in the 1930’s. After centuries of assimilation, the remaining Livonians experienced a brief national revival between the world wars, including some literature and even voluntary school instruction of Livonian. World War II and the Soviet era put an end to these. In the last few years, a group of activists, including some young people, have organized attempts to save or revive the Livonian language.
The Finnic languages 209
References Abondolo, Daniel [ed.] 1998: The Uralic Languages. London/New York: Routledge. Ariste, Paul 1981: Keelekontaktid. Eesti keele kontakte teiste keeltega. Tallinn. CIFU-8 = Congressus Octavus Internationalis Fenno-Ugristarum. Jyväskylä 1995. Comrie, Bernard. 1981. The languages of the Soviet Union. Cambridge. Décsy, Gyula. 1967. Is there a Finnic substratum in Russian? Orbis 16: 150–160. Forsberg, Hannele. 1998. Suomen murteiden potentiaali. Muoto ja merkitys. Helsinki. FUF = Finnisch-Ugrische Forschungen. Zeitschrift für finnisch-ugrische Sprach- und Volkskunde. Helsinki. Grünthal, Riho. 1997. Livvistä liiviin. Itämerensuomalaiset etnonyymit. (English summary: The Finnic ethnonyms.) Castrenianumin toimitteita 51. Helsinki. Fromm, Hans 1997: Germanen im bronzezeitlichen Mittelschweden? FUF 54: 127–150. Hahmo, Sirkka-Liisa 1997: Finnische nominale Ableitungssuffixe fremder Herkunft. In: Hahmo et al. (eds.), 1997: 95–102. Hahmo, Sirkka-Liisa et al. (eds.). 1997. Finnisch-ugrische Sprachen in Kontakt. Vorträge des Symposiums aus Anlaß des 30-jährigen Bestehens der Finnougristik an der Rijksuniversiteit Groningen 21.–23. November 1996. Maastricht. Häkkinen, Kaisa. 1996. Suomalaisten esihistoria kielitieteen valossa. Tietolipas 147. Helsinki. Halling, Tiina. 1997. Der Dativ im Livischen im Vergleich zum Lettischen. In: Hahmo et al. (eds.), 1997: 103–110. Hausenberg, Anu-Reet. 1985. Die ostseefinnisch-permischen Kontaktwörter in Zeit und Raum: VI. Internationaler Finnougristenkongress (Syktyvkar, 24.–30. Juli 1985). Preprint KKI 30. Tallinn. Helasvuo, Marja-Liisa. 1996. A discourse perspective on the grammaticization of the partitive case in Finnish. SKY 1996 (Suomen Kielitieteellisen Yhdistyksen vuosikirja = Yearbook of the Linguistic Association of Finland): 7–34. Helsinki. Helimski, Eugene. 1995. Proto-Uralic gradation: continuation and traces. CIFU-8 I: 17–51. Hint, Mati. 1983. Viroko neljän kvantiteettiasteen sävelkieli? (Summary: Is Estonian a language with four contrastive quantity degrees and four tones?) Virittäjä 87: 166–177, 328–348. Hint, Mati. 1991. Viron sanaston ja kieliopin ideologisoinnista. Kieli, valta ja eriarvoisuus. Esitelmiä 18. kielitieteen päiviltä. Kieli 6. University of Helsinki, Dept. of Finnish. 137–148. Honti, László. 1996. Numerusprobleme. Ein Erkundungszug durch den Dschungel der uralischen Numeri. FUF 54: 1–126. Itämerensuomalaiset kielikontaktit = Läänemeresoome keelekontaktid. Itämerensuomalainen symposium 7. kansainvälisessä fennougristikongressissa Debrecenissä 27.8.–1.9.1990. Kotimaisten kielten tutkimuskeskuksen julkaisuja 61. Helsinki. Itkonen, Terho. 1983. Välikatsaus nykysuomen juuriin. (Zusammenfassung: Über die Wurzeln der finnischen Sprache: Zwischenbetrachtung.) Virittäjä 97. Itkonen, Terho. 1997. Reflections on Pre-Uralic and the “Sami-Finnic protolanguage”. FUF 54: 229–266. Janhunen, Juha. 1982. On the structure of Proto-Uralic. FUF XLIV. Kangasmaa-Minn, Eeva. 1993. Aspektista ja sen sukulaisilmiöistä suomalais-ugrilaisissa kielissä. In: Yli-Vakkuri (ed.), 1993: 13–23. Kasik, Reet. 1989. Suomen ja viron verbinjohto. Sananjalka 31: 81–92. Kauppinen, Anneli. 1998. Puhekuviot, tilanteen ja rakenteen liitto: tutkimus kielen omaksumisesta ja suomen konditionaalista. Suomalaisen Kirjallisuuden Seuran Toimituksia 713. Helsinki.
210 Johanna Laakso
Koivisto, Vesa. 1995. Itämerensuomen refleksiivit. Suomalaisen Kirjallisuuden Seuran Toimituksia 622. Helsinki. Koivulehto, Jorma. 1997. Die Datierung der germanisch-finnischen Kontakte, revidiert. In: Hahmo et al. (eds.), 1997: 11–34. Koivulehto, Jorma. 1999. Varhaiset indoeurooppalaiskontaktit: aika ja paikka lainasanojen valossa. In: Paul Fogelberg (ed.), Pohjan poluilla. Suomalaisten juuret nykytutkimuksen mukaan. Bidrag till kännedom av Finlands natur och folk 153. Helsinki. 207–236. Koponen, Eino. 1998. Eteläviron murteen sanaston alkuperä. Itämerensuomalaista etymologiaa. [Abstract: The origin of vocabulary in the South Estonian dialect. Finnic etymology.] Mémoires de la Société Finno-Ougrienne 230. Helsinki. Korhonen, Mikko. 1978. Oliko suomalais-ugrilainen kantakieli agglutinoiva? (Summary: Was Proto-Finnic an agglutinative language?) Virittäjä 82. Korhonen, Mikko. 1981. Johdatus lapin kielen historiaan. Helsinki. Korhonen, Mikko. 1988. Uralilaisten kielten jälkitavujen vokaaliston historiaa. (Zusammenfassung: Zur Geschichte des Vokalismus der nichtersten Silben in den uralischen Sprachen.) Virittäjä 92. Korhonen, Mikko. 1996 [1992]. Remarks on the structure and history of the Uralic case system. Typological and historical studies in language by Mikko Korhonen. Mémoires de la Sociéte Finno-Ougrienne 223. 219–234. [First published in Journal de la Société Finno-Ougrienne 1992.] Kulonen, Ulla-Maija. 1989. The passive in Ob-Ugrian. Mémoires de la Société Finno-Ougrienne 203. Helsinki. Kytömäki, Leena. 1992. Suomen verbiderivaation kuvaaminen 1600-luvulta nykypäiviin. Publications of the Department of Finnish and General Linguistics of the University of Turku, 40. Laakso, Johanna. 1989. Muodon ja funktion suhteesta itämerensuomen verbinjohtimistossa. (Summary: On the relationship between form and function in the Baltic-Finnic system of verbal derivative suffixes.) Virittäjä 93: 50–69. Laakso, Johanna. 1997. On verbalizing nouns in Uralic. FUF 54: 3: 267–304. Laalo, Klaus. 1983. Nykysuomen kaksitavuisten nominien nuorin kerrostuma: puhekielen uudennokset. Virittäjä 87. Laanest, Arvo. 1982. Einführung in die ostseefinnischen Sprachen. Hamburg. [First version published in Estonian: Sissejuhatus läänemeresoome keeltesse, Tallinn 1975.] Lehiste, Ilse. 1998. Partitive or illative? Proceedings of the Finnic Phonetics Symposium, August 11–14, 1998, Pärnu, Estonia. [Published as third number of the quarterly] Linguistica Uralica, Vol. 34: 162–171. Lehtinen, Tapani. 1983. Suomen konditionaalin morfologisesta ja semanttisesta motivaatiosta. (Zusammenfassung: Zur morphologischen und semantischen Motivation des finnischen Konditionals.) Virittäjä 87. Lehtinen, Tapani. 1984. Itämerensuomen passiivin alkuperästä. (Suomi 129.) Helsinki. Lehtinen, Tapani. 1985. Venäjän verbisemantiikkaa itämerensuomessa: karjalan ja vepsän inkoatiiviverbit. Sananjalka 27. Lehtinen, Tapani. 1990. Venäjän aspektien heijastumia vepsän verbijärjestelmässä. In: Itämerensuomalaiset kielikontaktit. Lehtinen, Tapani. 1994. Suomen kielen esihistoriaa. (Draft version of a book in preparation; available as a photocopy at the Dept. of Finnish, University of Helsinki.) Markianova, L. F. 1987. O nekotoryx tipologicˇeskix osobennostyax glagoloobrazovanija v pribaltijsko-finskix jazykax severnoj gruppy. XVII Vsesojuznaja finno-ugorskaja konferencija. I. Jazykoznanie. Tezisy dokladov. Ustinov. 147–149. Metslang, Helle. 1997. Unterschiedliche Tendenzen in den grammatischen Systemen des Estnischen und des Finnischen. In: Hahmo et al. (eds.), 1997: 165–174.
The Finnic languages
Nikkilä, Osmo. 1998. nop-ea Æ nop-sa, sel-keä Æ sel-vä: suffiksinvaihto ja suomen sanojen etymologiointi. In: Urho Määttä & Klaus Laalo (eds.), Kirjoituksia muoto- ja merkitysopista. Folia Fennistica & Linguistica 21. Tampere. 77–101. Ojutkangas, Krista. 1997. Suomen ja viron U- ja ne-johdosten käytön vertailua (Summary: Comparison of the use of verb derivatives with the suffixes -U- and -ne- in Finnish and Estonian). Virittäjä 101: 358–379. Pajusalu, Karl. 1996. Multiple linguistic contacts in South Estonian: Variation of verb inflection in Karksi. Publications of the Department of Finnish and General Linguistics of the University of Turku, 54. Paunonen, Heikki. 1993. Suomen mieli — oikea kieli. Virittäjä 97: 81–88. Paunonen, Heikki. 1995. Puhesuomen muuttuva omistusmuotojärjestelmä. (Summary: Morphological changes in spoken Finnish possessive forms). Virittäjä 99: 501–531. Posti, Lauri. 1954. From Pre-Finnic to Late Proto-Finnic. FUF XXXI. Pyöli, Raija. 1996. Venäläistyvä aunuksenkarjala. Kielenulkoiset ja -sisäiset indikaattorit kielenvaihtotilanteessa. University of Joensuu Publications in the Humanities, 18. Raag, Virve. 1998. The Effects of Planned Change on Estonian Morphology. Studia Uralica Upsaliensia 29. Acta Universitatis Upsaliensis. Uppsala. Salminen, Tapani (ed.). 1993. Uralilaiset kielet tänään. Snellman-instituutin julkaisusarja A 13. Kuopio. Salminen, Tapani. 1998. Pohjoisten itämerensuomalaisten kielten luokittelun ongelmia. In: Oekeeta asijoo. Commentationes Fenno-Ugricae in honorem Seppo Suhonen sexagenarii. Mémoires de la Société Finno-Ougrienne 228. Helsinki. 390–406. Sammallahti, Pekka. 1988. Historical phonology of the Uralic languages with special reference to Samoyed, Ugric, and Permic. In: Sinor (ed.), 1988. Sammallahti, Pekka. 1995. Language and roots. CIFU-8 I: 143–153. Sarhimaa, Anneli. 1995. Karjalan kansat ja kielet kontakteissa (Summary: Karelian peoples and languages in contact). Virittäjä 99: 191–223. Savijärvi, Ilkka. 1981. Sanajärjestystyyppi pääverbi–kieltoverbi viron kielessä. (Summary: The word order type main verb–negation verb in Estonian.) Virittäjä 85: 109–117. Seilenthal, Tõnu. 1981. A névutók fejlödésútjai az ugor nyelvekben. Congressus Quintus Internationalis Fenno-Ugristarum VI: 467–475. Turku. Shore, Susanna. 1986. Onko suomessa passiivia. Suomi 133. Helsinki. Sinor, Denis (ed.). 1988. The Uralic languages: description, history and foreign influences. E.J. Brill, Leiden/New York/København/Köln. Söderman, Tiina. 1996. Lexical Characteristics of the Estonian North Eastern Coastal Dialect. Studia Uralica Upsaliensia 24. Uppsala. Suhonen, Seppo. 1995. Finno-ugristische Sprachwissenschaft 1989–1994. Congressus Octavus Internationalis Fenno-Ugristarum. Pars I: 191–203. Tikka, Toivo. 1992. Vepsän suffiksoituneet postpositiot. Studia Uralica Upsaliensia 22. Uppsala. Tommola, Hannu. 1993. Ambipersoonainen suppressiivi: diateesista suomessa ja virossa. In: YliVakkuri (ed.), 1993: 41–96. Trosterud, Trond. 1994. Auxiliaries, negative verb and word order in the Sami and Finnic languages. In: Ago Künnap [ed.]: Minor Uralic languages: structure and development. Tartu. 173–181. UEW = Károly Rédei: Uralisches etymologisches Wörterbuch. Unter Mitarbeit von Marianne Bakró-Nagy, Sándor Csúcs, István Erdélyi, László Honti, Éva Korenchy, Éva K. Sal und Edit Vértes. 1–2. Harrassowitz, Wiesbaden. Vare, Silvi. 1993. Konversioonist eesti keele sõnamoodustuses. (Referat: Zur Konversion in der estnischen Wortbildung.) Virittäjä 97: 34–50.
211
212 Johanna Laakso
Viitso, Tiit-Rein. 1998. Fennic. In: Abondolo (ed.), 1998: 96–114. Winkler, Eberhard. 1997. Krewinisch. Zur Erschließung einer ausgestorbenen ostseefinnischen Sprachform. Veröffentlichungen der Societas Uralo-Altaica 49. Wiesbaden. Yli-Vakkuri, Valma (ed.). 1993. Studia comparativa linguarum orbis Maris Baltici 1. Tutkimuksia syntaksin ja pragmasyntaksin alalta. Publications of the Department of Finnish and General Linguistics of the University of Turku.
DOCINFO
AUTHOR ""
TITLE "Early history of the Circum-Baltic languages"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Part 2
Early history of the Circum-Baltic languages
AUTHOR "Östen Dahl"
TITLE "The origin of the Scandinavian languages"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
The origin of the Scandinavian languages* Östen Dahl
1.
Introduction: The Common Nordic Hypothesis
At the beginning of this century, the Swedish linguist Adolf Noreen gave an account of the origin of the Scandinavian languages, which, in slightly reworded translation, goes as follows: Although the time for the settlement of the Germanic population in Scandinavia cannot yet be exactly determined, it quite definitely took place before the birth of Christ, most probably as early as the beginning of the Late Stone Age (during the 5th millennium B.C. or even earlier). If this is correct, the Nordic or Scandinavian languages can be traced back1 almost 6,000 years. Until the beginning of the Christian era, however, nothing is known about the language of the old Scandinavians, which at that time was spoken not only in all of Denmark (including Schleswig) and large parts of southern and middle Sweden and Norway but also in several areas in Finland and Estonia.2 In spite of this fairly large geographical extension the language seems to have been fairly similar over this entire area, and it is regarded as a uniform language, the origin of all the modern Scandinavian languages, and it is accordingly called the Proto-Nordic language.3
This is an early formulation of what I shall here call the Common Nordic Hypothesis, which has become a standard assumption in the literature on the history of the Scandinavian languages. According to this hypothesis, there existed up to the Viking Age a uniform Germanic language which was spoken over most of Scandinavia and which is the origin of all modern Scandinavian languages and dialects. The quotation just given differs from most later formulations in that it is quite explicit about the beginning of the Germanic-speaking period in Scandinavia. Nowadays, hardly anyone dares apply the word “Germanic” to the Neolithic or Late Stone Age period (4,000–1,800 B.C.E.). If a date is given for the point in time when the Germanic languages split up, it is usually some time during the first millennium B.C.E. However, many authors combine this with the assumption that Common Germanic was spoken in an area that either coincided with or included that of
216 Östen Dahl
Common Nordic. A relatively strong version of this Scandinavian Urheimat hypothesis is found in a recent textbook, Robinson (1993: 15–16): Information about the early history of what are now called the Germanic peoples is also rather sketchy. The weight of evidence … points to an ancient homeland in modern Denmark and southern Sweden. As to when they got there from the Indo-European homeland, the evidence is primarily archaeological. Although humans had inhabited the region since about 10,000 B.C., the Germanic tribes are usually associated with a archaeologically distinct group, known as the “Battle-Axe culture”, who invaded the area in the third (?) [question mark in the original] millennium B.C…
This position reduces the history of Germanic in Scandinavia from 6,000 to 4,000 years; since the more tangible facts still start around the beginning of our era, even on this revised time-scale, it very much looks as if the same language has been spoken in Scandinavia since time immemorial. Accordingly, most general treatments of the history of the Scandinavian language do not indicate any time-point at which Proto-Nordic began to be spoken but almost invariably say that it was “spoken all over southern Scandinavia with only minor dialectal variations up to the Viking age”. In this paper, I shall argue that the Common Nordic hypothesis is neither plausible given what we know about language and language change in general nor supported by the linguistic data at hand. To a large extent, I shall be focusing on the question of what language was spoken in the Swedish provinces around Lake Mälaren, the assumed “cradle of the Swedish nation”. I shall argue that the apparent homogeneity in the language in the central parts of Sweden, Denmark and Norway around the previous millennium shift was quite a recent phenomenon and due to the spread of a koiné from a political and economical centre in the south.
2.
Linguistic preconditions: General assumptions about speech communities and linguistic change
One assumption about human languages that most linguists would agree to assign axiomatic status to is that change is a universal and unavoidable phenomenon. Put in a different way, if one leaves a speech community alone for a sufficient length of time, one will find that its language has changed in some ways. Furthermore, if a speech community is split in two and all contacts between the halves are cut off, the language will inevitably develop in different directions in the new communities, which means that if one waits long enough, the speakers will no longer be able to understand each other. We also know that this kind of pure split is seldom found in real life. Communities tend to be in contact even after splitting up, and they tend to
The origin of the Scandinavian languages 217
influence each other linguistically. Moreover, the notion of a speech community is itself problematic. The ideal case of a speech community would be a group of people where everyone speaks equally often to everyone else, and nobody ever speaks to anyone outside the group. In small, non-hierarchical societies, such as the hunter-gatherer groups that humankind is supposed to have consisted of once upon a time, speech communities are perhaps not far from this ideal, but in most other situations, they tend to be stratified internally and not well delimited externally. However, we may still use the ideal speech community as some kind of null hypothesis in the sense that we see deviations from it as something that has to be explained. More concretely, if we find that two groups of people who live at a considerable distance from each other do speak the same language, we have to find some facts in their history which explain why they do, whereas if they speak different languages, it is just the normal situation and needs no explanation. We may use linguistic innovation as a general term for changes in the language or languages a group of people speak, including both situations restricted to individual linguistic features — that is, what is usually called language change — and situations where a group of people switch to a new language — language shift. These two main types of linguistic innovation share a number of properties with each other and with non-linguistic cultural innovations. For instance, they tend to start in an economic and political centre and then spread over larger or smaller areas which are dependent on this centre. But language change and language shift also differ from each other in many important respects. Above all, while individual linguistic features generally may spread very easily from one group of people to another, even if the contact between them is relatively limited, language shifts take place only in rather specific situations, usually presupposing extensive bilingualism within a community, together with a lack of balance in power and/or numbers of speakers between the languages involved. There are also differences in the results. The diffusion of individual linguistic features within a geographical area is often incomplete; moreover, the areas covered by two such spreads are seldom identical. The result is the well-known criss-cross pattern of isoglosses seen in most dialectological maps. In this way, the result of language change is often a net increase in linguistic diversity. Language shift, on the other hand, tends to decrease linguistic diversity, in that a uniform language is introduced into a previously heterogeneous area. This is of course what has happened in the recent spread of standard languages in most European countries and many other parts of the world. We may assume that similar processes have taken place earlier in history, obliterating without trace much of the diversity that has undoubtedly existed. When reconstructing the development of the languages in an area, historical linguists often seem to employ (maybe unconsciously) what can be called the “cracking monolith” model. That is, it is assumed that the language in a given geographical area was originally uniform and then became differentiated through
218 Östen Dahl
“cracks” that successively divided the “monolith” into smaller chunks. A consequence of this model is that linguistic diversity is seen as increasing over time. But, in the cases where we are able to observe longer periods of language development, what we can observe is an even degree of linguistic diversity as long as the political and economical system does not change. A certain scepticism against “cracking monolith” hypotheses thus seems in place, in particular as these hypotheses seem to flourish especially well with respect to periods where there is little or no documentation of the actual linguistic situation. We should, at least, always consider the plausibility of the resulting picture in the light of what we know about similar sociolinguistic situations at other times (this could be seen as an application of the uniformitarian principle of geologists — do not assume that the past obeyed other laws than the present).
3.
Demographic and political structure of southern Scandinavia during the Iron Age (500 B.C.E.–800 C.E.)
It is hard to get a clear picture of the demographic, political and ethnic situation in Scandinavia during the Iron Age. Some generalizations are still possible: It seems that the population was rather sparse in the pre-Roman Iron Age and continued to be so in large parts of the region even in later periods. In a few areas, however, there was strong population growth especially in the Late Iron Age. In Sweden, this goes especially for the Mälar provinces (primarily Uppland and Södermanland). Hyenstrand (1982) presents population estimates for a number of Swedish provinces according to which there would have been around 10,000 people in the Mälar provinces and as many again in all the other core Swedish provinces4 put together in 500 C.E. These figures are given in the context of a hypothetical argument in Hyenstrand’s text and can therefore at best be seen as a rough indication of the order of magnitude that we are dealing with. Having said this, it is still worth noting that with the size of the core Swedish provinces being about 120,000 square kilometres, these figures would imply an extremely low population density (0.2 inhabitants per square kilometer), comparable to that of present-day Alaska. With regard to the political structure, the generally accepted opinion seems to be that there is no reason to assume any organization transcending province boundaries (as traditionally defined) in Sweden before the Viking Age. Local chiefdoms are generally assumed to have been the dominating type of political entity. There would be little room for any more advanced stratification of society. (The political organization in Denmark may have developed earlier than that in Sweden.) What we know from archaeology is in general compatible with the picture of the ethnic composition of Scandinavia obtained from various contemporary authors. The fullest description is probably that of Jordanes from around 500 C.E.
The origin of the Scandinavian languages 219
Jordanes mentions by name about thirty tribes in Scandinavia, most of them in the southern part. Fairly plausible hypotheses may be made about the identity of a number of these; the details do not matter here. What is relevant, though, is the general resulting picture of Scandinavia as being populated by a large number of groups, well enough defined to have names known also to foreigners, consisting of at most a couple of thousand people and living at relatively large distances from each other. As an example, consider the originally Danish province of Halland, which is essentially a 150 kilometer long strip of land along the Swedish west coast. Three of Jordanes’ groups (ferur, hallin, ahemil) have been assumed to have lived in Halland; as Hyenstrand notes (1982: 64), it is striking how well this fits with archaeological evidence of the distribution of the Iron Age population. It means that the relatively concentrated settlements that are to be assumed in Halland at this time would be separated from each other by gaps of up to 50 kilometers. Another example, geographically close to Halland, is the present-day province of Småland, which takes its name from the ten “small lands” it originally consisted of (one or two of which figure in Jordanes’ description). There is also archaeological evidence that shows cultural differences between these, such as different styles of graves (Hyenstrand 1996: 28–30). What assumptions can be made about the linguistic situation in an area with the demographic and political structure we have been talking about? Experience from other parts of the world and other historical periods strongly suggests that Jordanes’ groups would, in general, correspond to separate speech communities, each with their own language development. How different the languages of the groups were is, of course, another question; the answer depends on a number of factors that we shall discuss below. One point to be made here, however, is that the same factors that favour linguistic differentiation and small speech communities also make rapid language shift without major population replacement unlikely. In particular, it is rather improbable that a large number of small and dispersed groups of people would simultaneously switch to another language.
4. The first wave of expansion: Germanic As was noted above, the idea that the Germanic Urheimat was situated in Scandinavia has enjoyed considerable popularity over the years. This is not the place to go into a detailed discussion of this hypothesis, which seems to have been favoured both by various obscure statements made by classical authors and by patriotic and romantic sentiments. It should be noted that the region in Scandinavia from where the Germanic peoples supposedly originated tends to be rather vaguely delimited, as “southern Scandinavia” or “Denmark and southern Sweden”. The part of Sweden this would involve may be taken either to be the Scanian provinces — Skåne
220 Östen Dahl
(Scania), Blekinge, and Halland, which all belonged to Denmark before the 17th century — or include all of the traditional partitions Götaland and Svealand, in particular, the core Mälar provinces Uppland and Södermanland. Even vaguer formulations indicate “northern Germany and southern Scandinavia” as the original abode of the Germanic peoples. The time-point of the assumed exodus — of direct importance for the dating of the first split in the Germanic speech community — is also somewhat unclear. Robinson (1993), talking of “a gradual splitting-up of the Germanic people and their languages, along with a migration southward out of their original homeland in southern Scandinavia”, says that Germanic tribes had spread “from northern Belgium in the West to the Vistula in the East” by 300 B.C.E. If we want to allow reasonable time for this operation, we have to assume that the migration southward started back in the Late Bronze Age, that is, before 500 B.C.E. However, recent research has tended to associate the genesis of the Germanic peoples as an identifiable grouping with the rise of iron technology in northern Europe, notably the Jastorf culture whose centre was situated in present-day Lower Saxony approximately between 600–300 B.C.E. While this leaves room for outliers in Schleswig and Jutland, the hypothesis formulated in Udolph (1994) places the (pre-Jastorf) Urheimat firmly between the Elbe, the Erzgebirge and the Thuringian Forest, a fair distance from Scandinavia. Udolph bases his conclusions on placename evidence; according to him, Scandinavia does not, by and large, share any of the oldest layers of Germanic place names. The hypothesis that a relatively homogeneous Germanic language was spoken at this time is compatible with the degree of similarity between the oldest attested forms of Germanic dating from about 500 years later, and the fact, to be discussed later, that no systematic difference between North and West Germanic can be observed in these forms. Using the (admittedly controversial) method of lexicostatistics, Arndt (1959) comes to the conclusion that the first split in the original Germanic speech community took place around the beginning of the Common Era. It is also plausible to assume that a phenomenon like the Germanic expansion would be preceded by a local population increase such as would be induced by the introduction of new technology — which speaks in favour of the association of the Jastorf hypothesis. Furthermore, it is in this kind of social situation that radical linguistic change could be expected. It may be noted that this dating of “Common Germanic” leaves open what happened before this point in time, and in particular, what the ancestry of “Common Germanic” was. Of course, if we knew more about the Bronze Age of Northern Europe, we could certainly identify the linguistic forerunners of Germanic — which would not, however, necessarily have been spoken by the main contributors to the gene pool of the present-day Germanic-speaking peoples. In two recent papers, Elert (1993, 1997) has discussed the origin of the Scandinavian languages, partly along similar lines of argument as those found in this paper.
The origin of the Scandinavian languages 221
His conclusion is that Germanic languages came to be spoken in Scandinavia as the result of a language shift in the period 1000–500 B.C.E. While this is a late placement compared to most earlier datings by Scandinavianists, it looks as if it leaves rather little time for a move from the continent if e.g. Udolph’s ideas are accepted. It appears probable that Germanic was spoken around the westernmost part of the Baltic — in present-day Denmark and Germany — somewhat before the beginning of our era, although evidence in the form of writing does not appear until the second century C.E. Another question concerns when Germanic reached further into the Baltic. It has been argued that Germanic loan-words in the Baltic-Finnic languages prove early Germanic-Finnic contacts.5 Udolph (1994) combines this argument with onomastic evidence and assumes that the area of contact was located in the Åland islands, south-east Finland and Estonia around the beginning of our era. It thus seems reasonable to assume that there were also groups of Germanic speakers in the Swedish coast provinces at this time. Approximately at the same time, the Goths seem to have established their presence around the Vistula estuary in present-day Poland. Beyond this point, however, we risk getting lost in pure speculation — there is very little substantial information in the literature about these things. It may be noted, however, that the Germanic groups that took part in the Baltic expansion need not at this time have been specifically “North” or “East” Germanic.6 Also, such an expansion need not have gone via Denmark but could have gone directly from Germany to southern Sweden. The references to the Suiones in Tacitus’ Germania, written around 100 C.E., have traditionally been understood as implying the existence of a Swedish kingdom already at that time; in view of Tacitus’ somewhat confused notions of the geography of the region, however, what he says at most suggests the existence of a people with that name somewhere in or around the Baltic. I do not think this really contradicts the timetable given here. What was said in the previous section about language shifts in non-centralized societies makes it probable that the introduction of Germanic took rather a long time, and that there must have been groups speaking other languages for several centuries after the advent of the first Germanic speakers, perhaps still in Jordanes’ times.7 We may note that the Iron Age population seems to have been concentrated to the lowland parts of Scandinavia, to a very large extent coinciding with the areas that are nowadays the major agricultural regions. A common pattern found all over the world is for “new” languages to be predominant in areas of economic expansion, usually centrally positioned within an area, while smaller residual groups speaking “old” languages are found in peripheral and economically less dynamic parts. There is no lack of suitable locations for such residual groups in Scandinavia. Also, being illiterate and small in numbers, it is not very likely that such groups would leave very substantial traces of their languages.
222 Östen Dahl
In contradistinction to most proto-languages, which are merely reconstructions, the common ancestor of the Scandinavian languages is usually claimed to be attested, mainly through runic inscriptions using the 24-symbol Older Futhark, dated from the 2nd to the 7th century. Accordingly, the language used in these inscriptions is simply referred to as “Proto-Nordic” (urnordisk(a)). Since one of the aims of this paper is to question the status of “Proto-Nordic”, I cannot very well use that name but shall call it by the more non-committal name Early Runic instead. The identification of Early Runic with the assumed Scandinavian proto-language has a central role in the Common Nordic hypothesis. However, the specific Scandinavian character of Early Runic was actually questioned in the fifties by the German scholar Hans Kuhn, who noted that the early runic inscriptions contain very little that show that they are Scandinavian rather than early West Germanic. Accordingly, Kuhn postulated a “North-West Germanic” unity that would have been broken up only around the middle of the 1st millennium (Kuhn 1955: 29), whereas the connections between Gothic and the other Germanic languages would have been severed much earlier. Later discussion of the issues seems to have qualified rather than contradicted Kuhn’s claims (cf. Makaev 1996 [1965], Antonsen 1975 and the papers in Marold & Zimmermann 1995). Arndt 1959 also includes a North-West Germanic unity (explicitly not including Old High German) in his lexico-statistically based time-table of the development of the Germanic languages. This, of course, does not mean that Early Runic is the ancestor of all North and West Germanic languages; rather, what seems likely is that the Germanic dialects spoken around the western end of the Baltic at the time when the written form of Early Runic was fixated — presumably around the 3rd century — were not very different from each other and probably formed some kind of speech community, which was finally destroyed only with the Slavic expansion.8 It is, however, natural to ask how far such a speech community may have extended — did it really cover all of southern Scandinavia, as is usually claimed? This takes us to the general question of what conclusions about the spoken language can be drawn from the Older Futhark inscriptions. The Older Futhark inscriptions are few and in general very short. Many of them consist only of the futhark itself or of a single word, often a name. Their interpretation is sometimes obscure; some have not been interpreted at all. The latter obviously cannot be taken as evidence for a uniform Scandinavian language — in fact, for all we know, their language could be non-Germanic. It has been noted repeatedly that the apparent uniformity of Early Runic as a written language may be partly explainable by the general conservatism of written norms and/or the existence of professional scribes, who would not necessarily be taken from the local population.9 The name of the Russian scholar Ènver A. Makaev has become associated with the “Runic koiné” hypothesis, according to which Early
The origin of the Scandinavian languages 223
Runic was “the first Germanic literary language” which “passed along from master to master, from one school to another” (Makaev 1996 [1965]: 47).10 In all fairness, it should be noted that dialectal differences in Early Runic times have been postulated earlier (such as the alternation u : o in words like ko noted in Wessén 1930). It seems, though, that scholars tend to postulate dialectal differences in early inscriptions only in cases where they correspond to independently attested sound-developments.11 The possibility that local features may come and go, being wiped out by a standardizing wave, probably does not occur to them. There are also some points of a general methodological nature. We are dealing with a material that is extremely limited (about 200 characters) and only part of it has been given a reliable interpretation. As long as this is the case, it is of course impossible to say how much the language deviates from that found in other inscriptions. Also, any deviation would be likely to show up only once, due to the brevity of the texts, and one could never know if it represents a mistake or a true difference in the language. Finally, as long as the “Standard Early Runic” character of inscriptions is seen as axiomatic, this will influence their interpretation, even if the results are in themselves less plausible. Antonsen (1986) argues that the ways in which some runic stones have been read have been influenced by the investigators’ expectations that the texts would have a “Nordic” character.12 It is easy to get the impression when reading the literature that the Older Futhark inscriptions were more or less evenly sprinkled all over Scandinavia. This, however, is not the case — in fact, show a very definite clustering in time and space. One may discern three basic groups: – – –
an early group (before 300) of movable objects (clasps, spearheads etc.), mainly in Denmark and the Scanian provinces a late group (after 350) of movable objects (overwhelmingly bracteates and other gold objects), also mainly in Denmark and the Scanian provinces another late group (after 300) of stones and cliff inscriptions — these are concentrated to Norway (including the Swedish province of Bohuslän, earlier belonging to Norway); some are also found in Sweden (all after 400).
In Table 1, I give some statistics based on Krause (1966), whose listings are slightly dated but so far the most complete ones.
224 Östen Dahl
Table 1.Inscriptions in the Older Futhark according to Krause (1966)
Denmark incl. the Scanian provinces Gotland Norway incl. Bohuslän Rest of Sweden Germany All other countries
Inscriptions Bracteates excl. bracteates
Inscriptions before 400 C.E.
20 5 54 12 20 10
11 1 6 0 1 4
38 11 4 4 4
As can be seen, there are relatively few inscriptions (and even fewer bracteates) from the core Swedish provinces, and none of them is dated before 400 C.E. Actually, the full texts of these inscriptions (about 200 characters) can easily be shown on a map (Map 1). Particularly striking is the low number of inscriptions in the economically and politically central provinces of Uppland and Södermanland. To emphasize this I give here a complete listing of the attested inscriptions from the Mälar region: –
–
–
the Möjbro stone, commonly dated to the 6th century, with an inscription traditionally rendered as frawaradar r anahaislaginar r (a number of interpretations have been suggested, e.g. “Frawaradar was killed on the horse”) the Krogsta stone, dated to roughly the same time, with the two-line inscription mwsïeij × sïainar (the first word has not been interpreted, the second is usually taken to be the word for ‘stone’) the recently found Vallentuna dice, with the following partially unreadable sequence: xxxhi/lAhAhAukrAlbuxx (interpretation unclear)
in Södermanland: – the Skåäng stone with the text harija leugar (interpreted as proper names) – the Berga stone with the two separate words saligastir r and fino (interpreted as proper names). I think there are basically two conclusions to be drawn from the data summarized here. The first is that the scarcity of inscriptions from Sweden, in particular the Mälar provinces, makes it difficult to make any definite claims about the status of writing in these parts before 800 C.E. and even more so before 400 C.E. It seems plausible that the use of the Older Futhark came significantly later to Sweden than to Denmark and Norway, and that its use may have been less extensive than in other places. Such a view appears compatible with what archaeologists have demonstrated in other domains, namely that the culture of southern Scandinavia was not homogeneous. Such considerations are of some importance when we try to get a picture of the linguistic situation in general. If the communication barrier between Denmark and Sweden was strong enough to keep back the introduction of runic
writing for a couple of hundred years, is it then plausible to assume that the contacts were at the same time intensive enough to preserve linguistic uniformity? The second conclusion to be drawn is that the discussion about the homogeneity of Early Runic up to now in a way has missed the point. If we had a large number of inscriptions from the Mälar provinces, it would make sense to ask to what extent they reflect the spoken language in the area. However, as we have seen, the material from this region is so scanty that it hardly says anything at all. The only inscription that contains something intelligible that is not just a proper name is the Möjbro stone, and, as has already been mentioned, its interpretation is far from clear. In fact, if the Möjbro stone should be taken to tell us anything at all, it could equally well be seen as evidence of language diversity as of uniformity. Let us look at the details. The first seven or eight runes in the second (physically the first) line of the Möjbro stone (anahahai…) have not received an uncontroversial interpretation. Both Krause (1971) and Antonsen (1975) assume that the meaning is ‘on the horse’. For this hypothesis to hold, the following circumstances must not be seen as obstacles: (i) the preposition ana ‘on’ has no other clear attestation with a final a in Early Runic (only in Gothic); (ii) no noun like haha meaning ‘horse’ is otherwise attested in Scandinavia (hanha- occurs in some compound names on the continent); (iii) the dative ending -i would at the date commonly assumed for this stone normally have fused with the final -a to -e. Mainly relying on the last-mentioned fact, Antonsen (1975) places the Möjbro stone as early as the beginning of the 4th century, making it about a hundred years older than any other inscription in the core Swedish provinces, and at least one hundred and fifty years older than most other scholars have assumed. He does not discuss the possibility that the language
226 Östen Dahl
on the Möjbro stone reflects a more recent but also more conservative language variety, yet, this supposition seems no less likely than the one he makes. If the Möjbro stone in fact reflects the spoken language of Uppland in the 6th century, and that language was derived from the Germanic dialect spoken by groups arriving in the area about half a millennium earlier, the deviations from the Early Runic norms are only what we should expect. In particular, we may note similarities with Gothic, which of course do not imply that 6th century Upplandic was Gothic or even in any sense an “East Germanic” language.13
5.
The second wave: Nordic
The language found in the later runic inscriptions (Late Runic) differs very markedly from that in those written in the Older Futhark. It is highly questionable if a person from the 11th century would have understood the speech of Hlewagastir, the author of the famous Gallehus horn from around 400 V.C.E. Wessén (1941) enumerates nine major changes that took place during the period 600–1000 C.E.: (1) syncope (*dagar > dagr ‘day’); (2) umlaut of various types (*gastir > gestr ‘guest’); (3) disappearance of j and w before back vowels (*jungar > ung ‘young’); (4) disappearance of final -n (*geBan > gefa); (5) various assimilations of consonants (*gulþa > goll); (6) unification of 2nd and 3rd person verbal endings (bindr ‘binds’); (7) introduction of middle forms in -sk (kallask ‘to be called’); (8) introduction of suffixed definite articles (dagr-inn ‘the day’); (9) renewal of the personal pronoun system: hann, hon ‘he, she’. Not mentioned by Wessén but also important is the rise of the Scandinavian tone accent system. One should also add lexical changes to the list. The standard way of interpreting the linguistic innovations enumerated in the preceding paragraph is as a set of changes that applied to a basically uniform language — “Proto-Nordic” — and yield another, which however either immediately (partly via those changes themselves) or shortly afterwards split up in two dialects: “West Nordic” (Norwegian, Icelandic) and “East Nordic” (Danish, Swedish). The first assumption that may be questioned is the identity of the language that was the input to the process, in other words, whether the language of the early runic inscriptions deserves its standard name “Proto-Nordic”. If Early Runic was, as various scholars have suggested, a standardized written language not necessarily corresponding to any spoken dialect, the question should rather be how different the actual progenitor of the Nordic languages was from this written standard. I shall return to this question below. Whereas considerable attention has been given to the temporal ordering of the changes, the role of space has largely been neglected. Indeed, even in recent works,14 attempts are made to date inscriptions from the transitional period relative to each
The origin of the Scandinavian languages 227
other on the basis of their language. Yet, of course, major changes of this kind must take considerable time to spread over such a large area. Indeed, one may go further and question the plausibility of the idea that such a radical and specific set of changes could apply in such a uniform way to a set of dialects spoken at considerable distances from each other. (Notice in particular the improbability that such a idiosyncratic and typologically uncommon tone accent system as the one found in the Scandinavian languages15 would spread uniformly.) As noted earlier, each individual innovation in languages tends to have its own geographic distribution. Here, although differences between the West and East Nordic areas are acknowledged, the changes are assumed to take place in basically the same way from Jutland in Denmark to Hälsingland in Sweden. Scholars before me have already questioned the plausibility of the assumption that the Scandinavians preserved a common language over many centuries. But the most implausible assumption is really another: that they also changed their language all at the same time and in the same fashion, as if conforming to a EU regulation on the length of cucumbers. What, then, is the alternative? To me, it seems much more natural to think of the implementation of the changes enumerated above as part of a process of language shift, more specifically, the spread of a “prestige dialect”. Thus, the explanation of the uniform language that we find in the runic inscriptions from the 11th century onwards, in written documents from the 13th century onwards and indirectly in later spoken and written Scandinavian dialects is that at least the ruling classes in the central parts of the Scandinavian countries were using a common language which had spread very recently. The final result of this spread was the obliteration of the dialects or languages spoken earlier in the peripheral parts of the area. Consequently, instead of increasing linguistic diversity (by giving rise to the split between West and East Nordic) as suggested by the traditional account, the outcome was a decrease in diversity, a unification of the languages spoken in Scandinavia. The period in which the transition from Early to Late Runic occurred is a “dark” period with little known about linguistic or non-linguistic developments. This makes it difficult both to map the details of the linguistic situation and to discuss their societal preconditions. In any case, it is fairly obvious that the conditions for the spread of a common language in Scandinavia were rather different towards the end of the first millennium than they were in its first half. An important technological factor that made possible more efficient communications between the different parts of Scandinavia was the introduction of sailing ships around 600 C.E. Larger political entities start to arise approximately at the same time. The first settlements of an urban character arose in the 8th century. We may thus assume that conditions were created for a mobile élite, whose members would be more tightly connected to each other than to other groups in the places where they
228 Östen Dahl
happened to live — thus preparing the ground for the spread of a trans-regional language. The question is then if we can identify when and where this spread began. Earlier research tended to focus on the Mälar provinces as a centre of political power during the centuries before the Viking Age, but lately the role of Denmark as the dominating power in Scandinavia at least from the 8th century on, if not earlier, has come to the fore. In this perspective, the Mälar provinces are naturally seen as a peripheral part of the Danish sphere of influence, rather than as a centre in their own right. Consider in this context the statement by the historian Peter Sawyer that “the wide currency of the Scandinavian tongue, which was known as the Danish tongue even in Iceland…may itself be a consequence of Danish overlordship in the north, comparable with the spread of Latin or English in the Roman and British empires” (1991: 285). Although one should perhaps not draw too far-reaching conclusions from the use of the adjective ‘Danish’ in the denomination of the language, Sawyer’s formulation is interesting in that it implies a rather different picture of the genesis of the Scandinavian linguistic unity than the traditional “cracking monolith” model, viz. one where the common language spread with Danish hegemony rather than being there from time immemorial. In fact, the Swedish scholar Bengt Hesselman as early as in the thirties put forward a hypothesis which, with some modifications, fits well into the picture I have just suggested. In Hesselman (1936), he postulates the existence of a specific language variety called “Birka Swedish” (Birkasvenska). Hesselman’s point of departure is the existence of alternate forms such as jär of the demonstrative adverb här ‘here’. Such forms are or were found in Nordic dialects spoken in various parts of Scandinavia, including Upper Norrland and Dalecarlia in continental Sweden, Ostrobothnia in Finland, Gotland in the Baltic and the Swedish dialects in Estonia, but also in Danish dialects in an area of southern Jutland and Schleswig. Hesselman provides evidence that forms beginning with j- were earlier found over a larger area and draws the conclusion that there was a sound change ¯e > ja which spread from the Mälar region with Birka as the centre and was in fact one feature of “Birka Swedish”, a language variety supposedly spoken “in a contiguous area around the Baltic Sea from Överkalix in the north to Slesvig (Hedeby) in the south” (Hesselman 1936: 158; my translation).16 This change would never have reached Southern Sweden, however. Hesselman further suggests that the original centre of innovation may have been Hedeby rather than Birka. Hedeby, close to the present-day city of Schleswig on the east coast of Jutland, is believed to have been founded in the middle of the 8th century and is known as a major commercial and political centre from that time onwards. Birka seems to have been founded around the same time and may be seen as part of the same network. In Hesselman’s times, it was generally assumed that Swedish Vikings for some time had the political power in Hedeby — a hypothesis which has later come into disrepute — but which made it possible for Hesselman
The origin of the Scandinavian languages 229
to attribute a major role to the Swedes also in the linguistic process. He suggests, however, that the e > ja change may have arisen under non-Scandinavian influence, more specifically it would have taken inspiration from Old Frisian, Old Saxon and Old Frankish. Hesselman’s theory seems to have received rather limited attention until it was taken up and further developed by Gun Widmark almost sixty years later (Widmark 1994). Widmark discusses a number of other changes (such as the monophthongization of au to o and the “breaking” illustrated by developments like *singwa > sjunga), that could be connected with the Hedeby/Birka language, which she characterizes as a “prestige language that spread over large areas” (1994: 199; my translation). As she notes, the “Hedeby Nordic” hypothesis does not really depend upon the idea of a period of Swedish supremacy in Jutland: rather, in her view, the general distribution of the phenomena she is looking at is compatible with a more general spread in Scandinavia. She also points to an important issue that Hesselman more or less manages to avoid: the later fate of “Hedeby Nordic”. Since the traits in question are no longer characteristic of the language varieties spoken in the central regions of Denmark and Sweden, it seems to follow that “Hedeby Nordic” was later superseded by some other prestige variety, which may well have spread from other centres, although presumably still in southern Scandinavia. Hedeby traits would only survive in the more peripheral regions, such as Upper Dalecarlia, Upper Norrland, and Ostrobothnia. Widmark points out that it is “somewhat shocking” that it should be precisely these areas that preserve traces of a foreign influence. Since parts of the regions in question were not populated until the 13th or 14th centuries, it is further necessary to assume that “Hedeby Nordic” survived for a long time in at least those parts of Sweden from which these areas were settled, at the same time as the new prestige variety got the upper hand in the central regions. Like Hesselman, Widmark points to the contacts between Hedeby and continental Europe, mentioning peoples such as Saxons and Frisians. She says that Hedeby must have been a town “with a considerable mixture of languages” (1994: 197). Neither Hesselman nor Widmark mention another important neighbouring people, viz. the Slavic-speaking Obotrites, who at this time inhabited the territory east of the Elbe. There is a record in the Frankish annals according to which the Danish King Godfred resettled a number of merchants from the Obotrite town Reric to Hedeby in 808 C.E. Whether this is true or not, contacts between Hedeby and the Slavic peoples are quite probable and it is not impossible that Hedeby’s population was partly Slavic-speaking. Thus, the linguistic situation in Hedeby may well have been even more complex than Hesselman and Widmark suggest. It may indeed be a “somewhat shocking” conclusion that the prestige language of the Viking Age arose in a small urban settlement with a mixed Germanic-Slavic population, situated on what is now German territory. The possibility of
230 Östen Dahl
mutual Germanic-Slavic linguistic influence has been discussed in the literature; usually, however, the contacts are assumed to have taken place in the eastern part of the Slavic territory. Of course, there is much in all this that is still very unclear. Even if the general “Hedeby Nordic” hypothesis is accepted, we do not know how far it spread and what its competitors were like. As was noted above, the close similarities between the languages in the entire Scandinavian-speaking area as they reveal themselves in the earliest written records suggest a rather late general spread of a common language variety also in the western parts. It is tempting to identify this variety with Hedeby Nordic, although there is no compelling evidence for this conclusion. We may note, however, that at the beginning of the 9th century, there was supposedly a Danish overlordship over at least parts of Norway, and there were also urban settlements in Norway at this time with connections to Hedeby (Kaupang or Skiringsal). On the other hand, the strong Danish influence in the Mälar provinces does not exclude other centres of political influence there. Many historians and archaeologists still assume a strong local power in Uppland. What can be said is that the linguistic evidence suggests that the Danes prevailed in the end. An integration of external and internal groups is also not excluded. On the whole, early Germanic society seems to have been characterized by a general fluidity with respect to political and ethnic groupings; we may assume that the linguistic situation was correspondingly unstable.
6. The third wave: East Nordic As the prestige language spread, it is bound to have undergone local modifications (Widmark 1994: 199). Part of the characteristics of “Birka Swedish” may have been the results of such modifications. At the end of the Viking Age, there may already have been significant dialect variation within the prestige language. At this point, however, there are grounds for assuming a second wave of influence on the language of the Mälar region from Denmark. The role of cultural and commercial centre passed at the millennium shift from Birka to the newly founded town of Sigtuna. As has been proposed in various recent works, this might be seen as the beginning of an intensive period of Danish political and cultural influence, one consequence of which might be the rapid spread of runic stones in Uppland (another being the introduction of Christianity). This is the period of the (mainly phonological) changes that supposedly led to the definite separation of East and West Nordic such as the monophthongization in East Nordic of the old diphthongs, e.g. ai > e. Here, even the traditional accounts describe this change as a spread, starting in Denmark and later continuing to Sweden. According to Wessén (1968: 32) the development in Uppland was rather confusing, starting with some monophthongal spellings in the beginning of the 11th century, then going back to
The origin of the Scandinavian languages
diphthongs, and ending up with monophthongs at the end of the century. Wessén ascribes the initial monophthongs to Danish influence. We may interpret these vacillations as a reflection of the competition between two language varieties where the more conservative one, representing the local tradition, for a short time managed to make its way into the written language but then had to yield to the innovative variety coming from the south. Apparently, however, the older varieties survived for a relatively long time in less central parts of Sweden, as was noted above: the diphthongs that disappeared in Uppland in the 11th century still survive in many peripheral dialects.
7.
Conclusion: Why do Swedes speak Danish?
Here is in brief what I think is the most probable scenario for the origin of the Scandinavian languages: Germanic-speaking groups arrived to the very westernmost corner of the Baltic (present-day Germany and Denmark) somewhat before the beginning of our era. A little later they expanded eastwards as far as Uppland on the north side and the Vistula estuary on the south side of the Baltic. During the ensuing half millennium, the languages of the different Germanic groups became differentiated, exactly how much we do not know. At the same time, such preGermanic groups as still remained would slowly be Germanized, a process that we shall probably never find out the details of. As Denmark emerged as the major political power in Scandinavia, the language of its leading classes spread to critical parts of the other Scandinavian countries, most probably in several waves. The result was a relatively homogeneous language situation at the beginning of the historical period, which has been mistakenly extrapolated backwards in the traditional account of the history of the Scandinavian languages. Thus, we upend the old claim that “the same language was spoken all over Scandinavia with only minor variations up to the Viking Age and then gradually split up”, concluding that the previously heterogeneous linguistic situation in Scandinavia up to the Viking Age was replaced by a homogeneous language spoken at least in the central parts of each country. The Baltic Sea plays a crucial role in this account, in particular the rise of networks of urban settlements at its rim. There is of course a continuation, in the ensuing development of the Hanseatic system, which might be seen as setting the scene for the creation of the modern standard Scandinavian languages, which are again much more similar to each other than they ought to be if they had simply developed on their own out of the medieval languages spoken in the respective countries. But that is another story.
231
232 Östen Dahl
Notes * In this paper, I have ventured into fields that are rather far from the ones where I have been working earlier. I am grateful to a number of people who know more than I do about these subjects and who have gracefully tried to keep me from going too far astray: Birgit Arrhenius, Claes-Christian Elert, Rune Palm, Staffan Hellberg, Åke Hyenstrand, Barbro Söderberg, Tore Janson. I hope the final result will not disappoint them too much. 1. Noreen says “har nästan 6,000-åriga anor”. Anor is literally “forbears” but the phrase Noreen uses has a vaguer significance — it is not entirely clear if it should be understood as “history” or “history and pre-history”. 2. The spread of Nordic (or Swedish) to Finland and Estonia is now assumed to have taken place much later, so these areas are no longer included in the Proto-Nordic territory. 3. The primary source for the quotation is Noreen’s article “Nordiska språk” in the encyclopedia Nordisk familjebok, but almost exactly the same formulations (although in German) are found in Noreen (1913). 4. “The core Swedish provinces” is not a paraphrase of “the cradle of the Swedish nation” but is only a convenient way of referring to Svealand and those parts of Götaland that belonged to Sweden before the 17th century (see Map 1). 5. It has been suggested (e.g. Koivulehto 1976) that some Germanic loan-words must have entered Baltic Finnic as early as the Bronze Age, which would, of course, entail a rather different chronology. Without really entering this issue, which is certainly outside my competence, I would like to note that the arguments rest on the correctness of the traditional datings of the Finnic languages. 6. The most popular hypothesis about the Goths is of course that they “came from Scandinavia”, or more specifically, from the Swedish province of Västergötland. It may be noted in this connection, however, that what Jordanes originally said was that they came from “Scandza”, which might equally well be taken to mean “Scania”. A look at the map shows that a stop-over in Scania is quite compatible with a general expansion eastwards in the Baltic. Going via Västergötland, on the other hand, would mean a more radical detour. (See Ellegård 1986 for a critical discussion of the traditional view of the Goths.) 7. It is tempting to suggest Småland as a candidate for a region which may have been germanized relatively late. One of the few relatively uncontroversial ethnic names quoted by Jordanes is finnaithae. Later Runic stones have almost identical forms (finaiþi or finhiþi) for what is assumed to be the “small land” today called Finnveden, one of the areas assumed to have had a specific culture during this period (Hyenstrand 1996: 30). The first element of the name might, of course, be taken to indicate a Finnic population; Hellqvist (1939: 211), however, suggests that finninitially referred to the non-Indo-European original population of Scandinavia in general. But all this remains on the level of speculation, however fascinating. 8. Again, a look at a map is useful. The habit of thinking of Scandinavia as a natural unit cut off from the rest of the world makes one forget how narrow the Baltic is west of Bornholm, in particular in the western half of this portion. 9. It is actually possible that even local people used a non-local standard for writing. Consider e.g. the case of the Dalecarlian runes. In some parts of the province of Dalecarlia (Dalarna), in particular the parish of Älvdalen, runes were used until quite recently. The same parts of Dalecarlia are famous for having what is probably the most deviant dialect (better, vernacular) in Sweden, not comprehensible to outsiders. Yet, most Dalecarlian runic inscriptions are written in Standard Swedish, with some minor interference from the spoken vernacular. Although the
The origin of the Scandinavian languages 233
situation is not quite parallel to that of the Migration Age — after all, the runic script was used as an alternative to the Latin one, it illustrates the strong impact of written standards. 10. For thirty years, Makaev’s work was virtually inaccessible to Western scholars, and most people who have cited him have most probably only read Antonsen’s review of it (Antonsen 1968). In fact, Makaev’s discussion of the nature of the koiné is rather brief and he has hardly anything concrete to say about the spoken language. Makaev’s koiné idea is not wholly original. In 1949, a Russian translation of Wessén (1941) was published. According to the then prevailing custom, it was equipped with a preface in which Soviet readers were warned of all the reactionary mistakes in Wessén’s teachings. The preface was written by the Soviet linguist S. Kacnel’son in the spirit of Ja. Marr, whose theories were at the time still the official truth in Soviet linguistics but were to be denounced by Stalin just the following year. Talking about the idea of a common Scandinavian protolanguage, Kacnel’son says that it would be an unforgivable mistake to identify it with the language of the early Runic inscriptions. Among other things, he says, one has to take into account the sacred character of the inscriptions, which shows that “we are dealing with a special caste language of priests or scribes, cultivated in a special environment and spread over a relatively large territory with a population that talked different languages and dialects” (1949: 13; my transl.). Going further back, the basic idea was formulated as early as 1905 by Bugge (1905–13, 215), who says that his idea that the Older Futhark inscriptions were created by peripathetic scribes of Herulic origin provides “a better explanation than that previously given of the fact that the language of the inscriptions from prehistoric times everywhere in Scandinavia is the same, in spite of the fact that the inscriptions are so far from each other in space” (my transl.). (Makaev does acknowledge Bugge’s contribution.) 11. In a third case, Krause (1971) does acknowledge the possibility of dialect differences. He notes that the first syllable of the verb form tôjeka ‘I prepare’ on the Noleby stone from Västergötland suggests a monophthongization of au to o “unter dem Einfluss des Westgötischen” (!), taking this as an indication of dialectal differences showing themselves in “late Proto-Nordic”. 12. Here is an example of the shakiness of judgements about the linguistic affinities of runic inscriptions. One of the few Swedish inscriptions written in the Older Futhark, the Skärkind stone from the province of Östergötland, contains the single word skiþaleubar, the first part of which is interpreted by Antonsen (1975: 49) as meaning ‘judicial decision’ (yielding ‘one who loves justice’ as the interpretation of the whole), whereas Krause/Jankuhn (1966: 195) assume that what is meant is a word skinþa ‘fur’ where the -n- is deleted due to a “runic spelling rule”. It may be noted that neither of these interpretations is supported by evidence from other Older Futhark inscriptions. Still, Krause and Jankuhn give the categorical verdict: “Die Sprache ist gut urnordisch”, using the hypothesized cluster -nþ- (which was reduced to -nn- at later stages of the development of Scandinavian) as evidence. 13. Another stone whose dating is controversial is the Ellestad stone from the province of Östergötland. The first line starts with eka, interpreted as ‘I’. Again, the final -a creates problems: there are about 25 instances of ek ‘I’ in initial position among the Older Futhark inscriptions, making this one of the most well-documented grammatical morphemes in Older Runic, but the final -a otherwise only shows up when the pronoun is enclitic, in four cases. One of those is also on the Ellestad stone, another on the Noleby stone from Västergötland, and the two remaining ones are from Sjælland and Scania. Krause (1966) thinks that the person who wrote the runes on the Ellestad was trying to imitate an archaic form of the language; in Krause (1971) he assumes that the stone is from 800, while Antonsen dates it to before 600. Again, the possibility that eka correctly reflects a local form is not considered.
234 Östen Dahl
14. Birkmann (1995) thus puts the stones from Björketorp and Istaby in Blekinge and the stone from Eikeland in Rogaland (Norway) in order on the basis of their language, although the distance between the locations is at least 700 kms. 15. Standard Danish has later got rid of the distinction between the two tones, but it is still reflected in the distribution of stød. 16. As Widmark 1994 notes, this is clearly an exaggeration: the northern border of Scandinavianspeaking settlements most probably did not go as far north as Överkalix at this time.
References Antonsen, Elmer H. 1968. Review of Makaev 1965. Language 44: 132–5. Antonsen, Elmer H. 1975. A concise grammar of the Older Runic inscriptions. Tübingen: Max Niemeyer. Antonsen, Elmer H. 1986. Die ältesten Runeninschriften in heutiger Sicht. In: Beck, Heinrich (ed.), Germanenprobleme in heutiger sicht, 321–343. Berlin: De Gruyter. Arndt, W. W. 1959. The performance of glottochronology in Germanic. Language 35: 180–92. Birkmann, Thomas. 1995. Von Ågedal bis Malt. Die skandinavischen Runeninschriften vom Ende des 5. bis Ende des 9. Jahrhunderts. Berlin: De Gruyter. Bugge, Sophus. 1905–13. Norges Indskrifter med de Ældre Runer. Indledning: Runeskriftens oprindelse og ældste Historie. Christiania: A. W. Brøggers Bogtrykkeri. Elert, Claes-Christian. 1993. “Hur länge har de nordiska språken talats i Sverige?”. In Wollin, Lars (ed.), Studier i svensk språkhistoria 3. Skrifter utgivna av Institutionen för nordiska språk vid Uppsala universitet 34. 69–75. Elert, Claes-Christian. 1996. Språket i södra Skandinavien under bronsåldern: finsk-ugriskt, baltiskt eller germanskt? In Åström, Patrik (ed.), Studier i svensk språkhistoria 4. mins 44. Stockholm: Inst för nordiska språk, Stockholms universitet. 77–86. Ellegård, Alvar. 1986. The ancient Goths and the concepts of tribe and migration. In: Dahlström, Gunnar, Hallberg, Paul & Holmberg, Åke (eds.), Vetenskap och omvärdering: till Curt Weibull på hundraårsdagen 19 augusti 1986, 32–62. Göteborg: Göteborgs universitet. Hellqvist, Elof. 1939. Svensk etymologisk ordbok. Lund: Gleerup. Rev. ed. [1st edition 1922]. Hesselman, Bengt. 1936. Några nynordiska dialektformer och vikingatidens historia. En undersökning i svensk och dansk språkutveckling. In: Hesselman, Bengt (ed.), Ordgeografi och språkhistoria. Nordiska texter och undersökningar 9. Stockholm and Copenhagen: Gebers and Levin & Munksgaard. 127–162. Hyenstrand, Åke. 1982. Forntida samhällsformer och arkeologiska forskningsprogram. Stockholm: Riksantikvarieämbetet. Hyenstrand, Åke. 1996. Lejonet, draken och korset. Sverige 500–1000. Lund: Studentlitteratur. Kacnel’son, S. 1949. Preface to Vessen, È. Skandinavskie jazyki. Moskva: Izdatel’stvo inostrannoj literatury. [Translation of Wessén 1941] Koivulehto, Jorma. 1976. Vanhimmista germaanisista lainakosketuksista ja niiden ikaamisesta. Virittäjä 80.33–47. Krause, Wolfgang. 1966. Die Runeninschriften im älteren Futhark I. Text. Mit Beiträgen von Herbert Jankuhn. Göttingen: Vandenhoeck & Ruprecht. Krause, Wolfgang. 1971. Die Sprache der urnordischen Runeninschriften. Heidelberg: Winter. Kuhn, Hans. 1955. Zur Gliederung der germanischen Sprachen. Zeitschrift für deutsches Altertum und deutsche Literatur 86: 1–47.
The origin of the Scandinavian languages 235
Makaev, Ènver A. 1996. The language of the oldest Runic inscriptions. A linguistic and historicalphilological analysis. Kungl. Vitterhets Historie och Antikvitets Akademien, Filologiskfilosofiska serien 21. [Translation of Jazyk drevnejšyx runicˇeskix nadpisej. Lingvisticˇeskij i istoriko-filologicˇeskij analiz. Moskva: Nauka, 1965]. Marold, Edith & Zimmermann, Christiane (eds.). 1995. Nordwestgermanisch. Berlin & New York: Walter de Gruyter. Noreen, Adolf. 1913. Geschichte der nordischen Sprachen: besonders in altnordischer Zeit. Grundriss der germanischen Philologie 4. Strassburg. Robinson, Orrin W. 1993. Old English and its closest relatives. A survey of the earliest Germanic Languages. London: Routledge. Sawyer, Peter. 1991. Kings and royal power. In: Mortensen, Peder & Rasmussen, Birgit M., Fra Stamme til Stat i Danmark 2. Høvdingesamfund og Kongemagt. Jysk Arkæologisk Selskabs Skrifter X11thI: 2. Højbjerg: Jysk Arkæologisk Selskab. Udolph, Jürgen. 1994. Namenkundliche Studien zum Germanenproblem. Berlin & New York: Walter de Gruyter. Wessén, Elias. 1930. Runinskriften på Gallehus-hornet. Några anmärkningar. Fornvännen 25: 165–74. Wessén, Elias. 1941. De nordiska språken. Stockholm: Filologiska fören. vid Stockholms högsk. Wessén, Elias. 1968. Svensk språkhistoria. I. Ljudlära och ordböjningslära. 8th edition. Stockholm: Almqvist & Wiksell. Widmark, Gun. 1994. Birkasvenskan — fanns den? Arkiv för nordisk filologi, 109, 173–216.
AUTHOR "Lars-Gunnar Larsson"
TITLE "Baltic influence on Finnic languages"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Baltic influence on Finnic languages Lars-Gunnar Larsson
1.
Contacts between Finno-Ugric and Indo-European languages
Scandinavia and the Baltic countries constitute an area, where Indo-European and Finno-Ugric languages have been in contact and influenced one another since time immemorial. The majority of Finno-Ugric languages are found in eastern Russia and western Siberia. Around the Baltic Sea they are represented primarily by the Finnic group, i.e. Estonian, Finnish, Karelian and a few other languages around the Gulf of Finland, all of them quite closely related to each other and deriving from a common language stage known as Proto-Finnic. Also Sami, the second language to be mentioned here, could equally well be regarded as a group of languages presenting considerable variation. There has been a lot of discussion about the position of Sami in the Finno-Ugric group, but nowadays the theory according to which Sami is the closest relative of the Finnic group, seems to dominate, at least among Scandinavian scholars. The question is complicated above all by the fact, that after the assumed split from Proto-Finnic, Sami has been under almost constant influence of Finnish. It is therefore quite difficult to distinguish between inherited forms, borrowed forms, and borrowed, adapted forms. As will be seen later on, this question is of great importance for the dating of the earliest Baltic influences. The third Finno-Ugric language to be mentioned here is Mordvin, which is codified as two literary languages: Erza and Moksha. Mordvin seems to represent the eastern boundary of Baltic influence on Finno-Ugric (Mägiste 1959). There are obviously Baltic loanwords in Mordvin that are not attested in Finnic, e.g. Mrd. pejel’ ‘knife’, cf. Lith. peilis. Of course, it is possible that such a loanword has totally fallen out of use in every Finnic language and dialect. It seems, however, more adequate to take this word as inconsistent with the idea that contacts with Baltic tribes in prehistoric times always would have taken place in the immediate vicinity of the Baltic Sea (cf. Kalima 1936: 191). Earlier the Baltic peoples have obviously had a wider distribution in what is now Slavic speaking Russia and Belarus (Larsson 1983: 14f. with literature). The Indoeuropean languages in the area belong to the Baltic, Germanic and Slavic groups of languages. According to the traditional opinion in Finno-Ugric studies the earliest Indoeuropean influences on Proto-Finnic were Baltic. These
238 Lars-Gunnar Larsson
were succeeded by Germanic influences and finally East Slavic. When Proto-Finnic had split up into dialects and languages, these had separate language contacts, resulting in, e.g., Swedish influence on Finnish, Russian on Veps, German on Estonian and Latvian on Livonian. Nowadays, after the work of Koivulehto, there is, however, reason to question, whether three such distinct prehistoric periods of influence can be distinguished. Furthermore, the source language of early Germanic influence has also been discussed (e.g. Sköld 1996). Scandinavian Finno-Ugristics has throughout the years paid considerable attention to the early Germanic contacts of Baltic-Finnic and Sami. This has, as a matter of fact, been one of the main interests of scholars for more than a century (Qvigstad 1893; Wiklund e.g. 1892; Karsten e.g. 1915; Collinder e.g. 1932; Sköld 1961; Koivulehto e.g. 1988). The interest of researchers has not only been directed towards the contacts of a Germanic majority language with a Finno-Ugric minority language, but the mutual influence of the Finno-Ugric languages Finnish and Sami has also been studied (e.g. Hansegård 1967). Even today, questions pertaining to the contacts between Finnish and Sami on the one hand and the majority languages Swedish or Norwegian on the other (e.g. Nesser 1986; Huss 1991; Svonni 1993) preponderate over other questions. The Russian influence has also been the object of considerable interest (e.g. Plöger 1973; Ojanen 1985), but for various reasons the rôle of the Baltic languages in the development of Proto-Finnic seems to have been less attended to. The interest of Swedish and Norwegian scholars in Germanic influences on Finnish and Sami and the interest that Russian experts in Slavic languages have paid to Slavic elements in Finnic languages have simply been stronger than that of Latvian and Lithuanian scholars. Certainly, there have been eminent Finno-Ugric researchers working in the field of the Baltic influence on Finnic, people like Thomsen, Kalima, Posti, Suhonen, Uotila and Vaba, but they seem to have been outnumbered by the experts in Germanic and Slavic linguistics. Furthermore, some of them have mostly been treating the more recent contacts between Latvian and the southernmost Finnic languages: Livonian and south Estonian dialects. Contacts between Balts and Baltic Finns have, however, a long history, going back into remote prehistoric times. The purpose of this article is to give a survey of the contacts of Baltic and Finnic languages. The earliest linguistic influence can only be investigated on the basis of the traces it has left and by means of historical-comparative methods. In such investigations the testimony of the Lithuanian language, being in general more archaic than Latvian, plays a more important rôle than in treatments of later contacts.
2.
Vocabulary
Examples of Baltic borrowings connected to innovations in the technology and culture of the Finnic peoples can be found in the terminology of animal husbandry,
Baltic influence on Finnic languages 239
e.g. Fin. härkä ‘ox’ < Balt., cf. Lith. žirgas ‘horse, steed’; Fin. jäärä ‘ram’ < Balt., cf. Lith. ˙eras ‘lamb’; Fin. vuohi ‘goat’ < Balt., cf. Lith. ožys; Est. lauk ‘blaze’ < Balt., cf. Lith. laukas ‘with a white spot on the forehead’; Fin. muli or mulipää ‘hornless (about a cow)’ < Lith. šmulas; Fin. paimen ‘shepherd’ < Balt., cf. Lith. piemuo; Fin. karsina ‘pen, box’ < Balt., cf. Lith. gardinys; Fin. heinä ‘hay’ < Balt., cf. Lith. šienas; Fin. luhta ‘marsh-meadow’ < Balt., cf. Lith. lukštas ‘Sumpfdotterblume’, Ltv. luksts ‘marsh-meadow’; Est. piim ‘milk’ < Balt. *pijimas, cf. Lith. pyti ‘give milk’; Fin. villa ‘wool’ < Balt., cf. Lith. vilna, nom.pl vilnos and Fin. vuota ‘umprepared hide’ < Balt., cf. Lith. oda ‘skin; hide’ (Kalima 1936; Larsson 1984). Borrowings of nouns denoting wild animals, on the other hand, suggest more intensive contacts between Balts and Proto-Finns. The Finno-Ugric tribes were hunters and fishers and therefore they must have been familiar with the fauna, if only from the perspective of a hunter. In those areas, where the Finno-Ugric languages were originally spoken, in their “Urheimat” (cf. Larsson 1990), there are different species of birds that are known by ordinary people, in our time as well as in earlier times: geese, thrushes and magpie. The geographical distribution of these bird species has probably not been subject to any radical changes, so there could very well have existed words denoting these birds in the language of the Finnic tribes even before their contacts with the Baltic tribes. The form *lunta could have been such an old word for ‘goose’ (Larsson 1993). Nevertheless, the speakers of Proto-Finnic borrowed Baltic words for all these birds, Fin. hanhi ‘goose’, rastas ‘thrush’ and harakka ‘magpie’. Furthermore, such Baltic words can be shown to have been borrowed once more: in their Latvian shape, two of them were later taken over by the Livonians, Liv. strazd ‘thrush’ and žagat ‘magpie’ (cf. Larsson 1981 for an extensive list of examples). Further evidence of the intensive contacts between Proto-Finns and Balts is provided by the borrowings of body-part terms, e.g. Fin. kaula ‘neck’ < Balt., cf. Lith. kaklas; Fin. napa ‘navel’ < Balt., cf. Ltv. naba; Fin. reisi ‘thigh’ < Balt., cf. Lith. rietas and Fin. hammas ‘tooth’ < Balt., cf. Lith. žambas ‘(sharp) corner, edge’, Ltv. zuobs ‘tooth’ (Kalima 1936). In principle, there is certainly a possibility that the Proto-Finns took over from the Balts a new way of dividing the semantic field of the body. The Russian word noga covers a part of the body that in Swedish and many other Germanic languages is denoted by two separate words: Swd. fot and ben, Grm. Fuß and Bein etc. In the Baltic loanwords above this was, however, obviously not the case. Instead, one gets the impression that indigenous Finno-Ugric words have simply been replaced by Baltic words. The Finno-Ugric word for ‘tooth’, e.g., attested in ErzaMrd. pe^, Kom. pin´, Hng. fog, has in Finnish been superseded by the Baltic loanword, Fin. hammas, and the old Finno-Ugric word only denotes the tooth of a hay-rake. The taking over of terms for parts of the body and for relatives is understandable, if one reckons with an intensive cultural influence, leading to marriages across the Proto-Finnic/Baltic linguistic border and thus to bilingualism.
240 Lars-Gunnar Larsson
The old Baltic loanwords in Finnic languages clearly support such an assumption. Loanwords like Fin. heimo ‘tribe’, cf. Lith. šeima ‘family’, Fin. morsian ‘bride, fiancée’, cf. Lith. marti ‘daughter-in-law; bride’ and Fin. sisar ‘sister’, cf. Lith. sesuo, seem to be significant evidence for such contacts. Even Baltic loanwords denoting matters that were most probably well known to the speakers of Proto-Finnic, could hardly be explained in any other way than by assuming heavy cultural influences and bilingualism. This is valid for words such as, e.g., Fin. halla ‘frost’, cf. Lith. šalna; Fin. hanhi ‘goose’, cf. Lith. ža˛sis; Fin. harakka ‘magpie’, cf. Lith. šarka; Fin. käki ‘cuckoo’, cf. Lith. gege˙, Est. vähk, vähi ‘crayfish’, cf. Lith. ve˙žys; Fin. lohi ‘salmon’, cf. Lith. lašiša; Fin. käärme ‘snake’, cf. Lith. kirmis ‘worm’; Fin. kaarna ‘bark’, cf. Lith. karna ‘bast, bass’. Even the word Fin. metsä ‘forest’ is once again regarded a Baltic loanword by some scholars, cf. Lith. medis ‘tree’ (SSA s.v.).
3.
Baltic loanwords in Sami
The hypothesis that loanwords such as the examples above had already been taken over into Proto-Finnic is based on two arguments: their distribution and their sound shape. The words belonging to this old stratum are found in all Finnic languages in the same way as inherited Finno-Ugric words; in neither case is every word attested in every language. The general picture of their spread indicates, however, that the words in question existed in Proto-Finnic. These old Baltic loanwords have participated in the same Proto-Finnic sound-changes as inherited Finno-Ugric words have. Consequently, Baltic loanwords such as Fin. heinä ‘hay’ and Fin. morsian bride, fiancée’ tell us about very old and very close contacts between Balts and Proto-Finns. Linguistics has no reliable way of dating these prehistoric contacts (Larsson 1981). Estimates have been hundreds and even thousands of years apart. To make the hypotheses more reliable, linguistics, archaeology and other disciplines must cooperate. Such a cooperation is, however, not uncomplicated. The advantage of archaeology is its ways of giving more or less absolute dating by means of methods like C-14 and dendrochronology. The problem is to connect an archaeological artefact with linguistic material. Von Hertzen (1973: 85) tries to combine the (assumedly) Baltic boat-axe culture with the old Baltic loanword attested in Fin. kirves ‘axe’, which could open up a way of dating the contacts. The basis of his theory is that this word originally denoted a new type of axe, taken over — together with the word — from Baltic peoples. This is, however, most uncertain, since terms for parts of the body and well-known objects in nature were also taken over. The linguistic material tells about a general Baltic influence, but it does not allow detailed conclusions. Even if the absolute dating of the Baltic contacts is uncertain from a linguistic
Baltic influence on Finnic languages 241
point of view, there are linguistic arguments as to the time when these contacts began. This is connected with the question of the position of Sami among the Finno-Ugric languages. The origin of the Sami people and their language has been the subject of extensive dispute in Uralic studies. Here I would like to underline the confusion that the mixing up of people and language has caused. In his excellent handbook on Sami language history, Korhonen (1981: 23f.) mentions three theories that have dominated and still dominate the discussions about the position of the Sami language among the other Finno-Ugric languages. The first possibility is that Sami has developed from a hypothetical Early Proto-Finnish language stage, i.e. is very closely related to Finnish and other Finnic languages. The second possibility is that the similarities between Sami and Finnic languages mainly depend on Finnish influences on Sami, whereas the original affinity of Sami within the Finno-Ugric group cannot be ascertained. The third possibility is the “Proto-Lapp” theory, which takes its starting point in the anthropological differences between the Sami people and the Finnic peoples that contrast with their linguistic similarities. To solve this opposition it is claimed that the Sami people once, in distant prehistoric times, exchanged their original language and took over Proto-Finnic. Now, it seems to have been constantly overlooked that this third theory about an assumed but not attested “Proto-Lapp” language of the Sami people is not founded on any linguistic evidence whatsoever. It is a theory brought forward as an explanation of anthropological differences between two groups of people. Not one single linguistic circumstance in the Sami language supports this assumption of a language shift. Furthermore, if one were to explain anthropological differences between speakers of Sami and speakers of Finnic languages by assuming a language shift, then it should — as pointed out by Collinder (1945: 143) — first of all be discussed which one of the groups that has changed its language. Modern genetic research does not support the theory of a language shift of the Sami (Sajantila et al. 1995: 47ff.). Nevertheless, I insist that no linguistic circumstance demands such a theory and that it is the task of linguistics to deal with linguistic matters. What is of great interest in this connection is the existence of Baltic loanwords in Sami. Some of these words in Sami show the same development as the inherent Finno-Ugric words of Sami. Such words are, e.g., North-Sami suoldni ‘dew; haze, mist, steam’/Fin. halla ‘frost’ < Balt., cf. Lith. šalna ‘frost’; North-Sami suoidni ‘herb-like flowering plant; hay’/Fin. heinä ‘hay’ < Balt., cf. Lith. šienas; North-Sami luossa ‘salmon’/Fin. lohi < Balt., cf. Lith. lašiša. Ever since Thomsen’s (1890) investigation these loanwords are regarded as very early indirect borrowings into Sami, that is to say, these Baltic words are assumed to have been taken over by Proto-Finnic and later borrowed from that language by Proto-Sami. This should — it is assumed — have happened at such an early stage in the development of ProtoFinnic, that no differences in their linguistic form would show the Proto-Finnic
242 Lars-Gunnar Larsson
origin of the loanword. In some other words, there are traces of Proto-Finnic sound changes, e.g. Fin. silta ‘bridge’ < Baltic, cf. Lith. tiltas ‘bridge’. Here the combination *ti has developed into si in Proto-Finnic and the Sami form of the word, North Sami šaldi, reflects this Finnic sibilant. Since Proto-Sami did not take part in the change *ti > si, a direct borrowing into Sami would have resulted in the form *daldi. Such an example can be clearly judged on the basis of linguistic criteria. The theory of borrowing via Proto-Finnic, current since Thomsen’s times, should be confronted with the theory of an Early Proto-Finnic language stage. As already mentioned, this theory argues that Proto-Sami and Proto-Finnic developed from one common source language. Baltic loanwords in Sami that show no sign of indirect borrowing via Finnish but expose the sound-changes of inherited words indicate that the Baltic influence began in Early Proto-Finnic times. This takes the first Baltic contacts back to an even more remote time than has been traditionally assumed. From a strictly linguistic point of view it seems strange that Baltic words in Sami, showing no Proto-Finnic fingerprint at all in their sound shape, have been regarded for such a long time as indirect borrowings. Obviously, this idea is not founded on linguistic evidence but on the “Proto-Lapp” theory, i.e. on anthropological theories, and upon a wish to differentiate between the Finnic and Sami peoples. Leaving comprehensive hypotheses about the prehistory of Sami and Baltic-Finnic peoples aside, a linguist can only state that the oldest Baltic loanwords of Sami behave in the same way as inherited Finno-Ugric words. This seems to me to be a clear indication that the earliest Baltic influences date back to the Early Proto-Finnic period. This way of looking at the problem certainly presupposes that the similarities between Sami and the Finnic languages are, in fact, due to a close relationship between these languages and not due to Finnic influences on Sami. The likelihood of this opinion has grown thanks to the works of, first and foremost, Erkki Itkonen, but it is still one of two theories. The other theory, current e.g. in Hungarian research, must not be disregarded. Summing up: the fact that some of the Baltic loanwords in Sami behave like inherited Finno-Ugric words tells us that contact began at a very early time. The character of the loanwords in Finnic languages bears witness of profound Baltic influence on the speakers of Proto-Finnic and close contacts between the speakers of these languages. In fact, words for parts of the body tend to be preserved in language (cf. Larsson 1996b) and if speakers of a language exchange such inherited words, it is not without good reason. Intensive influence is one such reason. The fact that Finnic languages have a larger number of Baltic loanwords than Sami could be explained by the change of culture of the speakers of Proto-Finnic, whereas the Sami people maintained their traditional hunting and fishing (Larsson 1981: 50).
Baltic influence on Finnic languages 243
4. Influence on the Proto-Finnic sound system? When the lexical influence is as profound as in the case of Baltic and Proto-Finnic, it is reasonable to assume traces of Baltic influence also in other parts of the ProtoFinnic language. In phonology, the change of *ti > si of Proto-Finnic is regarded by Posti (1953: 54f.) as a result of the Baltic influence. According to traditional opinion this change is regular both in inherited Finno-Ugric words and in Baltic loanwords but not in later strata of words, cf. Fin. syvä ‘deep’ /NSam. davvi ‘north’; Fin. sika ‘pig’ /ErzaMrd tuvo, from *tiwä and *tika respectively, and Fin. silta ‘bridge’ < Balt., cf. Lith. tiltas. As the reason for this change, Posti (1953: 54f.) suggests the palatalization of consonants in Baltic languages, Lith. tiltas ‘bridge’, e.g., being pronounced [t’i-]. The palatalized [t’], alien to the Proto-Finnic phonological system, was supposedly regarded as prestigious by the speakers of Proto-Finnic — in the same way as Baltic words for familiar notions — and therefore they imitated it but pronounced it more like an affricate [ts] from which the sibilant s later developed. Posti (1953: 54) is well aware that his hypothesis depends on the age of the palatalization of plosives before front vowels in Baltic languages. He bases his theory on Vaillant, according to whom the palatalization in this case would have been at hand in Proto-Balto-Slavic or even in periods before that. Today this hypothesis does not appear to be quite that uncomplicated. First of all, a development, like that assumed by Posti, i.e. *ti > *t’i > *t’s´i, has taken place also in East Baltic languages, cf. Lith. žodis ‘word’ (NSg.): žodžio (GSg.). There are, however, strong arguments against the high age of this development, e.g. writings of Lithuanian place-names in older sources, which do not show any affricate. This “indicates that in the 13th c[entury], perhaps even in the 14th c[entury, A.D.!]” they were still pronounced as plosives (Zinkevicˇius 1996: 99). Furthermore, the representation of this sound combination in the Žemaitish dialects of Lithuanian does not indicate that the change is of Proto-Baltic origin. This is probably supported by the testimony of Prussian texts — probably since it is difficult to tell what sound the writing in Prussian represents (Stang 1966: 103ff.). The early palatalization of *t’ in Baltic assumed by Posti on the basis of Vaillant consequently seems rather uncertain. The development of the affricate in Lithuanian is obviously a very late phenomenon and even if Stang (1966: 88f.) does not deny the possibility of *t’ in Proto-Baltic, it seems to me to be a far more important argument that palatalization is a spreading phenomenon. There is reason to believe that palatalization of consonants was not common in Proto-Baltic but spread in later times under the influence of Slavic languages. If this is correct, then Posti (1953) in his classic article has not given any example of Baltic influence on the phonological system of ProtoFinnic but several examples of Germanic influences.
244 Lars-Gunnar Larsson
5.
Baltic influence on Proto-Finnic syntax
Loanwords are not simple additions to the inherited stock of words, but they also influence the usage and functions of already existing vocabulary. This is obvious, if one considers borrowed adjectives denoting colours, e.g. Fin. keltainen ‘yellow’ < Balt., cf. Lith. geltas and Fin. harmaa ‘grey’ < Balt., cf. Lith. širmas. There are no natural borderlines in the spectrum of colours and the taking over of an adjective denoting a colour leads to restructuring of this semantic field. A good example of the close connection between a loanword and changes in syntax is Fin. ja ‘and’. Originally, Finno-Ugric languages had no conjunctions. Instead of subordinate clauses, nominal constructions were used, cf. Fin. Kallen tullessa Pekka lähti ‘when Kalle arrived, Pekka left’, lit. ‘in Kalle’s coming…’, and to express the co-ordination of elements clauses could be repeated, cf. Chanty ta]k6 wet6s, n´o˘h6s wet6s ‘he killed squirrels and sables’, lit. ‘squirrel he killed, sable he killed’; Nnts nisew ha¯´s, nebew ha¯´s, næ¯w ha¯´s ‘my father, my mother and my brother died’, lit. ‘my father died, my mother died, my brother died’ (examples from Collinder 1960: 249). Against this background I regard the obvious syntactical similarities that can be found between Finnic and Baltic languages as a result of the restructuring of the inherited, FinnoUgric usage depending on the heavy Baltic influence on Proto-Finnic that is testified by the loanwords. 5.1 The case of the subject and the object The Finnic variation of nominative ~ partitive as the cases of the subject and accusative ~ partitive as the cases of the object resembles the variation of nominative ~ genitive and accusative ~ genitive respectively in Baltic languages. In contexts, where Finnic languages have the partitive, Baltic languages often have the genitive. One cannot expect any exact correspondence, since we are dealing with the traces of an ancient language contact, that took place maybe 3000–4000 years ago, and since the morphological systems of the languages in question differ from each other. The differences in morphology can be illustrated by the variable of definiteness. In the Finnish standard language, definiteness can be expressed in several ways in different contexts: by word order, by the variation nominative ~ partitive (in existential clauses) etc., but there is no unambiguous morphological means of expressing definiteness. In Lithuanian, however, there is such a means, viz. when a noun has an adjective attribute, cf. baltas namas ‘(a) white house’: baltasis namas ‘the white house’. In both instances Finnish would have valkoinen talo, and word order would express difference in definiteness. Some examples will illustrate the similarities.
Baltic influence on Finnic languages 245
Partitive/Genitive as the case of the subject: (1) a.
b.
c.
d.
(2) a.
b.
Finnish Metsässä on susia. forest:iness.sg be:3sg.pres wolf:prtvpl Lithuanian Miške vilku˛ yra. forest:loc.sg wolf:gen.pl be:3.pres ‘There are wolves in the forest’ cf. Finnish Sudet ovat metsässä. wolf:nom.pl be:3pres forest:iness.sg Lithuanian Vilkai yra miške. wolf:nom.pl be:3pres forest:loc.sg ‘The wolves are in the forest’ Finnish Minulla ei ole isää I:adess neg:3sg be father:prtv.sg Lithuanian Ne˙ra man te˙velio beneg I:dat father:gen.sg ‘I have no father’
Partitive/Genitive as case of the object: (3) a.
b.
(4) a.
b.
(5) a.
Finnish Juon vettä. drink:1sg.pres water:prtv.sg Lithuanian Geriu vandens. drink:1sg.pres water:gen.sg ‘I drink water’ Finnish Tapasin vieraita. meet:1sg.impf visitor:prtv.pl Lithuanian Radau svecˇiu˛. meet:1sg.impf visitor:gen.pl ‘I met (some) friends’ Finnish En anna isälle rahaa. neg.1sg give father:all.sg money:prtv.sg
246 Lars-Gunnar Larsson
b.
(6) a.
b.
Lithuanian Aš te˙vui neduodu pinigu˛. I father:dat.sg neg:give:1sg.pres money:gen.pl ‘I don’t give (any) money to father’ Finnish Etsitkö veistä? search:1sg.pres.int knife:prtv.sg Lithuanian Ar peilio ieškai? int knife:gen.sg search:2sg.pres ‘Do you look for the knife?’
The use of the Finnic partitive and the Baltic genitive as the case of the object is not restricted to the verb ‘search’ but can also be found at other verbs denoting an aterminative action, e.g. Fin. tarvita, Lith. reike˙ti ‘need’; Fin. odottaa, Lith. laukti ‘wait (for)’; Fin. surra, Lith. gede˙ti ‘mourn’; Fin. pyytää, Lith. prašyti ‘ask (for)’; Fin. katsoa, Lith. žiu¯re˙ti ‘look at’; Fin. opiskella, Lith. mokytis ‘learn, study’. There are certainly more such examples, but it is necessary to define each case very carefully. Vagueness in this respect would justify criticism about a Finnic taking over “en masse”. To avoid such criticism I have in my own study (Larsson 1983) tried to delimit exactly the similarities to be treated. The similarity between clauses like the examples (1)–(6) seems systematic indeed and cannot be explained as a result of pure coincidence, especially not when taking the old Baltic loanwords from Proto-Finnic into consideration. Finnish words of Baltic origin, such as morsian ‘bride’ (cf. Lith. marti), sisar ‘sister’ (cf. Lith. sesuo), reisi ‘thigh’ (cf. Lith. rietas), heimo ‘tribe’(cf. Lith. šeima ‘family’), clearly indicate close contacts across the language border in very ancient times. I do not see any other explanation for the systematic, syntactical similarities between the Baltic and Finnic languages. The distributions of old Baltic loanwords and of the use of subject and object cases are similar: both are spread fairly evenly among the Finnic languages and can be clearly distinguished from the influences that depend on later contacts. One basic result of my investigation (Larsson 1983) is that the use of the partitive as the case of the subject and object in Finnic languages is, in fact, very similar to that of the Baltic genitive, whereas it differs considerably from the distribution of the etymologically corresponding case in Mordvin. In Mordvin, a use of the object case can be shown that is in clear contrast to the use in Finnic, e.g. MrdErza Sas´ viravas´ i karmas´ ´simime piva ‘the Forest Mother came and began to drink beer’. The object here, piva ‘beer’, is in the nominative. In such a clause Finnic would definitely take a partitive object: the verb ‘drink’ in itself can have both nominative/accusative and partitive objects, but after a verb denoting movement the partitive would be preferred (Larsson 1983:139f.). Furthermore, the object word ‘beer’ is in the partitive case, unless it denotes a definite quantity.
Baltic influence on Finnic languages 247
This inevitably leads to the conclusion that the change that has taken part in Proto-Finnic has made its case syntax more similar to that of Baltic languages. Even if the circumstances in Sami may be taken as counterevidence to such an explanation, rejecting it would entail a double standard in that lexical similarities are seen as evidence of near contacts and influence whereas syntactical similarities between the same languages are ascribed to mere chance. Mordvin and Finnic form the endpoints of a scale of development and a gradual approaching to the Baltic system. The next task is to put the Sami language on this scale, taking into consideration its position within the Finno-Ugric language group, Finnish influence on Sami and its indirect and direct Baltic loanwords etc. A traditional diachronic study of language change in the Baltic geographical area seems to me to offer a better explanation than theories based on more diffuse Sprachbund ideas (cf. Mathiassen 1985: 274). The Sprachbund theory should be supported by more evidence of a mutual influence in the area (cf. Larsson in press), but the direction of the loanwords in the Baltic-Finnic language contact clearly speak in favour of a Baltic influence on Proto-Finnic, whereas relatively few words have wandered in the opposite direction. In the lexicon of Baltic languages there do exist words that probably are of Finnic origin, e.g. Lith. bure˙ ‘sail’ and laivas ‘ship, vessel, craft’, cf. Fin. purje and laiva, but these few examples are an exception. There is no doubt whatsoever that the dominating direction of influences has been from Baltic to Finnic. 5.2 The agent participle of Finnish The resemblance in distribution of the subject and object cases is not the only example where Baltic and Finnic languages go together. One more instance should be mentioned, even if it is more complicated. In Finnish, there is what is called an agent participle, ending in -mA(-) that in its function as well as its form is very similar to the passive present participle of Lithuanian (Larsson 1996a), e.g.: (7) a.
b.
c.
Finnish Isän ostama auto. father:gen.sg buy:part car Lithuanian te˙vo perkamas automobilis father:gen.sg buy:part.masc.sg car ‘the car bought by father’ Finnish Auto on isän ostama. car be:3sg.pres father:gen.sg buy:part
248 Lars-Gunnar Larsson
d.
Lithuanian Automobilis (yra) te˙vo perkamas. car (be:3pres father:gen.sg buy:part.masc.sg ‘The car is bought by father.’
There is also in other Finno-Ugric languages a participle ending in *-m(V). In many languages it has preterite meaning and can be formed from both transitive and intransitive verbs, e.g. Udmurt gyrem busy ‘a ploughed field, a field that has been ploughed’ and lyktem kišnomurt ‘the arrived lady, the lady who has arrived’. Languages with both a present and a preterite participle tend to connect the present participle with active meaning and preterite participle with passive meaning, e.g. Hungarian and Swedish. The m-participle in Mari denotes a preterite passive, e.g. in MarEast omtam poˇc´ mo ‘the door (has been) opened’. The Mari language is crucial in the explanation of Finnish ma-participles. The Mari m-participles can function as Finnish agent participles, e.g. MarEast t6j kal6k6n mond6mo ulat ‘you are forgotten by the people’, having the agent in the genitive as in Finnish, but they also differ from Finnish, e.g. in being formed also from intransitive verbs, e.g. MarEast memnan [gen.] tolmo korno ‘the road that we have come’, lit. ‘the by us come road’. So even if the material, the participle in *-mA(-), is of Finno-Ugric origin, its functions in Finnic could have been influenced by Baltic languages, thus providing us with a good parallel to the distribution of the partitive in Finnic languages. Quite a few indications point to this as an instance of Baltic influence. First of all, the mA-participle in Finnish cannot be derived from intransitive verbs as it can in other Finno-Ugric languages. It should also be observed that there is no participle with the ending -m(V) in Mordvin, although it is one of the closest cognates of Finnic. Furthermore, in descriptive Finnish grammars the participle in -mA(-) tends to end up outside the participle system. The system is instead regarded as consisting of four forms, e.g. from the verbal stem sano- ‘say’:
Present Perfect
Active
Passive
sanova ‘saying’ sanonut ‘having said’
sanottava ‘that has to be said’ sanottu ‘having been said’
Apart from these participles there is also, according to the general description of Finnish grammar (e.g. Karlsson 1979: 193, 206f.), an agent participle in -mA. The root of the trouble is, of course, the term passive in Finnish, which denotes an impersonal form (by many researchers called impersonal passive). Fin. sanotaan, the present passive of the verb ‘say’, means ‘they say, people say’, i.e. a form that can in some languages be rendered by a passive verb form, e.g. Swd. det sägs or Lith. sakyma. Such so-called passive forms can in Finnish be formed also from intransitive
Baltic influence on Finnic languages 249
verbs, e.g. tullaan ‘they come’, and the frequent use of this so-called passive has somehow turned the attention away from the agent participle. Many studies have been devoted to the question of whether there is a passive in Finnish, but they have all basically dealt with the impersonal form. Now there certainly is one real passive in Finnish, as can be seen from the example: (8) a.
b.
Finnish Kalle ostaa auton. Kalle:nom buy:3sg.pres car:acc.sg ‘Kalle buys a car’ Finnish Auto on Kallen ostama. car:nom.sg cop Kalle:gen.sg buy:part ‘The car is bought by Kalle’.
It can be questioned whether a passive construction like this, in a language that is said to have no passive, could be interpreted as a sign of its foreign (Baltic) origin. One problem is that this participle in -mA is missing in Estonian. Another problem is that the Baltic participle in -ma does not represent the most frequent IndoEuropean ending of a passive participle, even if it has, in fact, parallels in other Indo-European languages (Szemerényi 1957: 107f.). The similarity between Lithuanian and Finnish is, however, so obvious, that the possibility of a Baltic influence cannot be neglected.
6. Later, separate contacts According to the traditional picture of language contacts in the Baltic area, the period of Baltic influence on Proto-Finnic was followed by a period of Germanic influences. Later on, when both Baltic and Proto-Finnic were split up into separate languages and dialects, influences from different directions can be distinguished in the Finnic language area: Swedish influence on Finnish, Russian on Veps and Karelian, and Latvian on Livonian and south Estonian dialects. Today, Livonian is spoken on the shores of northern Courland, but, until the middle of the 19th century, it was also spoken east of the Gulf of Riga. Since Livonian has been yielding to the Latvian language for a very long time, it is easy to understand that it has been heavily influenced by Latvian and that it is, e.g., abundant in loanwords from Latvian. Kettunen’s (1938) Livonian dictionary contains some 12,000 words. More than one fifth of these are Latvian loans. In such a large number of words one will, of course, find words pertaining to most areas of human activity: professions and occupations, housing, food and clothing, personal names, fauna and flora. Verbs make up a quarter of the Latvian loanwords. In spite of the
250 Lars-Gunnar Larsson
high absolute numbers of Latvian loanwords (more than 20% of the lexicon!), these loans amount to only 5% in a running Livonian text (Suhonen 1973). The Latvian loanwords in the southern dialects of Estonian have been thoroughly investigated by Vaba (1977). He has found more than 1,100 words of Latvian origin, but only a few — 254 words — are to be found in the dialects spoken in Estonia. The absolute majority of these words are restricted to Estonian dialects spoken in Latvia (Vaba 1977: 254). When compared to the influence on Livonian this number is surprisingly low. Both in Estonian dialects and in Livonian one can find examples of words that have been borrowed twice: first into Proto-Finnic and much later from Latvian into Estonian and Livonian separately (cf. Section 2 above). As a further example of such words one could mention Est. kauss ‘wooden bowl’ (Vaba 1977: 42) and Liv. koùš ‘Schüssel’ (Suhonen 1973: 136), which was also taken over by Proto-Finnic, as is attested by Fin. kauha ‘scoop’ (Kalima 1936: 112f.). Such words bear witness to the long duration of the contacts between the Baltic and Finnic languages. In the sound system of Livonian it has been suggested, though perhaps not convincingly (Kettunen 1938: xxxviii), that the glottal stop of Livonian is due to Latvian influence. The extensive use of the dative case in Livonian is obviously connected with Latvian influence, even if the material — once again — is inherited (Kettunen 1938: xli; Laanest 1982: 169). There is one case where an influence from Latvian on the southern Finnic languages and a Swedish influence on Finnish seems to explain a difference between the Finnic languages, namely the use of reflexive possessive pronouns (Nilsson 1988). In Estonian one reflexive possessive pronoun is used for all persons, as in Lithuanian, e.g.: (10) a.
b.
(11) a.
b.
Estonian Ma loen oma raamatut. I read:1sg.pres refl.pron book:prtv.sg Lithuanian Aš skaitau savo knyga˛. I read:1sg.pres refl.pron book:acc.sg ‘I read my book’ Estonian Sina loed oma raamatut. you read:2sg.pres refl.pron book:prtv.sg Lithuanian Tu skaitai savo knyga˛. you read:2sg.pres refl.pron book:acc.sg ‘You read your book’
Baltic influence on Finnic languages
(12) a.
b.
Estonian Tema loeb oma raamatut. he/she read:3sg.pres refl.pron book:prtv.sg Lithuanian Jis skaita savo knyga˛. he read:3pres refl.pron book:acc.sg ‘He (/She) reads his (/her) book’.
This system differs considerably from the Finnish system which distinguishes the possessor only in the 3rd person, as does to the Swedish system, in cases like (13) Finnish a. Hän luki kirjaa-nsa. he read.3sg.pret book:prtv.sg-poss.3sg ‘He read in his book’ (the reader = the owner of the book) b. Hän luki hänen kirjaa-nsa. he read his book:prtv.sg-poss.3sg ‘He read in his book’ (the reader π the owner of the book).
The word oma certainly exists also in Finnish, but there it expresses possession: (minun) oma kirjani is my own book and belongs to nobody else but me. In contrast, Est. oma, as is the case also with Lith. savo, does not stress any ownership but only signals that the subject person has the book at hand.
7.
Further fields of study
The study of linguistic contacts between the Baltic and Finnic languages is by no means exhausted by the examples given above. As a matter of fact, it seems that the Baltic influences on Finnic have had less attention paid to them than the Germanic influences, one of the reasons for this being that a larger number of Finnic scholars are trained in Germanic languages. This could explain, why, e.g., there is an almost total unanimity in regarding Fin. ranta ‘beach’ and its counterparts in other Finnic languages as a Germanic loanword. Nobody can, however, tell, what sounds were omitted in the initial consonant cluster of the source language when the word was taken over: it could have been st-, cf. Swd. strand ‘beach’, as is usually assumed, but it could equally well have been k-, cf. Lith. krantas ‘bank, sea-shore’, as suggested — to my knowledge — only by Jaunius (1972: 264–265, 323) and Terent’ev (1990: 30). There are certainly still more similarities between Baltic and Finnic languages to be discovered. Baltic influence has played an important rôle in the development of the present day Finnic languages from an older Finno-Ugric structure.
251
252 Lars-Gunnar Larsson
References Collinder, Björn. 1932. Die urgermanischen Lehnwörter im Finnischen. 1. Skrifter utg. av K. Hum. Vetenskaps-Samfundet i Uppsala 28: 1. Uppsala. Collinder, Björn. 1945. The Affinities of Lapp. In: Språkvetenskapliga Sällskapets i Uppsala Förhandlingar. Uppsala Universitets Årsskrift 1945: 12. Uppsala & Leipzig. Collinder, Björn. 1960. Comparative Grammar of the Uralic Languages. Stockholm. Hansegård, Nils Erik. 1967. Recent Finnish Loanwords in Jukkasjärvi Lappish. AUU: Studia uralica et altaica upsaliensia 3. Uppsala: A&W. Hertzen, Erik von. 1973. Itämerensuomen lainasanakerrostumien ikäämisestä. In: Journal de la Sociéte Finno-Ougrienne 72, 77–105. Helsinki: Suomalais-ugrilainen seura. Huss, Leena. 1991. Simultan tvåspråkighet i svensk-finsk kontext. AUU: Studia Uralica Upsaliensia 21. Uppsala: A&W. Jaunius 1972 = Kalbininko Kazimiero Jauniaus rankrastinis palikimas. Katalogas ir publikacijos. Vilnius. Kalima, Jalo. 1936. Itämerensuomalaisten kielten balttilaiset lainasanat. Helsinki. Karlsson, Fred. 1979. Finsk grammatik. 2nd ed. SKST 339. Helsinki: Suomalaisen kirjallisuuuden seura. Karsten, Torsten Evert. 1915. Germanisch-finnische Lehnwortstudien. Ein Beitrag zu der ältesten Sprach- und Kulturgechichte der Germanen. Helsingfors. Kettunen, Lauri. 1938. Livisches Wörterbuch mit grammatischer Einleitung. Lexica societatis fenno-ugricae 5. Helsinki: Suomalais-ugrilainen seura. Koivulehto, Jorma. 1988. Lapin ja itämeren suhteesta. In: Virittäjä 92, 26–51. Helsinki. Korhonen, Mikko. 1981. Johdatus lapin kielen historiaan. SKST 370. Helsinki. Laanest, Arvo. 1982. Einführung in die ostseefinnischen Sprachen. Hamburg: Buske Verlag. Larsson, Lars-Gunnar. 1981. Three Baltic Loanwords in Finnic. On dating pre-literary linguistic material. I: Acta Universitatis Upsaliensis. Acta Societatis Linguisticae Upsaliensis. Nova series. 3: 2, 17–55. Stockholm: A&W. Larsson, Lars-Gunnar. 1983. Studien zum Partitivgebrauch in den ostseefinnischen Sprachen. AUU: Studia Uralica et Altaica Upsaliensia 15. Uppsala: A&W. Larsson, Lars-Gunnar. 1984. Estnisch piim und finnisch piimä — ein baltisches Lehnwort? In: Fenno-ugrica suecana 7, 129–140. Uppsala: Finsk-ugriska institutionen, Uppsala universitet. Larsson, Lars-Gunnar. 1990. Zum Problem der uralischen Urheimat. I: Linguistica Uralica 26, 230–238. In Memoriam Paul Ariste. Tallinn: Akad. Nauk Est. Larsson, Lars-Gunnar. 1993. Zum Problem der urfinnisch-ugrischen Vogelbezeichnungen. In: Hajdú Péter 70 éves, 255–259. Szerk. Bakró-Nagy M. & Szíj E. Budapest: MTA Nyelvtudományi intézet. Larsson, Lars-Gunnar. 1996a. Bemerkungen zur Geschichte der mA-Partizipien im Finnischen. In: FU-8. Congressus Octavus Internationalis Fenno-Ugristarum. Pars III: Sessiones sectionum. Phonologia & Morphologia, 145–152. Red. H. Leskinen et al. Jyväskylä: Moderatores. Larsson, Lars-Gunnar. 1996b. Sprachen, die dem Lauf der Zeit trotzen? In: Ünnepi könyv Mikola Tibor tiszteletére, 222–226. Szerk. Mészáros E. & Janurik T. Szeged: JATE Finnugor tanszék. Larsson, Lars-Gunnar, in press: Quantitätsausdruck und Hauptwort in den ostseefinnischen und baltischen Sprachen. Methodologische Überlegungen. In: Symposium Konfrontation und Identifikation. Göttingen 14.–17. November 1997. Mathiassen, Terje. 1985. A Discussion of the Notion ‘Sprachbund’ and its Application in the Case of the Languages in the Eastern Baltic Area (Slavic, Baltic, and West Finnish). In: M. Flier & D. Worth (eds.), International Journal of Slavic Linguistics and Poetics 31–32. Slavic
Baltic influence on Finnic languages 253
Linguistics, Poetics, and Cultural History. In Honor of Henrik Birnbaum on his Sixtieth Birthday 13 December 1985, 273–281. ’s-Gravenhage. Mägiste, Julius. 1959. Gibt es im Tscheremissischen baltische Lehnwörter? In: Ural-Altaische Jahrbücher 31, 169–176. Wiesbaden: Otto Harrassowitz. Nesser, Anne. 1986. “Se oli kurja käya koulusa”. Bruket av inre och yttre lokalkasus i sverigefinska skolbarns språk. AUU: Studia Uralica et Altaica Upsaliensia 16. Uppsala. Nilsson, Torbjörn. 1988. Reflexive Possessive Pronouns in the East Baltic Area. In: Ural-Altaische Jahrbücher. N.F. 8, 18–37. Wiesbaden: Otto Harrassowitz. Ojanen, Muusa. 1985. Adjektiivikategoria venäläis-lyydiläisissä kontakteissa. Lingvistinen interferenssitutkimus. MSFOu 188. Helsinki: Suomalais-ugrilainen seura. Plöger, Angela. 1973. Die russischen Lehnwörter der finnischen Schriftsprache. Veröffentlichungen der Societas Uralo-Altaica 8. Wiesbaden: Otto Harrassowitz. Posti, Lauri. 1953. From Pre-Finnic to Late Proto-Finnic. In: Finnisch-ugrische Forschungen 31, 1–91. Helsinki. Qvigstad, Just Knud, 1893: Nordische Lehnwörter im Lappischen. Christiania VidenskabsSelskabs Forhandlinger for 1893: 1. Christiania. Sajantila, Antti et al. 1995. Genes and Languages in Europe: An Analysis in Mitochondrial Lineages. In: Genome Research 5, 42–52. Cold Spring Harbor Laboratory Press. Sköld, Tryggve. 1961. Die Kriterien der urnordischen Lehnwörter im Lappischen. 1. Skrifter utg. av Institutionen för nordiska språk vid Uppsala universitet. 8. Uppsala: Institutionen för nordiska språk vid Uppsala universitet. Sköld, Tryggve. 1996. Finn. meri ‘Meer’. In: Lapponica et Uralica. 100 Jahre finnisch-ugrischer Unterricht an der Universität Uppsala. Vorträge am Jubiläumssymposium, 311–316. Hrsg. von L.-G. Larsson. AUU: Studia Uralica Upsaliensia 26. Uppsala. SSA = Suomen sanojen alkuperä. Etymologinen sanakirja. 1-. 1992-. Helsinki: Suomalaisen kirjallisuuden seura. Kotimaisten kielten tutkimuskeskus. Stang, Chr. S. 1966. Vergleichende Grammatik der baltischen Sprachen. Oslo, Bergen & Tromsö: Universitetsforlaget. Suhonen, Seppo. 1973. Die jungen lettischen Lehnwörter im Livischen. MSFOu 154. Helsinki: Suomalais-ugrilainen seura. Svonni, Mikael. 1993. Samiska skolbarns samiska. En undersökning av minoritetsspråksbehärskning i en språkbyteskontext. Acta Universitatis Umensis. Umeå Studies in the Humanities 113. Umeå. Szemerényi, Oswald. 1957. The Problem of Balto-Slavic Unity. In: Kratylos 2, 97–123. Wiesbaden. Terent’ev, V. A. 1990. Corrections to the ‘Suomen kielen etymologinen sanakirja’ concerning Germanic, Baltic and Slavic Loanwords. In: Uralo-Indogermanica. II. Materialy 3-ej baltoslavjanskoj konferentsii, 18–22 ijunja 1990 g, 30–32. Moskva: Akademija Nauk SSSR. Thomsen, Vilhelm, 1890: Beröringer mellem de finske og de baltiske (litauisk-lettiske) Sprog. København. Vaba, Lembit. 1977. Läti laensõnad eesti keeles. Tallinn: Valgus. Wiklund, Karl Bernhard, 1892: Die nordischen lehnwörter in den russisch-lappischen dialekten. In: Journal de la Société Finno-Ougrienne 10. Helsingfors: Suomalais-ugrilainen seura. Zinkevicˇius, Zigmas. 1996. The History of the Lithuanian Language. Vilnius: Institute of Lithuanian language.
DOCINFO
AUTHOR ""
TITLE "Contact phenomena in minor Circum-Baltic languages"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Part 3
Contact phenomena in minor Circum-Baltic languages
AUTHOR "Stefan M. Pugh"
TITLE "The role of language contact in the formation of Karelian, past and present"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
The role of language contact in the formation of Karelian, past and present Stefan M. Pugh
1.
History and the Linguistic Context
The earliest history of the Karelians is one common to the Finnic group as a whole: that is, the Finns, Karelians, Estonians, Vepsians, Ingrians, Livonians, and Votes. The exact time of their appearance in the Baltic region cannot be identified with any great certainty; thus, while some scholars assume their arrival to have taken place as early as the III or IV millennium BCE, others place the largest waves of settlement in the first millennium BCE. Contact between Baltic Finns and Eastern Slavs (the Slavic branch to which Russian belongs) can be posited as early as the sixth century CE. From the earliest days the relationship between (Proto-) Russians and (Proto-) Karelians was one of apparently peaceful coexistence; the northern territories which they jointly inhabited were sparsely populated, and there was no evident conflict between them. Indeed, as the Karelians lived alongside Russians, they adapted to the growing Russian presence by participating in the Russian cultural sphere: they adopted (Russian) Orthodox Christianity and presumably spoke Russian — judging by the large number of early Russian borrowings in their language. By the beginning of the twentieth century, after a millennium of Russian growth and development, the position of the Karelian people gradually shifted from one of coexistence to de facto subordination (but without suppression) as an ethnic and linguistic minority within a large and powerful state. Centuries of contact between the (Proto-) Russian and Karelian peoples have left a distinct mark on the language of the latter. To the casual reader or listener, this mark will be perceived as lexical in nature, as the spoken — and now written — language of Karelians contains a vast number of lexemes borrowed from Russian. At the very earliest stages such borrowings would have been limited to semantic spheres specific to the Russian culture or to technological advances made by
258 Stefan M. Pugh
Russians, for which native Finnic words were lacking. The extensive and intensive nature of language contact between these peoples led very quickly, however, to the presence of massive numbers of Russian lexemes, from every semantic sphere, in the everyday speech of the Baltic Finns of Russia. Given that the Karelians are and have been bilingual in their L1 and Russian, it is extremely easy for Russian forms to be used in L1 without causing any difficulties of understanding. The Karelians who attended school were educated in Russian — the only language of education at all levels before the present day. Not all words of Russian origin used in Karelian can be termed ‘borrowings’, of course, and any number of labels can be attached to lexemes casually used in speech. I have applied the term ‘usings’ to such forms: a ‘using’ is a lexical form that is in use in the L1 of a given speaker, but that cannot with certainty be labelled a borrowing. A test that might confirm that a particular lexeme has achieved the status of a borrowing is its occurrence in a number of dialects of the same language, or in the language of a significant number of speakers of the same dialect. The investigative linguist who goes beyond the lexicon discovers that evidence of Russian influence on a language like Karelian extends to much deeper levels of the language, however: to phonology, morphology, and even to word-formation; in the following analysis I shall briefly examine each of these linguistic components in turn. The present day is the subject of the final section of the study, as it is concerned not only with the current state of the language, but with the direction it appears to be taking given current attempts to codify it, use it in newspapers and literature, and use it as a medium of instruction in schools.
2.
The Lexicon
2.1 The oldest layers of Russian borrowings clearly reflect the nature of the early Russian-Karelian relationship, as borrowings include such culturally important notions as ‘cross’ (Krl risti) and ‘pagan’ (Krl pakana), etc. Interestingly, the old Finnic word for ‘Bible’, based on ORus gramota ‘writing, written document’ has been superseded by Krl bibl’ii, based on Rus biblija (cf. Fin raamattu ’Bible’). This relationship continued and intensified over a period of centuries as the populations became more mixed — but mainly in urban centres; the influx of Russians in large numbers to the northwest began in earnest after the foundation of St Petersburg (1703) and Petrozavodsk (1777). With the growth of towns came changes to the early Finnic way of life; as a consequence, the development of the Karelian lexicon now came to be centred on such semantic fields as agriculture, government, the military, and the like. With the advent of the Russian revolution of 1917 came the last and most intensive phase of interaction: the imposition of absolute Soviet control over the economy and political life of the Soviet peoples led to the adoption
The role of language contact in the formation of Karelian, past and present 259
of great numbers of lexemes related to these two key spheres of life in the USSR. It must be remembered that all the while such contact was progressing, Karelian was an unwritten language, and that all lexical transferrence occurred in the spoken language. By the 1930’s many Soviet minority languages were ‘given’ the Cyrillic alphabet and codified; some of these were successful and continue to be successful literary languages, e.g., Komi and Udmurt. The success of any codification process is often a function of several factors, not least important of which is the number of speakers using a given language. Some of the codified languages, e.g., literary Komi and Udmurt, have survived and are widely used by relatively large numbers of people. The ultimate choice of alphabet was not, of course, made on purely linguistic grounds; in practical terms, writing in Udmurt in Cyrillic made the transferrence of Russian lexemes into literary Udmurt extremely easy. This somewhat cynical statement is of course valid only for lexical terms characteristic of sophisticated semantic domains: it does not apply to the hundreds or even thousands of Russian-based words that were already present in the spoken language. The attempt to codify Karelian failed, however, for a variety of reasons, political, sociolinguistic, and purely linguistic: 1) Finnish was instituted as a language of instruction and publication in 1940; 2) a suitable dialect base for the literary language was difficult — if not impossible — as literary Karelian was intended for speakers of three major dialects; 3) the geographic separation of speakers of these dialects; 4) the further dislocation and loss of speakers in World War II; and 5) the Cyrillic orthography — as it was applied to Karelian — did not suit the phonological structure of the language. For more detail on the sociolinguistic and linguistic issues involved, see Pugh (forthcoming) and Austin (1992). If one compares the lexical base of Karelian with that of, e.g., Udmurt, it was the sheer weight of the dominance and necessity of Russian that resulted in the presence of large numbers of Russian-based lexemes in Karelian, while the lexicon of Udmurt was expanded by Russian via both the bilingualism of its speakers and language planners/lexicographers. Evidence of the massive penetration of Russian lexemes into Karelian is found in all semantic spheres, far beyond those of politics and economics; thus, any given utterance by a Karelian can contain Russian-based lexemes for which native Finnic equivalents exist: (1)
Tihvin Krl a. voina ‘war’ < Rus vojna; cf. Fin sota Rukajärvi Krl b. pol’n’ittša ‘hospital’ < Rus bol’nica; cf. Fin sairaala
260 Stefan M. Pugh
All Krl dialects c. proit’i-e, proid’i-e ‘go, go through or past-INF’ < Rus projti
2.3 We also find that many Russian numerals are used in place of Karelian ones, especially large numerals in the expression of years, with a native Finnic case form at the very end: (2)
Aunus Krl a. sorok p’at -oil 40 5 -ADESS/ALL ‘in the year (19)40’ b.
d’evätnatcat -oil 19 - ADESS/ALL ‘in the year (19)19’
vuvv -el year - ADESS/ALL
vuvv -el year - ADESS/ALL
This particular phenomenon is described as reflecting a very significant level of interference by Thomason and Kaufman, and even as one of the signals of eventual language death (1991: 74–6). 2.4 Russian conjunctions (e.g. i, a ‘and’, no ‘but’, štoby ‘in order to’) and particles (že ‘emphasis’, [z]nacˇit ‘so, then’) are also heavily used in Karelian. 2.5 Even more interesting from a structural point of view is the occasional incorporation of Russian prepositions into the Karelian system, as they represent elements that have penetrated the system beyond the surface level of the lexeme, into the morpho-syntactic level: (3)
Salmi Krl a. kromi hän-däh except him-PRTV ‘except for him’, (Rus krome) b.
mesto hän-däh instead of him-PRTV ‘instead of him’ (Rus vmesto)
We note in both examples that it is the Karelian partitive that occurs with a following form, whereas it is the genitive that is required in Russian; this is not altogether surprising, as there are numerous Russian-Finnic parallel constructions in which these two cases occur as functional equivalents (e.g., the case of nouns/noun phrases following numerals; the Finnic partitive vs. Russian partitive genitive, etc.).
The role of language contact in the formation of Karelian, past and present 261
2.6 Finally, code-switching must be mentioned as crucial to Karelian speakers: Russian words (unchanged in form from the original, often including Russian case endings) and entire Russian phrases occur throughout the speech of many Karelians, especially of those living in larger and more russianized towns (for a detailed study of this phenomenon, see Pyöli 1996). Russian therefore continues to have an extremely high profile in everyday spoken Karelian, beyond the presence of Russian borrowings/usings discussed above.
3.
Phonology
Over the centuries, contact with Russian has changed the face of Karelian phonology. The two most important features of contemporary Karelian to be considered here are (1) consonantal palatalization and (2) consonant clusters. Consonant palatalization is present in a number of Baltic Finnic languages, partly as a result of internal processes, partly as a result of Russian influence; if palatalized consonants originally arose before i and j, they can now occur before e (and in Russian borrowings or usings before a, o, and u). This is now a feature not only of Russian borrowings/usings, but significantly also of native lexemes; there is some evidence that palatalized consonants are even gaining phonemic status (Pugh 1994: 47). Compare: (4)
Finnic-based forms ol’i ‘he/she was’ l’äks’imä ‘we left’ tul’i ‘he/she came’ n’enä, n’en’ä ‘nose’ n’euvonta ‘instruction’, etc.
Note that labial consonants generally resist palatalization in Karelian, although some speakers of one dialect (Tver’ Karelian) do have palatalized labials. The proliferation of consonant clusters in the Russian lexicon has led to their presence in large numbers in Karelian; they are often reduced by one element in the process of transferrence, but if the original Russian form contains a 3-element cluster, than the resulting Karelian word will still have a 2-element cluster that
262 Stefan M. Pugh
might have been unacceptable to the original Finnic system. In many cases we find that even 3-element clusters are retained in the Karelian form, sometimes side-byside with a form in which it has been reduced.The following make up but a small sampling of this widespread phenomenon: (6)
srojie
(e.g., Aunus Krl) ‘to build’ < Rus. stroit’; but strojie is also attested (Ludic, Valdai Krl) striettie (Ludic Krl) ‘to meet’ < Rus. vstretit’; note that there is still a 3-element cluster in the new form (the next stage, were there to be one, would presumably be *sriettie) tvettie (Valdai Krl) ‘to bloom’ < Rus. cvesti, stem cvet-
Finally, a feature of Karelian phonology — vowel harmony — is retained in that language. But Russian can also play a role in the realization of this feature in borrowings/usings: in verbs in particular, the presence of a prefix (which is not a structural feature of Finnic) can dictate the front-back colouration of the new Karelian form, as a function of the vowel of that prefix. Otherwise it is the desinence of vowel(s) of the root which is (are) important. (7) Examples without prefixes (where -i- is the marker of the verbal stem-type): a.
brod -i -u (root back vowel > desinential back vowel) wander -SM -3SG ‘he/she wanders’
b. sluav -i praise -SM ‘to praise’
-da (root back vowel > desinential back vowel) -INF
c.
-ü (root front vowel > desinential front vowel) -3SG
d’ern’ burst
-i -SM
d. sluuž -i serve -SM ‘to serve’
-mah -INF
(root back vowel > desinential back vowel)
e.
-mäh -INF
(root front vowel > desinential front vowel)
pl’et’t’ -i weave -SM ‘to weave’
(8) Examples with prefixes; most telling are those forms with a back vowel in the prefix, front vowel in the root (exx. c-e): the former overrule the latter:
The role of language contact in the formation of Karelian, past and present 263
a.
pere-živ -i -imö survive -SM -1PL ‘we survive’
(front prefixal vowel, front desinential vowel)
b. pr’i-miet’ -i -tää notice -SM -3PL ‘they notice’ c.
u-id’ -i leave -SM ‘he/she leaves’
d. za-d’ien’ put ‘we put’ e.
-u -3SG
(back prefixal vowel, back desinential vowel)
-i -mma -SM -1PL
na-l’ei -cˇemma pour -1PL (no SM in this stem-type) ‘we pour out’
It is striking that the prefix, a linguistic element that is not historically part of the L1 Finnic system, should play such a commanding role in this Karelian phonological domain. The inference that has to be made is that the verbal prefix is psychologically real to the Karelian speaker, and — perhaps because it is so semantically important in Russian, and therefore in Russian borrowings — it attracts phonological attention as well; perhaps more than ‘attention’, as the vowel of the prefix is clearly in control. In one respect, however, there is true interaction between the Russian prefixal and Karelian phonological systems: the vowel of the prefix does not control the quality of the desinential vowel if the prefixal vowel is i or e and a back vowel occurs in the root. In other words, these two ‘neutral’ vowels — which can occur with back or front vowels in Karelian and Finnish — will be overruled by a back vowel in the root. Compare example (f): f.
pri-nuud’ -i -u compel -SM -3SG ‘he/she compels’
(neutral-back-neutral-back)
For a more detailed study of the whole phenomenon, see Pugh 1996. As many of these and later examples reflect, Baltic Finnic consonant gradation is retained in Karelian, and in fact more widely applied than in standard Finnish; Russian-based forms with stem-final voiceless consonants take part in this process.
264 Stefan M. Pugh
4. Morphology In the strictest sense of the word, the ‘morphology’ of Karelian is unaffected by the influence of Russian — if we discount such immutable elements as conjunctions and particles. The native case system is stable (although syncretism between particular cases has occurred in various dialects, without external influence), as is the conjugational system of personal endings and tenses. However, there is a certain amount of systemic convergence between Karelian and Russian that has affected the morphological profile of Karelian to some extent, again in the verb. The existence of a large number of distinct Russian verbal stem types has meant that the transformation of Russian verbs into Karelian ones must involve a process of assimilation to native Karelian stem types. In particular, two large classes of verbs in Russian are those with stem-markers -i- and -Vj- (in which V = vowel). The key is that both of these stem types exist in the native Finnic system, but they represent considerably smaller sets of verbs in languages like Finnish (and originally in Karelian) than they do in Russian. In addition, all Russian verbs that are not of the -Vj-type (and therefore not just original Russian -i-type verbs) are assimilated into Karelian as -iverbs; this fact, and given that the Russian -Vj-class is so large, means that the profile of both types has grown considerably in Karelian. Thus, the change is not qualitative, but quantitative in nature. The following examples are given as stems (not the infinitive, as that form masks the identity of the stem for a great number of verbs), according to their verbal class in Russian: (9)
The role of language contact in the formation of Karelian, past and present 265
5.
Word-formation
Prefixes were mentioned in the context of phonology above, specifically in regard to the regulation of vowel harmony in Russian-based verbs. They could also have been mentioned in the context of verbal morphology, as the prefix is a vital element in the structure of Russian. Russian verbs are borrowed with prefixes, and the prefix is a semantically real element to speakers of Karelian; thus, I have shown that the use and distribution of Russian verbs of motion in Karelian is in fact often dictated by the meaning of a given prefix (Pugh 1991). There are also indications, however, that some Russian prefixes are able to be affixed to native Finnic stems: thus, we find pere- and do- in Vepsian, and pere- in isolated Karelian forms. Such a phenomenon is evidence of extreme systemic interference; it is impossible to tell at this point whether or not the tendency will grow or become more widespread, especially given the current attempts to codify such languages as Karelian and Vepsian (see 5 below).The following are examples of Russian prefixes found with Finnic stems: (10)
Vepsian: a. b. c. d. e. f. g.
do-panda do-soda i do-ajand pere-hüppähtada pere-panda do-r’ikot’he do-l’ämptab
Karelian h. pere-katšella i. pere-pat’t’ie
‘place as far as’ ‘finish eating’ ‘did not reach’ ‘jump across’ ‘place across’ (examples a-e from Oinas 1958) ‘finished off, killed’ ‘finish warming up’ ‘look at or examine thoroughly’ ‘thoroughly, utterly’ (examples f-i identified by SSMPugh; see Pugh forthcoming [b])
The native Finnic system of verbal (suffixal) derivation is clearly intact and productive; one comment we could make here is that the full participation of a Russian lexeme in such word-formational processes signals the complete assimilation of that lexeme into the Karelian system (and is therefore certainly a borrowing, not a using). Perusal of Karjalan Kielen Sanakirja (the Dictionary of the Karelian Language) reveals that vast numbers of Russian-based forms, especially verbs, do take part in this process (see Pugh 1992). (11)
muani> a. muan betray
-i -SM
-kse-nd-ele-FREQ+DUR (compound suffix, incorporating the notion of ‘passivity’) ‘be betrayed many times, for a long time’
266 Stefan M. Pugh
>
b.
muan -i betray -SM ‘be betrayed’
-vu-PASS
>> c.
muan -i -vu -kse-nd-elebetray -SM -PASS -FREQ+DUR ‘be betrayed many times, for a long time’
>
d.
muan -i betray -SM ‘dupe, deceive’
>> e.
>
muan -i -tbetray -SM -CAUS ‘deceive many times’
-ele-FREQ
>> f.
>
muan -i -ta betray -SM -CAUS ‘deceive many times’
-kse-nd-ele-FREQ
-t-CAUS
These are just a few of the many possible permutations that have been generated in the Karelian dialects; note that a number of different suffixes can express the same notion from dialect to dialect — and even within the same dialect.
6. The present The Karelian people finds itself at a linguistic crossroads as we speak. Since 1991 there has been a serious move to establish Karelian (as well as the much more seriously endangered Vepsian language) as a literary language, for use in literature, the media, and the schools. Primers have been prepared, in two major Karelian dialects, and are being used as of the first year of schooling in the Karelian Republic. There are, of course, immense difficulties involved in these developments, not least of which is the presence of large numbers of Russian speakers throughout the republic; this is most true of the capital, Petrozavodsk, but also of the relatively large town of Olonets, where at least a good half of the population speaks Russian as their L1 (Karelian is still commonly heard on the streets, however). Beyond the Russian presence there are financial implications — resources are scarce, especially for largescale production of books for use in schools. It is also not enough for primers and other basic language books to be introduced for the instruction of Karelian: if Karelian is to survive (at least in the schools), then it will have to become the medium of instruction for all or most basic school subjects: science, maths, and so on. For this to occur there are not only financial implications to consider (which would be substantial) but also questions of staffing; teachers must be found, trained in Karelian if they are not fluent L1 speakers already, and placed in schools, from
The role of language contact in the formation of Karelian, past and present 267
primary school up through to the high schools. This is obviously a daunting task, and only time will tell if it will succeed. At this point, however, there are also linguistic concerns. As mentioned above, materials are being prepared for speakers of two main dialects of Karelian. This was obviously the choice that had to be made: (1) codify one variant for all, which would entail reconciling features of two variants with each other for the greater good; and (2) codify two variants so that speakers would feel that this language really was theirs, and therefore could and would be usable. The disadvantage of option (1) is that the codifiers would be creating a supradialectal literary language, in a form that is actually spoken by no one Karelian. The disadvantage of (2) is clearly one of resources: two efforts are progressing instead of one, with a smaller audience for each, with greater training implications for the schools; but the fact that each variant mirrors the living language for the groups in question must outweigh the negative factors. A large concern for the future of the Karelian literary language(s) lies in the development or acquisition of new lexical items in semantic fields for which the Karelian lexical pool is unable to produce native alternatives. In other words, language planners will have to settle on which prestige language to use as the lexical source of new borrowings. In the field of popular culture, this may well turn out to be English, as the study of English is growing in popularity in Karelia as it is in the rest of the former Soviet Union. But in other fields a decision will have to be made. Logic dictates that Russian could continue to be this source: it is spoken throughout Karelia, will continue to be used in education (especially higher education), it has always been such a source (as a fact of life as well as by design in later years), and retains prestige. The latter point is not insignificant, as the Karelians are not essentially anti-Russian — as the populace might be perceived to be in such newly independent republics as Estonia or Lithuania. But in developing Karelian society, Karelians now look west, to Finland. Finland gives Karelia assistance in many ways, but especially in the area of the economy (as far as they are able to contribute, given current financial constraints in Finland) and the development or maintenance of Karelian national identity. Finnish is a prestige language, with a comparatively long literary and cultural history; it has been one of the official languages of Karelia (off and on) over the last several decades; and, what is extremely important, it is closely related to Karelian. Thus, Finnish is an obvious and easy resource. In fact, recent Karelian-language newspaper articles have indeed reflected a tendency to rely on Finnish as such as a source language. However, more importantly, the lexical items used represent no new coinages in new or developing semantic fields: the Finnish items in question were actually replacing already existing lexemes used in the spoken language that had come into the language from Russian. This phenomenon is unfortunate from the point of view of the linguist. On the one hand, it will in the end make Karelian more ‘Finnic’, if not more ‘Finnish’;
268 Stefan M. Pugh
but on the other, the replacement of lexemes in use in the spoken language with lexemes in use in the written language can ultimately result in a separation between the written and spoken forms of the language. In fact, it is not impossible to foresee the development of a diglossic situation in the medium-range future. Should this tendency become policy in any profound manner, it would mean adding diglossia to what can currently be described as ‘stable bilingualism’ (in Karelian and Russian) among many Karelians; in addition, there is no reason to suppose that there will be an end to their use of Russian, given demographic and practical considerations. This might be a recipe for eventual linguistic disaster for the maintenance of the Karelian language.
7.
Modelling language contact
There are numerous ways in which we could create a theoretical model for language contact, with diagrammes and pictures. Here I would like to reflect the RussianKarelian situation over the centuries in a very simplistic way — simplistic and yet relevant. In the following chart (chronologically, from the earliest phase at the top) K = Karelian, R = Russian, F = Finnish; when appearing in upper case, this represents the strength of the given component — vitality, stability, and degree of influence in language contact situations. Two letters appearing together, e.g. in the case of Kr, means that K is still strong, while R is playing a role in the development of K; the fact that it appears in lower case suggests less than major influence, whereas an upper case letter would suggest strong interference. If they are separate, arrows indicate in which direction the interference (however minor) is progressing. 1. 2. 3. 4. 5. 6. 7.
Early Phase (before 17th century) 17th century–1917 1917–1930’s Early 1940’s Late 1940’s–1991 1991–present The future
K < < < < < R, but becoming Kr Kr, becoming KR KR, becoming KRf then KR again KRf kR KRf KrF literary language??? KRf spoken language???
Such a description cannot of course be entirely accurate, because it concentrates on general tendencies, ignoring the stability of Karelian among older generations in the villages. In general, however, it is valid as a global perception of the state of Karelian in terms of language contact. Phase 3 recognizes the strong russifying activities of the Soviet government vis-à-vis small minorities; the addition of the component F reflects the widespread use of Finnish in Karelia (in schools and public life) until 1938, when F was banned from use (as a ‘protective’ measure against bourgeois/
The role of language contact in the formation of Karelian, past and present 269
fascist Finland!). Phase 4 adds F again, during the war years and occupation by Finnish forces. The post-war attempt to reintroduce Finnish in any thorough fashion failed in part because of the complete destruction of Finnish-language resources in 1938, therefore Phase 5 has lost F; the weakened state of K, whether a result of the Second World War, migrations, or government policy, is reflected here as well. Phase 6 recognizes the limited influence of F in the post-Soviet period, consisting currently of the replacement of some R-based K lexemes by F ones. What the future will hold is certainly up to language planners, grammarians, government support for indigenous languages, and the willingness of the people to use and maintain their language — and to pass it on to successive generations. Phase 7 reflects the potential co-existence of two forms of the language: for the literary language, it represents a move to choose F as a source for future lexical expansion, replacing R as the prime giver-language but not entirely supplanting it; given the bilingual status of Karelians, and the presence (alluded to above) of large numbers of Russians in the Republic, this could not reasonably occur. The spoken language, on the other hand, could continue as it is at present: strong R influence, a stable K component, and some F influence. Were this two-fold development to take place, the situation would clearly come under the category ‘potential diglossia’, the ultimate results of which cannot be predicted, as there are so many variables in play. In terms of the Circum-Baltic languages as a whole, Karelian (as it is currently in the process of being codified) represents one that is truly in a state of flux, to a greater extent than any other language of this region.
References Austin, Paul M. 1992. Soviet Karelian: the Language that Failed, in Slavic Review, v. 51, n.1 (Spring, 1992), pp. 16–35. de Bray, R. G. A. 1980. Guide to the East Slavonic Languages (Guide to the Slavonic Languages, Third Edition, Revised and Expanded, Part 3). Slavica Publishers: Columbus, Ohio. Kalima, Jalo. 1955. Die slavischen Lehnwörter im Ostseefinnischen. Osteuropa-Institut an der Freien Universität Berlin, Slavistische Veröffentlichungen, Band 8, Berlin. Oinas, Felix J. 1958. Russian and Eastern Balto-Finnic Linguistic Contacts, American Contributions to the Fourth International Congress of Slavists. The Hague: Mouton, 267–78. Pugh, Stefan M. 1991. A Balto-Finnic Semantic Field in Motion. Semantic and Structural Aspects of Language Contact. Finnisch-Ugrische Forschungen. Vol. L (50), 1, pp. 1–19. Pugh, Stefan M. 1992. Balto-Finnic Verbal Derivation and the Assimilation of Russian Lexemes. Linguistica Uralica, XXVIII 2, 89–101. Pugh, Stefan M. 1994. The Fennicization of Russian Verbs in Karelian and Veps: a Question of Degree, Ural-Altaische Jahrbücher, Neue Folge, Band 13, pp. 45–55. Pugh, Stefan M. 1996. Vowel Harmony in Karelian and Veps: Rules for Russian Verbal Borrowings. Papers from the Fourth World Congress for Soviet and East European Studies (Harrogate-July 1990). Volume III in the Occasional Series: Papers in Slavonic Linguistics. Ed. JI Press and FE Knowles. London. 213–223.
270 Stefan M. Pugh
Pugh, Stefan M. (forthcoming [a]). Minority L1 Literacy in Sweden and Karelia: Challenges and Strategies. Acquisition of literacy in a multilingual context: A cross-cultural perspective, in press, Erlbaum. Pyöli, Raija. 1996. Venäläistyvä Aunuksenkarjala. Joensuu. Thomason, S. G. and T. Kaufman. 1991. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press.
AUTHOR "Éva Ágnes Csató"
TITLE "Syntactic code-copying in Karaim"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Syntactic code-copying in Karaim Éva Ágnes Csató
1.
The Karaim language
For more than six hundred years, Karaim has been spoken as a community language in what is today Lithuania. Due to political measures taken by the postwar Soviet regime, the community is now dispersed and the maintenance of its language has become endangered. The number of Karaims in Lithuania is about two hundred, but only a fourth of them, mostly members of the eldest generation, still have a communicative competence in the language. Karaim belongs to the Kipchak branch of the Turkic language family, more closely to the sub-branch which also includes Karachay, Crimean Tatar and Kumyk. Three dialects of Karaim are distinguished. The eastern dialect, spoken in the Crimea until the beginning of this century, is now extinct. There are two western dialects, the northwestern Lithuanian dialect, which is often called the Troki (in Polish) or Trakai (in Lithuanian) dialect, and the southwestern dialect of Galicia in the modern Ukraine, is now spoken by only six Karaims in the town of Halich. The Lithuanian dialect shares a number of typological properties with cognate Turkic languages. It is also influenced by the non-Turkic dominant languages of the linguistic area. The adoption of new structural and lexical properties has led to significant convergence with the linguistic type dominating in the area. This study will focus on the nature of linguistic processes involved in contactinduced typological changes in general, and present the case of adpositional constructions in particular. My aim is to illustrate that processes of code-copying may introduce new grammatical patterns, i.e. that code-copying can lead to framechanging innovations at all levels of the grammar. For the following analysis, spoken material recorded recently in Lithuania has been analysed. All my informants speak the Lithuanian dialect, and their average age is above 50.1
2.
Processes of code-copying
Karaim speakers in Lithuania are multilingual, also having command of the regionally and super-regionally dominant languages: Lithuanian, Polish and Russian. A few
272 Éva Ágnes Csató
Karaims also have some fragmental knowledge of the religious language, Hebrew. The functional domain of the Karaim language is, however, restricted to everyday conversational situations in the family and with a few Karaim friends. Karaim also plays an important role in religious practice, since songs and prayers are both in Karaim and Hebrew. The speakers are trained in writing and reading Karaim, but their access to written literature is limited. Code-shifting is an everyday communicative habit of many speakers in multilingual Lithuania. The Karaim speakers would not be able to manage without it. Most families include speakers of different subsets of the dominant languages. The relatives living in Poland do not, for instance, speak Lithuanian but they understand Russian. The ones living in the Crimea understand neither Polish nor Lithuanian. Thus, the members of the families have to alternate from one code to the other according to the needs of a given speech situation, e.g. from Karaim to Polish, or from Polish to Russian. The processes I am going to discuss in this paper are not the ones involved in this type of code-shifting, i.e. shifting from one language to another, but those involved in code-copying, i.e., that a speaker of Karaim copies foreign linguistic items and properties into his Karaim speech. The notion of code-copying was introduced by Johanson (1993) and developed in several publications, see references for further titles. “The central concept of code-copying is that elements of one code are copied and the copies inserted into another code. Copies are not ‘switched’, not just juxtaposed to the elements of the other code, but adjusted to its structure in some way and to some degree, even if they are only used occasionally, as so-called ‘momentary copies’” (Johanson 1999: 39). The notation, code-copying, is used by Johanson instead of the terms “borrowing”, “interference” and “transfer” used by other authors. Partly because of their multilingual habits, and partly because of the insufficiency of the Karaim lexicon, Karaims have a well-developed habit of copying foreign lexical items. Examples (1) and (2) illustrate elements copied from Russian (< Russ) and Polish (< Pol) respectively. It is not in every case clear which Slavonic language served as the source language for copying. In such cases, the source will simply be called “Slavonic” (< Slav). (1) Nu to bašladï iš’l’˙am’˙a anar well then (< Slav) begin:past do:inf she:dat bu d’ixan’iye this respiration (< Russ) yattïrγïzdï yer üs’n’u, ˙ kuxn’˙ad’˙a lay down:DI-past floor on.postp kitchen (< Russ):loc ‘Well, then he started to give her artificial respiration, he laid her down on the floor, in the kitchen.’
Syntactic code-copying in Karaim 273
Vïzvat’ et’t’im skoranï, call (< Russ) DI-past.1sg ambulance (< Russ):acc k’el’m’it’. com:neg:A-nonpast ‘I called the ambulance, it doesn’t come.’ L’il’˙ag’˙a pazvont et’t’im, to Lilia:dat make a call (< Russ):DI-past.1sg to.part (< Slav) skoray k’el’d’i. ambulance (< Russ) come:DI-past ‘I called Lilia and the ambulance came.’ (2) Nu n’e maya γal’e avur bunu barïn, well (< Slav) what I:dat now difficult this:acc all:3poss.acc bunun ücˇ’un’ ˙ k’i t’uv’ ˙ ul’ ˙ barïn because that not all:3poss.acc m’en’ pam’entac’ et’˙am, I remember (< Pol): A-nonpast.1sg maya pl’ontac’ et’˙at’ barï. I:dat confuse (< Pol):A-nonpast all:3poss ‘Well, it is now difficult for me (to tell) it all, because I don’t remember everything, I confuse it all.’ Nu n’ecˇ’ik b’ir bala ošukac’ et’t’i well (< Slav) how a young man cheat (< Pol):DI-past b’ir tašcˇanï … Nu da ga]lax uže a girl:acc well (< Slav) and Catholic priest already (< Russ) spov’edac’ et’˙at’ alarnï, hear confession (< Pol):A-nonpast they:acc k’i pov’in’n’i ed’l’˙ar buley da buley … that ought to (< Pol):3pl be.cop:DI-past.3pl so and so ‘Well, how a young man gets a girl into trouble, and the Catholic priest hears their confession so that they ought to do so and so.’
The term “code-copying” is used here to refer to processes of creating a Karaim copy of a foreign lexical item or linguistic feature. The Karaim copy of the Slavonic lexical element is, however, never identical with the original item. It is, as a rule, both phonologically, grammatically and semantically adapted to the Karaim system. The adaptation is rule-governed. In the following example, the Karaim speaker uses a copy of the lexical unit zakuska ‘snack’ from Russian. (3) Iš’l’eyd’l’˙ar zakuskanï make:A-nonpast.3pl snack (< Russ):acc ‘They prepare the snack.’
The copied item, zakuska, is declined as a Karaim noun, and, consequently, the harmonic form of the accusative suffix, -nï, is attached to it. The copied item is
274 Éva Ágnes Csató
classified as a [−front] stem and the Karaim sound harmony rules are applied to it. This is a global copy, implying that both phonological material and certain grammatical, combinational, lexical and frequency properties are copied. Copying may also be selective, involving not all but only some selected aspects of a linguistic item. Example (4) includes an essive expression meaning ‘in the function of ’ that is formed in Karaim with the postposition -bA ‘with’. The use of this postposition with this function is the result of the selective copying of the syntactic function of the instrumental case which is used in corresponding Russian essive constructions. The function of the Russian case suffix but not the material form is copied selectively onto a Karaim postposition. The example also contains a global copy of the Russian expression vagonnï d’is’pecˇer ‘superintendent of waggons’. The selective copy and the global copy in this example together constitute a so-called mixed copy. (4) Iš’l’˙ad’im t’em’ir yolda, vagonnï work:past.1sg railway:loc waggon (< Russ adj) d’is’pecˇ’erbcˇ6. superintendent (< Russ):with.postp ‘I worked at the railway as superintendent of waggons.’
Syntactic patterns, for instance, the structure of a Polish copular sentence, may also be copied, as the following examples (5) and (6) illustrate. The predicate is introduced by the particle to in both the Polish and the Karaim sentences. (5) Polish Ten wysoki blondyn to mój brat. that tall blond man to.part my brother ‘That tall blond man is my brother.’ (Rothstein 1993: 729) (6) Karaim Bunda n’e jazγan, here what write:gan.part to astrï kart s’oz’l’˙ ˙ ar. to.part (< Slav) very old word:pl ‘The words which are written here are very old ones.’
Code-copying involves creative, productive and rule-governed processes. Codecopying strategies are a component of the language competence of Karaim speakers. Code-copying results in changes involving various aspects of the Karaim language system. The following discussion will discuss examples of contact-induced changes involving Karaim adpositional constructions.
Syntactic code-copying in Karaim 275
3.
Global and mixed copies of foreign prepositional syntax
Turkic languages exhibit postpositional phrases, which is in harmony with their typical basic SOV word order. Karaim also employs postpositions, as, for instance, üs’n’˙u ‘on’ in (7). (7) Cˇïpcˇïxlar olturadlar t’er’˙ak’l’˙ar üs’n’u. ˙ bird:pl sit:A-nonpast.pl fruit-tree:pl on.postp ‘Birds are sitting on the fruit-trees.’
The prepositional phrases used in Karaim are mostly copied from foreign languages. The following example (8) includes a global copy of a Slavic prepositional phrase, s Petrograda ‘from Petersburg’. (8) Troxta anuz yox ed’i Trakai:loc yet not existing be.cop:DI-past gramafon, a atam gramophone (< Slav) but father:1sg.poss k’el’t’ir’d’i s Petrograda bring:DI-past from.prep (< Slav) Petrograd (< Russ):gen gramafon. gramophone (< Slav) ‘There was no gramophone in Trakai, but my father brought a gramophone from Petersburg.’
Copied prepositions may also be used in mixed copies, i.e. in structures in which the copied preposition is used together with a Karaim noun. The combinational properties of the Polish preposition are copied onto the Karaim construction. Consequently, in the following example, the noun Basia (girl’s name) is in the genitive. (9) I uže bu fotograf and already (< Russ) this photographer (< Slav) turat kolo Bas’yanïn. stand:A-nonpast by (< Pol) Basia:gen ‘And this photographer is already standing near Basia.’
The mixed copy in example (10) contains the global copy of the Slavic preposition za ‘for’. The whole expression is a selective copy of an idiomatic Slavic expression meaning ‘What kind of …is it?’. (10) N’e bu za karay? what this for.prep (< Russ) Karaim ‘What kind of Karaim is this?’
As the above examples demonstrate, copied prepositions and prepositional phrases occur in Karaim. The use of such copies has not, however, changed the original typological property of Karaim. It has remained a predominantly postpositional language.
276 Éva Ágnes Csató
4. Disharmonic properties of the Karaim basic word order The other originally postpositional language spoken in the Baltic area, Estonian, has also preserved its predominantly postpositional nature. The resistance of these languages to change from postpositional to prepositional syntax is interesting given the fact that so many of their basic syntactic properties have been changed. In a previous study (Csató 1996), I have discussed syntactic changes in Karaim and concluded that its original Turkic SOV order has been changed to SVO. This might already have taken place before the Karaims settled in the Baltic area. Karaim has today word-order properties similar to those of the dominant languages spoken in Lithuania. It has adopted the word order noun+genitive (NG) as a free variant of the original genitive+noun (GN) order. In some stylistic variants, mainly in pre-war written language, the NG order may even be dominant. (11) Lithuanian Russian Polish Karaim
SVO SVO (free variation) SVO (free variation) SVO
Prep GN Prep NG Prep NG Postp/(Prep) NG/GN
AN AN AN/NA AN/(NA)
The Postp, NG, AN order in Karaim contradicts a word-order universal suggested by Greenberg (1966 and 1980), according to which prepositional languages have NG and postpositional languages GN as the basic order. The combination Postp+NG may, however, occur if and only if the adjective-noun order in the language is NA, thus Postp fi (NA fi NG). Karaim seems, however, to contradict this assumption, since it has a basic AN order, the NA order having a very low frequency. Copies of foreign prepositions and prepositional phrases are used in Karaim. One would expect that these copies would change the Karaim adpositional syntax. The fact that the use of postpositions is disharmonic with the other syntactic properties of Karaim gives further reason to assume that this component of Karaim syntax has undergone changes. In the following, I will present arguments supporting the assumption that the syntax of postpositional constructions has been changed in Karaim. The path of syntactic change in Karaim is, however, not postposition fi preposition, but rather noun-like postposition fi particle-like postpositions. Particle-like postpositions may be cliticised and thus come to constitute a phonological word together with the noun. This path is well documented cross-linguistically. There is historical evidence that case suffixes have developed from postpositions in several languages, for instance Hungarian.
Syntactic code-copying in Karaim 277
5.
The effect of selective copying on the syntactic properties of Karaim postpositions
5.1 Turkish postpositional constructions A closer look at the syntax of Karaim postpositional phrases reveals that the dominant type of postpositional phrases in Karaim is the one that is most compatible with SVO constituent ordering. I will first present the types of postpositional phrases typically found in Turkic languages by giving examples from Turkish and then argue that some constructions are more compatible with SVO word order than others. Several types of postpositional constructions can be distinguished in Turkish. I will present here two of these that have particular relevance with respect to the Karaim data. The first, very frequent, type has a structure that looks similar to possessive constructions. In spite of the similarities, there are significant differences between the two constructions (see Johanson 1974). Compare the postpositional structure in (12a) to the possessive construction in (12b). (12) Turkish a. ev(in) içinde house(:gen) in:poss.3:loc ‘in a/the house’ b. evin içi house:gen inside:3poss ‘the inside of the house’
The Turkish postposition içinde ‘in’ acts as the syntactic head of the nominal construction in (12a). The adverbial case-suffix (locative, dative or ablative) required by the semantics of the matrix construction is attached to this head. It also bears a possessive suffix agreeing with the attributed noun. The noun, ev ‘house’ can be assigned the genitive, as is usual in possessive constructions, but it can also be absent, which is an option only in postpositional constructions. This case-marking of postpositional phrases is an indicator of the altered relation between the postposition and the nominal category. Note that the structure ev içi [house inside:3poss] is a compound in Turkish. No syntactic category may intervene in either the postpositional construction without genitive-marking, e.g. ev içinde [house inside:3poss.loc], or in compounds, such as ev içi [house inside:3poss]. While the genitive-marked noun evin may take a postverbal position outside the postpositional phrase, ev without genitive-marking has a fixed position. Thus, whereas the postposition behaves as a nominal head in this type of postpositional constructions, there is a syntactic cohesion between the postposition and the attributed nominal category that is characteristic of nominal compounds. The postpositions used in the second type of postpositional constructions are morphologically unanalysable except for any synchronically segmentable possessive
278 Éva Ágnes Csató
or case suffix. Whereas the class of the first, noun-like postpositions is relatively open, this second type of particle-like postpositions constitutes a strictly closed class in Turkish, including only a few postpositions such as, için ‘for’, gibi ‘like’. Nouns may never be assigned the genitive when followed by a particle-like postposition, e.g. ev gibi ‘like a house’ and never *evin gibi [ev:gen like].2 No syntactic element may, as a rule, intervene between the noun and the postposition. Whereas a nounlike postposition may constitute a postpositional phrase on its own, i.e. without an overt attributed noun, a particle-like postposition may occur only together with a nominal category. Thus içinde [inside:3poss.loc] ‘in it’ is a well-formed expression but *Gibi ‘Like’ is normally not. The syntactic function of a particle-like postposition is to anchor a nominal phrase or a clause syntactically in the matrix construction, a function shared with case suffixes. The similarity to case-suffixes is even more conspicuous in the case of the postposition ile ‘with’, which can be cliticised to the noun and obey the morphophonological rules of sound harmony, e.g. babamla [father:poss.1sg.with] ‘with my father’, trenle [train:with] ‘by train’. Johanson (1974) observes that, in Turkic, the order of determination at the syntactic level is just the opposite of the order of determination at the morphological level. A left-branching syntax is based on a right-branching morphology.3 Types of postpositional constructions can be characterised as being closer to left-branching syntactic structures or to right-branching morphological ones. A scale may be defined by setting left-branching postpositional phrases headed by noun-like postpositions on the one extreme and right-branching postpositional phrases based on particle-like postpositions on the other. A number of intermediary sub-types, not demonstrated here, may be placed on the scale between these two extremes. Karaim has acquired right-branching syntactic properties, as demonstrated above. The left-branching properties of postpositional constructions based on noun-like postpositions are not in harmony with this new SOV order. Thus, we would expect Karaim to raise the frequency of the second type of postpositional constructions, which is not in conflict with the acquired right-branching principle. In the following, I will argue that this is actually what happens in Karaim. 5.2 Karaim postpositional constructions The dominant type of postpositional construction in Karaim is the one in which the case marking of the postposition is either standardised or missing. In addition to the postposition bïla ‘with’, the postposition usnu/üs’n’˙u ‘on, on to’ can also be cliticised to the nominal stem. They both have harmonic variants. See, for instance, the examples in (13). (13) a.
B’ir k’iš’i d’˙a n’erv’˙al’˙ar üs’n’u˙ olturmït. a person also nerve:pl on sit:neg:A-nonpast ‘Nobody annoys me [sits on my nerves].’
Syntactic code-copying in Karaim 279
b.
M’en’ macˇ’i k’ib’ik aγacˇ usnu I cat like tree on m’in’˙al’m’im. climb:pot.neg:A-nonpast.1sg ‘I cannot climb a tree like a cat.’
The form of this postposition is standardised. It is invariable in every person. In the following example (14), the postposition üs’t’˙un’˙a ‘at’ carries the fossilised third-person suffix in the expression m’en’im üs’t’˙un’˙a ‘at me’, and thus does not agree in person with the genitive attribute m’en’im [I:gen]. The postposition, which lacks a case suffix, can be used both in a locative meaning, ‘on’, and in a directional meaning, ‘onto’. Some postpositions may optionally bear a case suffix, as in example (14). The form without case suffix is, however, more frequent in spoken language. (14) S’en’ m’en’im üs’t’un’˙ ˙ a/ üs’n’u˙ you I:gen on:dat on baγïnmïs. look at:neg:A-nonpast.2sg ‘You are not looking at me.’
A dative case suffix may be an invariable part of the lexicalised form of certain postpositions. The postposition t’˙ub’˙un’˙a ‘under’ contains both a possessive and a dative suffix [under:(<3poss.dat)]. The suffixes do not, however, stand in any paradigmatic relation to other case suffixes. See example (15), in which the postposition t’˙ub’˙un’˙a ‘under’ is used, in spite of the dative suffix, with a locative meaning. (15) Yer üs’n’u˙ t’er’˙ak t’ub’ ˙ un’˙ ˙ a earth on fruit-tree under (< 3poss.dat) barïba ax. all:3poss.with white ‘The earth under the fruit-tree is completely white.’
Summing up, it can be concluded that the postpositions used in these Karaim examples have more particle-like than noun-like properties. The dominance of the particle-like postpositions can be regarded as a consequence of the adoption of right-branching syntactic properties. 5.3 Other changes induced by selective copying of right-branching
properties Global copying of prepositions and prepositional phrases has not led to a dominance of prepositional constructions at the cost of postpositional ones in Karaim. A significant change was, however, induced by the selective copying of foreign right-branching syntactic properties. This selective copying has resulted in a greater
280 Éva Ágnes Csató
frequency of particle-like postpositions, which is more in harmony with the adopted SOV-order. Some other changes can also be observed. Certain converbs are lexicalised as postpositions in Turkic languages, for instance, Turkish göre ‘according to, as to’; see example (16). (16) radyoya göre radio:dat according to (< see:conv) ‘according to the radio’
The corresponding Karaim converb may also function as an adposition taking a preposed nominal complement, as illustrated in example (17). (17) xalïma k’or˙ ˙ a condition:1sg.poss.dat according to ‘according to my condition’
However, a sentential complement, which is always finite in Karaim, must follow its head. Thus, a converbial adposition may function both as a ‘postposition’ in (17) and as a ‘preposition’ in (18). (18) a.
b.
c.
K’el’im baγïna come:R-nonpast.1sg depending on k’im anda bolur. who there become:R-nonpast.3sg ‘I will come depending on who will be there.’ Baγïna n’in’d’i t’il’in’d’˙an’ depending on which language:3poss.abl ‘Depending on from which language’ Baγïna n’et’˙ak et’m’˙ak t’iy˙at’ depending on how much bread cost:A-nonpast.dir m’en’ sonra satïn alïm. I later buy:R-nonpast.1sg ‘Depending on how much it costs, I will buy bread later.’
A new type of word order is illustrated by the following example (19) in which the genitive attribute is postposed, Consequently, the word order is N + Postp + G. (19) tayanïp kolcˇoγu usnu tal cˇïbuxnun lean on:ip.conv hand:dim.3poss on.post willow stick:gen ‘having leaned on the handle of the willow stick’
Syntactic properties of foreign correlative constructions have been selectively copied onto a corresponding Karaim postpositional construction based on the postpositon k’ib’ik ‘like’. As a result of this copying, a new type of syntactic construction has developed in Karaim. As the examples in (19) illustrate, the Karaim postposition k’ib’ik ‘like’ can function as a correlative particle that does not hinder the matrix
Syntactic code-copying in Karaim 281
verb in assigning the accusative to the nominal phrase öz’ bagcˇasïn [own garden:3poss.acc] ‘own garden’ and otnu [fire:acc]. (20) a.
b.
Dostum m’en’imk’i baγcˇamnï friend:poss1sg I:gen:KI garden:poss.1sg.acc öz’ baγcˇasïn k’ib’ik own garden:poss.3sg.acc as.postp saban s’ur’d’ ˙ u. ˙ plough:past.3sg ‘My friend ploughed my garden like his own.’ T’ok’t’ ˙ u˙ otnu k’ib’ik acˇuvun. pour:past fire:acc like anger:3poss.acc ‘He poured his anger like fire.’
The last example of syntactic change is induced by the selective copying of the Case assignment in Slavic essive constructions of the type mentioned above, see example (4). As a result of this type of copying, the postposition bïla ‘with’ has acquired a particular syntactic property typical of case suffixes. This new property involves the postposition being governed by the adverb yanaša ‘next to’ as in example (21) and illustrates a possible development in the grammaticalisation of postpositions as case endings. Compare this Karaim construction to the use of the Polish preposition przy ‘by, at’, which governs the instrumental case. (21) m’en’imb’6 yanaša I:gen.with next to ‘next to me’
6. Convergence and typology To conclude, two theoretical corollaries: the syntactic convergence between Lithuanian Karaim and the dominant languages of the linguistic area is a result of contact-induced processes of code-copying. There are several types of code-copying processes. Global copying of foreign material is a necessary precondition for selective copying of phonological, combinational, frequential and semantic properties. Selective copying may trigger frame-changing innovations. Another conclusion can be drawn concerning word-order typology. Identifying categories across languages is a necessary prerequisite for typological undertakings. Categories labelled as postpositions in grammatical descriptions, even in closely related languages, may, however, exhibit different syntactic properties. A careful analysis of the relevant differences may lead to a more sophisticated typology, in which also gradual changes, such as those which characterise Karaim, may be observed and set in relation to other typological features of the compared languages.
282 Éva Ágnes Csató
Notes 1. The field research in Lithuania, the findings of which have been evaluated in this paper, has been carried out within the framework of a project financed by the Deutsche Forschungsgemeinschaft. I thank Prof. Dr. Hans-Jürgen Sasse of the University of Cologne for his friendly support and willingness to supervise the project. 2. Pronouns are more conservative. Postpositions such as gibi, for instance, still govern the genitive with pronouns, thus senin gibi [you:gen like] ‘like you’. The same phenomenon can also be observed in Karaim. The pronoun is still marked by the genitive although the possessive suffix does not agree with it, for example m’en’im üs’n’u? [I:gen on] ‘on me’. This is also true for the postposition bïla, which is used almost exclusively as cliticised on the stem. However, it takes the genitive with pronominal stems. 3. “Beachtlich ist jedenfalls der Parallelismus zwischen Aufhebung der ursprünglichen Determinationsrichtung und einer formalen Neigung der ursprünglich syntaktisch übergeordneten Elemente, sozusagen sich dem Status des Suffixes anzugleichen. In suffigierter Form sind auch Postpositionen Elemente der spezifischen syntaktisch nach rechts orientierten, phonologisch vereinheitlichenden oikonomia des Wortes, die jedenfalls eine klare Tendenz zur sukzessiven (inhaltlichen) Determination aufweist. Es läßt sich mit Recht sagen, daß die Postpositionen als Gesamtheit eine formal labile Gruppe darstellen, und es könnte die Vermutung gewagt werden, daß diese Labilität damit im Zusammenhang steht, daß sie sozusagen als Pufferelemente genau an der Schnittlinie zwischen zwei — einander entgegengesetzt orientierten — syntaktischen Kraftfeldern (dem des Satzes und dem des Wortes) ihren Dienst leisten und daß die strukturellen Ordnungen dieser beiden Felder jeweils mit einer determinativen Ordnung einigermaßen korreliert ist” (Johanson 1974: 103–104).
References Csató, Éva Ágnes. 1996. Some typological properties of North-Western Karaim in areal perspectives. In: Boretzky, Norbert, Enninger, Werner & Stolz, Thomas (eds.), Areale, Kontakte, Dialekte. Sprache und ihre Dynamik in mehrsprachigen Situationen. [= Bochum-Essener Beiträge zur Sprachwandelforschung 24], 68–83. Bochum: Brockmeyer. Csató, Éva Ágnes. 2000. Some typological features of the viewpoint aspect and tense system in spoken North-Western Karaim. In: Dahl, Östen (ed.), Tense and aspect in the languages of Europe, 723–751. Berlin: Mouton de Gruyter. Greenberg, Joseph H. 1966. Some universals of grammar with particular reference to the order of meaningful elements. Universals of language, 73–113. MIT Press. Greenberg, Joseph H. 1980. Circumfixes and typological change. In: LaBrum, Rebecca, Shepherd, Susan & Traugott, Elizabeth (eds.), Papers from the fourth international conference on historical linguistics, 233–41. Amsterdam: John Benjamins. Johanson. Lars. 1974. Sprachbau und Inhaltssyntax am Beispiel des Türkischen. Orientalia Suecana 22: 82–106. Johanson, Lars. 1993. Code-copying in immigrant Turkish. In: Extra, Guus & Verhoeven, Ludo (eds.), Immigrant languages in Europe. Clevedon, Philadelphia, Adelaide: Multilingual Matters Ltd, 197–221. Johanson, Lars. 1996. Kopierte Satzjunktoren im Türkischen. Sprachtypologie und Universalienforschung (Berlin) 49: 1–11.
Syntactic code-copying in Karaim 283
Johanson, Lars. 1998a. Code-copying in Irano-Turkic. Language Sciences 20: 3 [Special issue on areal typology], 325–337. Johanson, Lars. 1998b. Frame-changing code-copying in immigrant varieties. In: Extra, Guus & Verhoeven, Ludo (eds.). Bilingualism and migration. [=Studies on Language Acquisition 14], 247–260. Berlin & New York: Mouton de Gruyter. Johanson, Lars. 1999. The dynamics of code-copying in language encounters. In: Brendemoen, Bernt et al. (eds.), Language encounters across time and space, 37–62. Oslo: Novus. Rothstein, Robert A. 1993. Polish. In: Comrie, Bernard & Corbett, Greville G. (eds.), The Slavonic languages, 686–758. London, New York: Routledge.
AUTHOR "Neil G. Jacobs"
TITLE "Yiddish in the Baltic region"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Yiddish in the Baltic region* Neil G. Jacobs
1.
Introduction
Yiddish is a relative latecomer to most of the Circum-Baltic [CB] region. As a result, the amount and extent of influences of a CB Sprachbund upon Yiddish along the eastern rim of the Baltic Sea are not profound.1 Yet Yiddish belongs in a volume dealing with CB languages for two reasons. First, because it (Yiddish) is there, that is, indigenously spoken in the region.2 Second, limited as they may be, the linguistic influences — in both directions — are there. Examination of the contact features found in Baltic Yiddish can provide the linguist with non-trivial data concerning the nitty-gritty of language contact, internal vs. external causation of change, etc. The main goals of the present paper are as follows. First, I wish to show to what extent the term Baltic region is relevant to Yiddish linguistics. This is done by placing the discussion of Baltic Yiddish within the linguistic and geographical contexts of the Yiddish speech territory, and seeing in which ways Yiddish spoken in the Baltic region does or does not fit into the integrated topography of Yiddish linguistic geography.3 The flip side of this task, then, is likewise to see in which ways Yiddish in the Baltic region has been influenced through CB linguistic contact. The second goal of this paper must thus necessarily be the mere fact of presentation and analysis of a number of noteworthy features in Baltic Yiddish [BY].4 The features enumerated in the source works (Weinreich 1923; Kalmanovitsh 1926; Mark 1951; Lemkhen 1995) will serve as a basis for current discussion. In addition, I will draw upon my earlier discussion of Courland Yiddish [CourlYid] (Jacobs 1994a), and that of Verschik (ms.) for Yiddish in Estonia. It is hoped that the present paper can serve as a framework for subsequent detailed research into Yiddish in the Baltic region in the context of CB linguistic research.
2.
Ashkenaz: The Yiddish speech territory
The traditional Jewish term for the geographic area of indigenous Yiddish language and culture is Ashkenaz.5 The term Continental Ashkenaz, coined by Miller (1990),
286 Neil G. Jacobs
is used to distinguish the historical European home territory of indigenous Yiddish language and culture from the subsequent areas of settlement in the Ashkenazic diaspora, e.g., North and South America, South Africa, Palestine/Israel, Australia, etc. According to the generally-accepted scholarship, the origins of Ashkenazic Jewry6 trace back appoximately one thousand years, to the settlement of Jews from “areas now designated east-central France and northern Italy” (M. Weinreich 1980: 39) along the Rhine River, in an area referred to by Jews as Loter (cf. nonJewish terms Lothringen, Lorraine). The primary settlements were in the communities of (modern-day names in German) Speyer, Worms, Mainz, Köln, and Metz (M. Weinreich 1980: 1). Weinreich (1980: 328) refers to Loter as “the cradle of Ashkenaz”, as over the next several centuries Ashkenaz expanded through settlement to the north, south, and — primarily — east. At its geographic height, Continental Ashkenaz constituted the second largest language/culture area in Europe (after Russian), elaborated in Herzog et al. (1992:7) as stretching “from Alsace, at the Franco-German border in the Southwest, and Holland at the English Channel in the Northwest, eastward through Germany, southeastward through Austria, Czechoslovakia, Hungary, Rumania and Ukraine, northeastward into the Baltic States, Lithuania and Latvia, and eastward through Poland and Belarus into Muscovy.” More recently, the traditional Weinreichian view has been challenged. King (1979) and Katz (1987) suggest that Yiddish arose on eastern German territory. Wexler (1987, 1991) has offered a more radical challenge, suggesting that Yiddish arose even more to the east, on bilingual Slavic-Germanic territory, as speakers of Judeo-Slavic shifted to Judeo-Germanic (i.e., Yiddish).7 Weinreich (1980: 733) provides in table form a periodization of the linguistic and geographic history of Yiddish. The Early Yiddish period (up to ±1250) is characterized geographically by the confinement of Yiddish to the German speech territory. The Old Yiddish period (±1250–±1500) saw increasing migration of Yiddish speakers to Eastern Europe, and the consequent language contact with Slavic speech. Significant Ashkenazic settlement in the Baltic region begins after the Old Yiddish period.
3.
The major dialects of Yiddish
Traditionally, the primary division made in Yiddish dialectology is between Western Yiddish [WY] and Eastern Yiddish [EY].8 WY — once dominant, and the main source of the earlier written language — largely died out in the nineteenth century, in the aftermath of the Haskalah (Jewish Englightenment). When we speak today of the modern spoken and written Yiddish language, it is generally in reference to EY-based varieties. EY is further subdivided into Central Yiddish [CY], Northeastern Yiddish [NEY], and Southeastern Yiddish [SEY].9 CY and SEY
Yiddish in the Baltic region 287
together comprise Southern Yiddish [SY], as opposed to NEY (which includes the Courland subdialect; see below). However, NEY and SEY share the fact that each has incorporated Slavic influences to a significantly higher degree than has CY (U. Weinreich 1958).10 The major dialects of Yiddish are illustrated in the following map from the Language and Culture Atlas of Ashkenazic Jewry (Herzog et al. 1992).
Realization of: O4 k[ ]fn to buy X[ ]š meat E4 O4 E4 a a: a: a ~ aj oj a ~ aj oj aj oj ej ej ej ej ~ oj ej eu ej ej öj
WY CY SEY NEY Co
Marvin I. Herzog, Uriel Weinreich z”l, and Vera Baviskar z”l (1992). The Language and Culture Atlas of Ashkenazic Jewry. Volume 1: Historical and Theoretical Foundations. Max Niemeyer Verlag: Tübingen. Reproduced with permission.
Map 1.
Of course, dialect classifications ultimately are the constructs of linguists making informed — yet subjective — decisions as to which features are to be considered diagnostic. The total picture of all features across the speech territory yields more of a topography, rather than sharp dialect divisions. Generally, linguists draw their dialect boundaries at places of abrupt differences, where speech varieties differ from
288 Neil G. Jacobs
one another in a large number of features. Even here, however, with time, a new, contact-induced levelling between the radically different varieties usually occurs, giving rise to new transition dialects. We see such transition dialects in Northern Poland (Herzog 1965) and in areas along the WY–CY divide (U. Weinreich 1964b; Herzog et al. 1992). The emergence of a modern Standard Yiddish [StY] language further eroded the use of dialect in Eastern Ashkenaz from the 19th c. onward, as traditional features — cultural, as well as linguistic — faded in favor of the emerging StY language (see Mark 1951; U. Weinreich 1952; Herzog 1964).
4. Yiddish in the Baltic region 4.1 General and historical Yiddish-speaking settlement in CB territory is historically recent. We must define when, where, and from where Yiddish speakers settled in CB territory. Roughly speaking, we can divide the area — historically — into three main sub-areas: along the southern rim of the Baltic (for the most part covering present-day northernmost Germany and Poland, as well as, historically, Denmark), the northern Baltic area (Scandinavia — Sweden and Finland), and the eastern Baltic rim (the modern Baltic countries of Lithuania, Latvia, and Estonia), reflecting distinct situations in Yiddish/Ashkenazic history. To the extent that Yiddish was spoken historically along the southern rim of the Baltic, it can, for the most part, be seen as a geographical extension of the WY speech territory. Yiddish in Sweden is largely to be seen in two distinct phases: originally, as an outpost of WY dating back to the arrival of WY-speaking Jews from Germany in the late eighteenth century; significant migration of EY speakers to Sweden began only much later (after the pogroms of the 1880s and beyond). The tiny Ashkenazic presence in Finland stemmed largely from Russian Jews conscripted into the Czarist army. The present paper focuses on Yiddish along the eastern Baltic for two main reasons. First, unlike Yiddish in Germany and Scandinavia, EY along the eastern Baltic did not dwindle to (near) extinction post-Haskalah. Rather, Yiddish along the eastern Baltic remained the language of the Ashkenazic masses. Second, Yiddish along the eastern Baltic is useful in that we can assess it in terms of (a) how it fits into the general topography of the EY dialect continuum; (b) the nature and extent of CB influences; (c) the comparative “pull” from StY and CB influences. The traditional NEY speech territory covered Lithuania, Latvia, Estonia, Belarus (Jewish term Raysn), parts of northern Poland, and the northern Ukraine. In Yiddish, NEY is generally referred to as litvish(er)11 yidish ‘Lithuanian Yiddish’, and the term Lite generally covers the above-mentioned geopolitical/ethnographic areas.
Yiddish in the Baltic region 289
Thus, Jewish Lite – non-Jewish Lithuania, Litwa, Lietuva, Litauen, etc. (see Katsenelnbogen 1951), where the non-Jewish terms generally are limited to either geopolitical or ethnographic Lithuania. For Jews, the terms Lite and Litvak ‘Jew from Lite’ are defined Jewishly.12 For example, Katsenelnbogen (1951: 309–310) notes that Jews from Zamet ‘Samogitia’ referred to themselves (in conversation with non-Litvakes) as Litvakes, seldom as zameter Jews; similarly, the Jews from Raysn (non-Jewish Belarus) and Letgalye (non-Jewish Latgalia) are also Litvakes. The following historical background is taken from Lemkhen (1995: 17ff). Ashkenazic settlement in the Kingdom of Lithuania dates from the 14th c., when Grandduke Vytautas granted privileges to Jews. By the 16th c. there were 15 Jewish communities in the realm, but most of these were located in Belarus, with only two in ethnographic Lithuania (in the Aukštaitian area). In 1566 the Jews in Lithuania numbered approximately 8,000–10,000. There was no Jewish settlement in Zamet in this period. By the 17th century the number of Jews reached approximately 32,000; by 1766, there were 157,000 Jews in Lithuania, of which approximately 75,000 lived in ethnographic Lithuania or in the districts of Vilnius, Trakai, and in Samogitia. After the Union of Lublin (1569) there was a movement of Jews into villages in Lithuania. By the 18th c. one half of Lithuanian Jews lived in villages. Lemkhen (1995: 17) writes that the relocation of Jews from villages into towns by Czarist force did not significantly change the sociolinguistic situation for the Jews initially. Lemkhen (1995: 18) claims that the significant Lithuanian influence upon Yiddish began first in the 18th c., and that is when the Lithuanian loanwords began to make their way into Yiddish. The number of Lithuanian loans decreased in the interwar period of the 20th c. under the influence of Yiddish books and newspapers (Lemkhen 1995: 18). 4.2 Broad NEY: Some features This section will deal with NEY in the broadest sense, that is, a basic commondenominator version of the NEY dialect area, in which the focus is on features generally shared throughout the whole region, rather than on the isoglosses which subdivide the NEY area. Broad NEY is set off from other Yiddish varieties by a number of key isoglosses, as has been thoroughly discussed in the Yiddish linguistic literature (see Herzog 1965; Herzog et al. 1992). However, there are likewise several things more of a sociolinguistic or social nature which set Yiddish in Lite apart from other EY; these are best described in terms of general tendencies, rather than in terms of one or the other specific isogloss. Thus, Mark (1951: 439ff; 464) claims that, in general, Yiddish speakers in Lite tended to have a higher understanding of learned Semitic-component vocabulary than did Yiddish speakers elsewhere. Mark (1951: 464) also calls noteworthy the high frequency of usage of Semiticisms in NEY women’s speech. Furthermore, the NEY area is set off from other EY by a number
290 Neil G. Jacobs
of non-linguistic features in areas such as food preparation, religious custom, etc.; see Herzog (1965: 18–28). Nevertheless, one may still recognize a number of isoglosses which, collectively, paint a general picture of NEY. Some of these will be given in this section; see Mark (1951: 433ff) for detailed discussion of broad NEY features. However, two cautionary notes should be sounded. First, the features given do not necessarily cover each and every location within the NEY area; variation at the subdialect level does occur. Second, there was — from roughly the 19th c. onward — an increasing tendency toward consolidation in EY. In the NEY speech territory this has led to two trends: the tendency of local dialects to give way to broader NEY (Mark 1951: 441), and/or to the emerging modern standard language (U. Weinreich 1952). Below follow some features which mark broad NEY from other varieties of EY. 4.2.1Phonology Herzog (1965: 164) outlines four main features which characterize the historical development of the NEY vowel system: (i) loss of length; (ii) (regional) unrounding of øy > ej; (iii) (regional) raising of au > ou; (iv) (regional) half-fronting of ou > oj. Three of the four features (ii, iii, and iv) are qualified as “regional”, i.e., as occurring only in a subregion of the broader NEY dialect area; the remaining feature, (i) made only partial inroads (phonologically) into zameter yidish [ZY]/CourlYid (see below). Thus, the more Yiddish approaches the Baltic, the more we see the maintenance of the earlier state of affairs (retention of vowel length, of rounded øy, etc.). However, in spite of the clear differences between northern (i.e., closer to the Baltic) NEY and non-northern NEY, the two sub-areas do share much common development, and Yiddish linguistic scholarship is justified in linking these to a common earlier NEY stage. The historical changes in the vowel systems across the Yiddish map are Yiddish-internal developments in this area, reflecting Yiddish-internal topography, rather than external contact. Herzog et al. (1992: 41) write: “In large parts of NEY and SEY, [ji] is a free or contextual variant of word-initial /i/.” One also frequently finds in NEY (and SEY) that StY jid ‘Jew’, jing6r ‘younger’ are realized as NEY id, ing6r. In most NEY, the insertion of j as a hiatus breaker is a general process; cf. StY r6fu6 ‘remedy’, to6s ‘mistake’, NEY r6fuj6, toj6s (see Herzog 1965: 209–210; Herzog et al. 1992: 22). Note that glide j thus often reappears as a hiatus breaker in NEY: id ‘Jew’, but di + idn as [di-j-idn] ‘the Jews’. Much of NEY also generally inserts a glide between a or e and a following velar nasal; thus, StY bre]g6n ‘to bring’, la]g ‘long’, NEY brej]g6n, laj]g (Herzog et al. 1992: 22). In the consonantism the broader NEY speech territory shared the historical confusion of hushers and hissers (š–s, ž–z, tš–ts, etc.) — the so-called sabesdiker losn ‘Sabbath speech’ (based on the historical presence of hushers in the words realized in StY as šab6s ‘Sabbath’, lošn ‘language, speech, tongue’). This feature was under
Yiddish in the Baltic region 291
retreat throughout the broader NEY area (though to varying degrees; additionally, the manifestation and description of sabesdiker losn varied in different NEY subareas; see Mark 1951: 436–438), under the influence of emerging StY (U. Weinreich 1952). Weinreich (1964a) showed the independence of Yiddish sabesdiker losn from a similar phenomenon in coterritorial languages (e.g., Polish regional mazurzenie). Lithuanian and Latvian maintain a clear distinction between hushers and hissers; thus, the presence of the (retreating) feature of sabesdiker losn in BY is not to be attributed to CB influence.13 The voicing sandhi whereby a sequence of obstruents takes the voicing of the last obstruent in the sequence (e.g., fus ‘foot, leg’ + benkl ‘bench’ Æ fu[z]benkl ‘footstool’) is attributed to Slavic influence (U. Weinreich 1958: 7), and is found generally across the NEY region. Thus — to the degree that this feature is found in BY — we are not justified in claiming a specifically CB source, in spite of the fact that this feature is present in Lithuanian (and Latvian). NEY generally preserves obstruent voicing distinctions word-finally (Mark 1951: 434–435), e.g., vajz ‘(I) show’, vajs ‘white’. 4.2.2Morphology and syntax Perhaps most striking in NEY vis-à-vis other Yiddish varieties is the restructuring from an earlier three-gender system (masculine, feminine, neuter) to a system lacking the neuter gender. Early treatments attributed the rise of the “two-gender” system in NEY to contact14 with Lithuanian and/or Belarusian.15 The Baltic languages coterritorial with BY — Lithuanian and Latvian — have two-gender systems — masculine and feminine.16 The gender system of modern Lithuanian (Mathiassen 1996: 37ff) is heavily semantically-based; for example, biological sex overrrides morphology in gender assignment. A residue of an older neuter is found in certain impersonal constructions. Furthermore, modern Lithuanian contains classes of nouns which only occur in the singular (singularia tantum) or the plural (pluralia tantum). The class of singularia tantum includes abstract nouns like ‘love’, ‘conscience’; collectives like ‘youth’, ‘intelligentsia’; materials like ‘iron’, ‘vodka’, ‘honey’, and names like Lietuvà, Vìlnius. The class of pluralia tantum includes names of ceremonies and feasts, many mass nouns like ‘paint’, ‘flour’, ‘gas’, compass directions, and more. Mathiassen (1996: 51) finds noteworthy the high number of mass nouns in the pluralia tantum group, and writes: “The reasons for this development remain obscure.” Some parallels in the NEY gender system may possibly be relevant to discussion of the Lithuanian pluralia tantum group; see also Lemkhen (1995: 44–45). In three-gender StY, nouns are assigned (in the singular) a gender — masculine, feminine, or neuter. There is definite article/adjective agreement for all cases. All plural noun phrases show (definite article di and) adjective suffix -6.
292 Neil G. Jacobs
Standard Yiddish definite article and adjective suffix
nom acc dat
masc
fem
neut
pl
der -6r dem -n dem -n
di -6 di -6 der -6r
dos -6 dos -6 dem -n
di -6 di -6 di -6
Allomorphs of adjective suffix {-n} are -n, -6n, and -6m. Positional variants of the -n are [n], [m], and []]. The uniform (singular) indefinite article is a/an for all genders and cases. With one exception, the substitution of the definite article by the indefinite article leaves the adjective suffix unchanged; thus: der gut6r man ‘the good man’, a gut6r man ‘a good man’. The exception is with neuters; here, the adjective ending is zero in indefinite NPs: dos šejn6 lid ‘the pretty song’, a šejn lid ‘a pretty song’. The indefinite article in the plural is zero; thus: di šejn6 lid6r ‘the pretty songs’, šejn6 lid6r ‘pretty songs’. Additionally, StY allows contraction of the masculine and neuter definite article dem to -n after a preposition; thus, mit ‘with’, af ‘on’, far ‘for’, etc. + dem Æ mitn, farn, afn, etc. With the prepositions in ‘in’, and fun ‘of, from’ the contracted forms are in6m, fun6m (or in–ø, fun–ø). With the prepositions ba ‘by’ and tsu ‘to’, the contracted forms are bam, tsum. Note that it is only the post-prepositional dem which may contract; thus, in StY only masculine and neuter NPs may show contraction, a point which is relevant to our discussion of the NEY system. U. Weinreich (1961) showed that the new NEY gender system that arose was more than simply a “two-gender system.” The new system contained masculine, feminine, mass nouns (vs. regular count nouns), in addition to regular plurals. In addition, masculine and feminine are further subdivided semantically into ±male, ±female, and/or morphologically marked/unmarked. Nouns which lack both semantic maleness/femaleness and morphological marking are either intermediate masculine or intermediate feminine.17 The pairings of definite article + adjective suffix in NEY are like those in StY: der -6r, di -6, dem -n (with no dos -6 pair). However, two points should be noted. First, NEY generally shows a reduction of the older (sg) acc-dat > obl. Mark (1951: 454) calls this combined case the objective case. He claims that in the pronouns the dative forms have survived, while in the definite articles the older accusative forms have survived. However, the situation is less clear-cut across the Yiddish map, as seen from Wolf ’s (1969) treatment of the relationship of gender and case systems. Second, NEY allows a contracted form even with intermediate feminines (as long as there is no adjective in the NP), e.g., di gas ‘the street’, afn gas ‘on the street’ (but: af der brejt6r gas ‘on the broad street’). Thus, the contracted intermediate feminine NPs allow a door for the reinterpretation of the gender of contracted
Yiddish in the Baltic region 293
masculines and neuters as intermediate feminine nouns, and of intermediate feminine nouns as (intermediate) masculine nouns. Suggestions are made that relate the origin of the neuterless gender system of NEY to a borrowed rule of case/gender syncretism in the Polish loc.sg, surfacing in NEY as the contracted form of prep + def (e.g., NEY intermediate feminineafn gas, masculine afn tiš ‘on the table’); see U. Weinreich 1961; Herzog 1965: 132). In light of Wolf ’s (1969) findings, suggestions of a Baltic source for this NEY development seem even weaker, since the Baltic languages in question (Lithuanian and Latvian, as well as Estonian) are rich in cases and case marking; rather, contact with Polish/Slavic is the preferable suggestion. Many of the structural changes in the NEY gender system can be shown to be internally motivated (U. Weinreich 1961; Herzog 1965: 124; Wolf 1969; Jacobs 1990a). While some parallelism exists between new gender assignment in varieties of NEY and the Lithuanian equivalents (see Mark 1944: 90; U. Weinreich 1961, cited in Herzog 1965: 122), there is also a substantial corpus of examples where the newly-assigned gender in NEY π Lithuanian gender. However, the influences of language contact — Yiddish and Lithuanian — are not to be ruled out as contributory factors in the development of the NEY gender system (U. Weinreich 1958: 12), particularly in the treatment of mass nouns,18 which appear similar to a sub-class of the Lithuanian pluralia tantum-substances, e.g., ‘water’, ‘jam’, ‘bread’, etc. Characteristic for the mass nouns in NEY is a plural NP, but a third person singular for the verb. Thus, in di gelt ‘the money, the definite article remains di after a preposition: mit di gelt. However, the verb is in the singluar: di gelt ligt afn tiš ‘the money is lying on the table’. In Jacobs (1990a: 97) the NEY gender system is represented as a series of flowchart decisions, as follows: 1. Is the noun [+count]? If NO, then assign it mass gender: di vas6r ‘the water’, di brojt ‘the bread, etc. 2. If YES ([+count]), is the noun [+singular]? If NO, then assign it noun-plural status. 3. If YES (i.e., noun is [+singular]), is it marked for biological sex [+sexus]? If NO, then ask a follow-up question: Does the noun contain any formal marking (phonological/morphological)? If NO, then assign to the class of EMPTY nouns, which ultimately receive intermediate masculine or intermediate feminine gender; here, the gender assignment is ad hoc, and open to regional (within the NEY area) gender reinterpretation. If, however, the [−sexus] noun does contain formal marking, then use that as a non-ad hoc basis for gender assignment, e.g., nouns ending in -6r are assigned masculine gender, nouns ending in -6 are assigned feminine gender. 4. If, however, the noun is [+sexus], then gender assignment is based on the biological sex, overruling noun-endings; thus: zejd6 ‘grandfather’ is masculine, in spite of its ending in 6, and švest6r ‘sister’ is feminine, in spite of its ending in 6r.
294 Neil G. Jacobs
For further details of the NEY system, the reader is referred to Jacobs (1990a), and the works cited therein. Yiddish has a present-participle construction consisting of verb “infinitive” + suffix -dik, e.g., laxn ‘to laugh’ + dik Æ laxndik ‘laughing’, redn ‘to speak’ + dik Æ redndik ‘speaking, talking’.19 When the verb takes the reflexive pronoun zix,20 as in, e.g., vašn zix ‘to wash (oneself)’, we observe a partial distinction between NEY and general EY. In the construction X-n-dik, non-Lithuanian EY shows either vašndik zix or zix vašndik, while Lithuanian Yiddish allows vašnzixdik. In non-NEY the suffix dik is appended directly to the verb form, forming a present participle; the reflexive pronoun zix may then either precede or follow the present participle. In NEY, however, the zix may be suffixed directly as part of the verb, to which the suffix dik is then attached. Mark (1951: 457) attributes this NEY construction to influence from Lithuanian. In Lithuanian gerund constructions the reflexive particle is inserted between the (last) prefix and the verb root (Mathiassen 1996: 154); this may have served as the model for the NEY innovation of laxnzixdik, bodnzixdik ‘while bathing (oneself)’, etc.21 NEY permits use of the present participle form to express “while X-ing”, whether referring to the subject of the main sentence or the subject of the embedded sentence (Mark 1951: 457). The distinction relies here on position in the sentence; thus: NEY (1) gej6ndik hob ix im g6zen walking have I him seen ‘I saw him (while I was) walking’ (2) ix hob im g6zen gej6ndik I have him seen walking ‘I saw him (while he was) walking’
Non-NEY, on the other hand, uses the gej6ndik form to refer only to the subject of the main sentence (I). When the walking is performed by the subject of the embedded sentence [= object of the main sentence] (he/him), this is expressed either by use of the infinitive gejn, or by a circumlocution (Mark 1951: 457): Non-NEY: (3) gej6ndik hob ix im g6zen walking have I him seen ‘I saw him (while I was) walking’ (4) a.
b.
ix hob im g6zen gejn I have him seen go.inf ‘I saw him (while he was) walking’ ix hob g6zen, vi er gejt22 I have seen how he goes
Yiddish in the Baltic region 295
Further, Mark (1951: 457) notes that the Yiddish literary language may express relative time of action by using either the present participle form (arajnkum6ndik) or the past participle form (arajng6kum6n): (5) arajn- kum6n- dik in štub, hob ix im derzen into come ing into room have I him pfv-seen ‘(while) entering the room I caught sight of him’ (6) arajn- g6kum6n in štub, hob ix im derzen into come[past.part] into room have I him pfv-seen ‘having entered the room, I caught sight of him’
Mark claims that in NEY (and in the non-literary language outside the NEY area) only the present participle form is used, and no distinction in time of action is possible; thus, arajnkum6ndik is used to express both present and past action. Here, NEY does not show influence from Lithuanian, where relative time of action between main and subordinate clause is marked in the gerund (Mathiassen 1996: 154).23 Historically, Yiddish forms the past tense by use of an auxiliary — a form of either hobn ‘to have’ or zajn ‘to be’ — plus the past participle of the main verb. There is a strong tendency in NEY to use the verb hobn ‘to have’ as the universal auxiliary verb in past tense constructions (Mark 1951: 457): NEY ix hob g6gang6n ‘I went’, ix hob g6zung6n ‘I sang’; cf. StY ix bin g6gang6n, ix hob g6zung6n. No suggestions of areal influences are made here, even though Swedish has undergone a similar change to universal use of ‘have’ as the auxiliary verb in past tense constructions (cf. English, as well). 4.3 Subdividing NEY While there is reason to speak of a broad NEY dialect, there is likewise reason to subdivide NEY. Thus, for example, U. Weinreich (1969) describes the situation of Yiddish in Belarus, as distinct from other NEY. Wolf (1969) demarcates a northern area from other NEY. This northern area — mentioned by others in the Yiddish linguistic literature as well — is what we might call here Baltic Yiddish. The term is justified as a shorthand notation; however, it is not clear that the term is justified as a separate classification. Mark’s (1951: 432) discussion of NEY distinguishes among Samogitian Yiddish — zameter yidish [ZY], “plain” NEY — Stam-Litvish, and Suvalker Yiddish. Mark (1951) will serve as the point of departure for the current discussion. Using Mark (1951: 440), we may summarize the general picture as follows. The main subdivisdion in NEY is between ZY to the west and Stam-Litvish to the east. ZY is bounded by Germany (actually, East Prussia) on the west, Courland to the north, Stam-Litvish to the east, and Suvalker Yiddish to the south. Mark (1951: 441)
296 Neil G. Jacobs
claims that the lines of demarcation between ZY and its neighbors are clear to the north, west, and south. To the east, there is a gradual transition from ZY to StamLitvish. Courland Yiddish (CourlYid) to the north is seen as a branch of ZY. Suvalker Yiddish to the south is basically Belarusian Yiddish, and is linked historically to the influences of the Yiddish of Grodne/Grodno.24 In subdividing NEY, Mark (1951:439ff) states that Semiticisms are more evident in ZY than elsewhere in NEY (and, as noted earlier, generally higher in NEY than in other EY). Mark also claims that ZY contains more Lithuanian loan words than does other NEY (see also Lemkhen 1995: 12). Thus, before we get into discussion of concrete linguistic elements and features, it can be noted that the NEY dialect area seems to subdivide culturally into two main areas: Zamet, and “not-Zamet”.25 In distinguishing the two culture sub-areas, Zamet and not-Zamet, it should be cautioned not to look in the first hand to an external, non-Jewish explanation. As Mark (1951: 447) points out, there is no direct parallelism between the main dialect subdivisions of Yiddish with those of the coterritorial dialects of Lithuanian. The Yiddish name Zamet corresponds to the Lithuanian Žemaitija. However, the geographic territories covered by Jewish Zamet and Lithuanian Žemaitija are not identical (though for conflicting views on the relative sizes of the respective areas, see Mark 1951: 447, and Lemkhen 1995: 12). In terms of Yiddish linguistics, ZY is a conservative, “older” type of NEY, while Samogitian Lithuanian is an innovative, “younger” dialect of Lithuanian. Furthermore, for the most part there is a lack of areal correspondence of specific linguistic features in Samogitian Lithuanian and Zameter Yiddish. For example, in some varieties of Samogitian Lithuanian original diphthongs > long monophthongs;26 ZY, on the other hand, has richly preserved original diphthongs (Mark 1951: 447).27 Lemkhen (1995: 21) does provide some examples of Samogitian Lithuanian loans in ZY; however, once having entered Yiddish, the geographic or social diffusion of Lithuanian loans is no longer constrained by the dynamics of Lithuanian linguistic geography, but rather, may diffuse along Yiddish linguistic-geographic pathways. Ashkenazic geography in general, and Yiddish linguistic geography in particular, must be looked at on their own terms, and not as subsets of the coterritorial non-Jewish cultural or linguistic geography (see Mark 1951: 450; M. Weinreich 1980: 47). When dealing with notions of contact, we must distinguish between internal and external contact. Thus, note Mark’s (1951: 442–443) general impression that (as far as Yiddish-Yiddish contact is concerned) ZY speakers for many generations had contact only with other NEY — in this instance, stam-litvishspeakers; the latter, on the other hand, were in ongoing contact not only with ZY speakers, but also with speakers of Polish Yiddish and Belarusian Yiddish. As regards external contact, Mark (1951: 442) notes that Stam-Litvish contains many more Slavic-origin words than does ZY, especially from Polish; e.g.:
Furthermore, Mark (1951: 450) notes that there is much more Slavic influence found in the Yiddish of Belarus (presumably, coterritorial Slavic only) than in Lithuania (historically, Yiddish in contact with both Baltic and Slavic, largely, Polish). The topography of NEY is shaped both by internal and external contact. We must keep in mind that even with our subdivision of NEY, there was an ongoing dialect levelling taking place. In general, urbanization, along with other political, social, national, and economic factors in the 19th and 20th cc. led to the loss of many older dialect forms in EY (Herzog 1964). For NEY, Mark (1951: 449–450) writes of four historical factors which contributed to dialect levelling in NEY: (1) the two major road-building projects of the 19th c.; (2) the spread of rail communications; (3) export business to Prussia brought in Jews from the eastern part of Lithuania; (4) major urbanization from shtetlekh ‘market towns’ to cities (e.g., Kovne/Kaunas, Shavl/Šiauliai, Ponevezh/Paneve˙žys) within the two most recent generations (pre-World War II). Thus, the face of NEY was undergoing significant change from the end of the 19th c. into the early 20th c., with consolidation of local dialects toward a broader, more generalized NEY (along with the tendency toward emerging StY). Mark (1951: 441) notes that ZY was once spoken over a wider area, and, in the decades before World War II, broader NEY gained ground. Mark (1951) provides a number of examples of features which distinguish varieties of NEY. We will focus at this point mainly on the features which distinguish ZY from Stam-Litvish. Page reference to Mark is given in the form: [YM: 439]. 1. Compared with Stam-Litvish, ZY largely retains the distinction between long and short monophthongs [YM: 439]: ZY bi˜n ‘bee’ — bin ‘am’, zu˜n ‘son’ — zun ‘sun’. Distinctive vowel length is also found in Lithuanian and Latvian (and Estonian), as well as in Baltic German. This might suggest that the coterritorial languages played a role in the maintenance of long vowels in northernmost NEY; however, contact with the coterritorial languages does not account for the instability of long vowels in CourlYid (see §4.5, below). 2. Vowels are frequently longer in stressed, open syllable in ZY [YM: 439–440]: mam6 ‘mother’, lib6 ‘love’. 3. Reduced or cliticized form of 3sg.fut marker [YM: 439–440]: in ZY -at, in Stam-Litvish -et (cf. non-reduced StY vet, reduced form -(6)t. Mark gives the
298 Neil G. Jacobs
sentences: ZY az m’at lebn, ‘at’n zen, Stam-Litvish az m’et lebn, ‘et’n zen ‘if one shall live, one shall see.’ 4. Survival of r-coloring in second-degree dim suffix28 in ZY mejdal6 vs. StamLitvish mejd6l6 ‘little girl’, from base form NEY mejd (StY mojd) ‘maiden, lass’. Note also lowering of schwa before x in Stam-Litvish mejdlax vs. ZY mejdl6x. 5. Maintenance of vowel color in clitic form of du ‘you’ in ZY, versus reduction to schwa in Stam-Litvish: ZY štejstu ‘you stand’ vs. Stam-Litvish štejst6. 6. Maintenance of diphthongal back/round glide u in ZY vs. tendency toward universal glide j in Stam-Litvish; thus: ZY
Stam-Litvish
gloss
heux feugl (cf. ZY plural [fejgl] boux
hejx fejgl bujx
‘high’ ‘bird’ ‘belly’
The presence of a back/round glide in Yiddish dialects correlates positively with the presence of phonemic long monophthongs; see Jacobs (1994b: 137). Thus, the presence of the back/round glide in ZY, and its absence in Stam-Litvish, adds support to the analysis of ZY as a dialect possessing phonemic vowel length, and to the analysis of Stam-Litvish as lengthless.29 4.4 Lithuanian influences in NEY 4.4.1Mark (1951) Mark (1951: 465ff) devotes discussion to Lithuanian influences in NEY. He claims that there are almost one thousand Lithuanian loanwords in NEY, and provides a partial list.30 Mark emphasizes social factors in Lithuanian-Yiddish language contact: Yiddish speakers in villages in contact with Lithuanian-speaking farmers, along with some contact among shtetl (market town) Jews with Lithuanian farmers who came into the shtetlekh. The shtetlekh also had their own Lithuanian-speaking element. Mark (1951: 466) writes that Lithuanian loan words in Yiddish are found in several semantic spheres, and do not lend themselves to precise summary; however, he claims that most of all these come from dorf (village) life, and heavily from daily life. Some examples from Mark’s (1951: 466ff) partial list of Lithuanian loans found in NEY include:
Yiddish in the Baltic region 299
NEY
Lithuanian
Yiddish gloss
pipk6n6n bug6 bájd6n6n zix gájš6n6n ~ gájs6n6n
pypkiúoti bugù baidýtis (su)gaı˜šti, (su)gaišìnti
‘to smoke, giving off much smoke’ ‘terror, fear’ ‘to be frightened (animal term)’ ‘to waste time’
The amount of Lithuanian loan words varied with the amount of contact, and many loans were localized — i.e., limited to one specific sub-area, and not known throughout Jewish Lite (Mark 1951: 465–466). Mark (1951: 466) provides the following examples of Lithuanian loans which are socially or geographically limited in their usage in Yiddish: only older (in 1951) Jews from Kovne and Yanev understood matréng6 ‘a flood’; only a true Zameter understood bukst6 ‘frightened (a horse term)’; possibly only a Jew from Palonge/Palangà called the wind from the north sóm6n6s, the wind from the south sáks6n6s;31 only cattle-cunning Jews understood grém6l6 (farlirn) ‘to cease chewing the cud’ or krék6n6s ‘first milk after calving’ < Lithuanian krekenà. Other scholars before Mark have noted Lithuanianisms in NEY. Prilutski (1924: 239) provides a small note on Lithuanian loans in Yiddish, recorded from the teacher Khane Rozin from Shavl; his list includes: Shavler Yid
StY
Lithuanian
gloss
set6n6 tak6 pelk6 šlaj6 plik6 gandriš šmag6n6s
brukv6 štegl kaluž6 a grojs6r šlitn nak6t bušl moj6x
‘turnip’ ‘footpath’ ‘puddle’ ‘a big sled’ ‘naked, bare’ ‘stork’32 ‘brains, mind’
However, we can go further than merely noting the presence of Lithuanian loans in NEY. Mark (1951: 466) notes that Lithuanianisms in Yiddish: (a) assimilate phonetically to Yiddish; (b) follow the grammar of the (Yiddish) dialect; (c) are adapted/integrated in idiomatic expressions; and (d) are thus Yiddish words. Thus, of deeper interest to us are the nature and degree of integratedness of Lithuanian influences in NEY. 4.4.2Lemkhen 1995 Lemkhen (1995) takes us significantly further in the discussion of Lithuanian influences upon Yiddish.34 However, even in Lemkhen’s detailed study the bulk of the Lithuanian influences to be found are lexical; pages 57–120 are devoted to a list
300 Neil G. Jacobs
of Lithuanian loan words in LithYid. Lemkhen (1995: 18) dates the appearance of Lithuanian loans in Yiddish to the 18th c., based on consideration of these loans in Yiddish in Lithuania, as well as in the Yiddish of Courland and Estonia. Like Mark (1951), Lemkhen (1995: 42) emphasizes the relatively large number of Lithuanian loan words in Jewish Zamet (as well as in the areas around Shavl, Ponevezh, and Birzh-Posvel). Lemkhen (ibid.) also gives examples of Lithuanian dialect variants in NEY, e.g., ZY/CourlYid kleb6 < Samogitian Lith kle˙b~ y s, non-ZY gleb6 ‘armful’ < Aukštaitian Lith gle˙b~ y s. This example shows a parallelism in this particular word between Jewish and non-Jewish Samogitia. Lemkhen (1995: 42) states that the variation in the Lithuanian loans within Yiddish is attributed to Yiddish borrowing from different Lithuanian dialects; however, once in Yiddish, the loan may diffuse beyond the geographic bounds of its Lithuanian source dialect. 4.4.2.1 Phonology. Lemkhen’s (1995: 26ff) discussion of phonology provides interesting data, both for issues of language contact and for the phonology of Yiddish. As expected, Lithuanian loans in Yiddish show, for the most part, adaptation to Yiddish phonology. Page references for data will be given in the form [L,26]. Liquids r, l. Lithuanian lingual r is realized with uvular r in Lithuanian loanwords in LithYid [L,26], except in CourlYid, which has a lingual r [L,34]. Similarly, NEY generally realizes Lithuanian non-palatal l as hard; CourlYid — presumably from contact with Baltic German — has a soft, alveolar l; thus, Lithuanian loans in CourlYid have a soft l. Lemkhen (1995: 33) notes that, in general, both Lithuanian l and Yiddish l are realized as progressively harder in the geographic sweep from west > east. Lemkhen (1995: 34–36) discusses the problem of husher-hisser confusion (sabesdiker losn; see discussion, above). He claims that for Yiddish in Zamet and the adjoining area (= bay-zamet), as well as in northeastern Lithuania, there was, traditionally, no set of hushers distinct from hissers; rather, there was an unclear inbetween sound. The confusion is not limited to older components in Yiddish (e.g., Semitic, Germanic, Slavic), but rather, is evident in Lithuanian loans as well. Thus, Lemkhen (1995: 35) states that he, as a native speaker, cannot say whether there is š or s in, e.g., NEY špejg6 ~ spejg6 ‘a big frost’ < Lithuanian speigas. Note also confusion in examples gandriš, šmag6n6s, mentioned above. Lemkhen (1995:29ff) discusses issues concerning vocalism. Lith au is integrated in various ways into Yiddish [L: 30], as ou or ei; thus, speakers “choose” identification with one or the other historical Proto-EY vowel. In CourlYid, Lith au remains au; thus, Lith ka˜upas ‘supremacy, preponderance, excess’ > LithYid koup ~ kejp ~ (CourlYid) kaup. The general NEY insertion of glide j between a, e and a following velar nasal is found in Lithuanian loans as well: LithYid trej]k6n6n (zix) ‘to shake’ < Lithuanian treñktis, trenkýti; cf. Gmc.-origin StY [la]g], NEY [laj]g] ‘long’.
Yiddish in the Baltic region 301
Lithuanian loans generally show reduction of the vowel to schwa in post-tonic position (L,29): Lithuanian vãnagas, rãgana, pãšaras > LithYid ván6g6 ‘sparrow hawk’, rág6n6 ‘witch’, (ZY) páš6r6 ‘fodder’. Only in a small number of words is the vowel color retained in post-tonic position (L,29): e.g., LithYid déndarz6s < Lithuanian di˜endaržis ‘enclosure for animals (in a field)’. This type represents a less integrated, newer loan. Lemkhen (1995: 29) also notes as an example of a newer loan — due to the non-reduction of a post-tonic vowel — Lith viršininkas > LithYid vírš6ni(n)k ‘boss’. However, it may be the case here that non-reduction of the posttonic vowel in the final syllable is related to higher-level phonological or morphological structure in Yiddish — that is, it may be seen as part of a suffix -ink or -(n)ik. Consideration of pretonic vowels in the Lithuanian loans also provides insight into foot structure in Yiddish. Commonly, pre-tonic e in Lithuanian loans becomes a in LithYid; thus: Lithuanian pelÍeda, kepùre˙ > LithYid paléd6 ‘owl’, kapúr6 ‘a nonfitting hat’ [L,30]. However, when the word contains two pre-tonic vowels, we find reduction of the second vowel to schwa for reasons related to Yiddish rhythm and foot structure; thus: Lithuanian šaltanõšiai, šimtakõjis > LithYid šalt6nós6s ‘blintzes, dumplings filled with cottage cheese’, šimt6kój6 ‘centipede.’ 4.4.2.2 Morphology. Lemkhen (1995: 44) deals with the issue of NEY gender, and opts for internal causation — rather than appeal to external influence — as the main impetus for the NEY development. Still, Lemkhen (1995: 44) is aware of the possibility of external, i.e., contact-induced change, and cites examples of change from historical neuter > feminine gender in Polish dialects of western Lithuania (around Wilkomir and Kejdan/Ke˙dainiai); in Polish spoken in the Vilnius area, the change of neuter > feminine gender was accompanied by the change of word-final -o# > -a# (generally these are markers of neuter and feminine, respectively, in Standard Polish), e.g., ucho, serco > ucha ‘ear (fem)’, serca ‘heart (fem)’. Likewise, Lemkhen (1995: 44) notes that if a Lithuanian loan ends in schwa in Yiddish, its gender is feminine in Yiddish, even if it was originally a masculine in Lithuanian.35 LithYid shows a rather free variation in how it deals with the Lithuanian noun suffixes -is, -es, -e. For example, LithYid can have an s-suffix where Lithuanian has -e: LithYid bjojríbis ‘rotten scoundrel’ < Lithuanian bjaurýbe˙ [L: 37]. Regionally, however, an s-suffix has established itself as productive in ZY, spreading to new Yiddish words, e.g. krextsis ‘a person who groans/moans a lot’ < krextsn ‘to groan/ moan’. See also kamatsis ‘cheapskate’ [L: 37], cf. StY kamtsn; here, there is direct replacement of one agentive suffix (-n, inherited from the Semitic component in Yiddish) with another (-is). Lith -ys is similarly used as an agent marker (Frank Silbajoris, p.c.). The use of -is as a type of noun marker is also found in Swedish; e.g., tjockis ‘fatso’, kändis ‘celebrity’ < adjectives tjock ‘thick, fat’, känd ‘known’.36 As concerns Lithuanian adjectival suffixes -is, -us, -as, however, LithYid shows a different development. Here, LithYid generally dumps the Lithuanian suffix
302 Neil G. Jacobs
entirely, and usually adds the native Yiddish inflectional suffixes (6, -6r, -n) as needed [L: 38]: Lithuanian plìkas > LithYid plik- ‘bare’. However, when the original Lithuanian adjective contained a (non-final) suffix -in-, LithYid incorporates it: Lith šešk-in-is ‘from a skunk’ > LithYid šesk6n6r.37 Lithuanian loan verbs frequently are placed in the Yiddish class of verbs containing a thematic vowel, and show -6n6n in the infinitive [L: 38].38 4.4.2.3 Semantics. Lemkhen (1995: 47ff) provides a list of Lithuanian loans in LithYid, organized by semantic area, e.g., nature, crafts, tools, agriculture, etc. Two examples will be noted here in the way of illustration. Lemkhen (1995: 56) notes that, often, a Lithuanian loan in Yiddish can take on a pejorative sense, especially when the Lithuanian loan coexists in Yiddish with an equivalent from the Slavic or Semitic component. He additionally notes semantic narrowing, as in Lith baidýti ‘to scare’ > LithYid bajd6n6n ‘to scare (animals).’ Lemkhen (1995: 56) also provides an example of a widespread Slavic-origin word in Yiddish taking on the meaning of its Lithuanian cognate. Generally, StY prost ‘simple’ derives from Slavic, and retains a meaning similar to the Slavic meaning. In Lithuanian prãstas has the meaning ‘bad’. Lemkhen gives the examples of NEY prost6 šix ‘bad shoes’, prost6 ib6rzetsung ‘bad translation’, etc., analogous to the Lithuanian meaning. 4.5 Courland Yiddish This section provides a brief discussion of the Yiddish of Courland, Western Latvia; see M. Weinreich 1923;39 Kalmanovitsh 1926; Jacobs 1994a. Linguistically, CourlYid is an offshoot of ZY. Jewish settlement in Courland is dated to the seventeenth (Kalmanovitsh 1926) or sixteenth century (Ariste 1937). The closing of Courland to new Jewish settlement in 1829 is important in that it means that Lithuanianisms in CourlYid may generally be assumed to predate that year (Lemkhen 1995: 19). Further, Kalmanovitsh (1926) emphasizes the importance of the presence of a significant Slavic component in CourlYid, tracing back to a population of Yiddish speakers in earlier contact with Slavic (Belarusian and Polish). Thus, CourlYid was “born” elsewhere, outside of Courland. There remains much work to do on the perceptual geography of Courland Jews, who referred to any Jew not from Courland as a Zameter (M. Weinreich 1923: 195). Lazerzon (1942) states that inland Jews from the Latgalia province called the Baltic-coastal Jews švejdn ‘Swedes’ or švejdl6x ‘little Swedes’, perhaps owing to the Swedish presence in the area from 1660. Lazerzon (1942) also provides interesting discussion of Yiddish toponyms in Latvia, which show distinct geographic patternings, cultural orbits, and orientations. The basic language setting in which CourlYid existed primarily consisted of contact with Baltic German (the language of official administration, commerce,
Yiddish in the Baltic region 303
urban life) and with Latvian. Russian did not arrive on the scene until much later. The German presence in the Baltic is divided into two main periods: a period of Low German settlement (Teutonic knights) from the thirteenth century, and a High German settlement dating from approximately the sixteenth century. High German eventually supplanted Low German by approximately the end of the eighteenth century, with survival of some Low Germanisms. Courland Jews were often seen as “Germanized” in speech40 and custom (see Rivkin 1951). The CourlYid word lists provided in Weinreich (1923) and Kalmanovitsh (1926) contain many (Baltic) German loans; e.g., CourlYid u(˜)r ‘clock, watch’, farpasn ‘to miss’, taug6ništ ‘good-for-nothing’ < German Uhr, verpassen, Taugenichts (cf. StY zejg6r, farzam6n gorništ). However, the German influences are for the most part lexical. In Jacobs (1994a), I argue that the changes of structural import occurring in CourlYid of the early twentieth century were mostly due to the “pull” of the general ongoing consolidation of EY speech during the modern period, toward broader NEY, or toward the emerging StY language.41 Thus, German loans show adaptation to Yiddish structure; e.g., German prefix ver- > CourlYid far-; German africate pf is adapted to Yiddish phonotactics: f in word-initial position, p non-initially: German Pfingsten ‘Pentecost’, Schnupfen ‘sniffles’, stumpf- ‘dull’ > CourlYid fingstn, šnupn, štump-. Like ZY, CourlYid represents a relatively conservative subdialect of NEY. On the one hand, CourlYid, like ZY generally, has preserved a number of features otherwise lost in NEY, notably, distinctive vowel length, the preservation of a frontround diphthong øy, and the non-merger of øy and ej (CourlYid [æj]), which are merged in broader NEY ej. While distinctive vowel length in CourlYid is still evident in the archival tape recordings of the Language and Culture Atlas of Ashkenazic Jewry (see Herzog et al. 1992), it is clearly a system in collapse, with occurrences of free variation (Jacobs 1994a: 95). This collapse is not attributable to influence from German or other CB languages, which maintain long vowels, but rather, to influences from broader NEY and/or the emerging standard language (Jacobs 1994a: 95–96). On the other hand, like broader NEY generally, CourlYid shows loss of the historical neuter gender, and the rise of intermediate-gender constructions (e.g., di land–afn land). Further, CourlYid shares with the rest of NEY the historical sabesdiker losn. However, here M. Weinreich (1923: 202) notes an interesting development: sabesdiker losn remains in CourlYid only in words of non-German origin (i.e., Slavic and Semitic); thus:
304 Neil G. Jacobs
CourlYid
StY
gloss
cep6blondz6n šuk6 šæjxl
ˇcep6blondž6n suk6 sejxl
‘badger, touch’ ‘to stray’ ‘booth for Sukkoth’ ‘logic; sense’
(< Slavic) (< Slavic) (< Semitic) (< Semitic)
In words of German origin the original hushing-hissing state of affairs has generally been restored, based — it would appear — on the influence of (Baltic) German; thus, CourlYid vas6r ‘water’, kusn ‘kiss’, šrajbn ‘write’, StG Wasser, küssen, schreiben; cf. StY vas6r, kušn, šrajbn, though the CourlYid doublet: kirzn6r ~ kiršn6r ‘furrier’ < German Kirschner provides an interesting counter-example (see Jacobs’ 1994a: 95 discussion of Weinreich 1923: 203). There is reason to look at CourlYid in a broader context of language contact in the Baltic region. For example, in Kiparsky’s (1936) discussion of Yiddish influences in Baltic German (see also Ariste’s 1937 review of Kiparsky), we note that Baltic German blondzen shows the results of NEY sabesdiker losn (cf. StY blondž6n ‘to stray, ramble (having lost one’s way)’ < Polish błondzic´). Furthermore, we find some loan words common to an array of languages in the region. Thus, Weinreich (1923: 235) claims that CourlYid raut ‘window’ traces back to a Middle High German form *rûte (the etymon is found in Courland German as well), but says that the source into Middle High German is unknown. Jacobs (1994a: 93) proposes Swedish ruta ‘quadrangle, window pane’ as the source for both Courland German and CourlYid. The claim of a Swedish source (rather than a Low German source, cf. Dutch ruit) is supported by the presence of this loan in Estonian ruut, which has a stem-final u in non-nominative forms, suggesting that the word was borrowed as an oblique form of an older Swedish weak noun (nom -a, acc/dat/gen -u), something typical of such Swedish loans in Finnish as well. M. Weinreich (1923: 224) notes that the Latvian influence on CourlYid is strikingly small, especially as compared with that of Lithuanian in the neighboring Kovner province. Weinreich suggests that this may owe to the presence of German in Courland. Lemkhen (1995: 22) claims to find only three Latvian loans in CourlYid which do not also have Lithuanian equivalents: priad6 ‘pine tree’, bruklin6s ‘mountain cranberry’, and kaudz6 ‘pile’. On the other hand, contact with Latvians should not be minimized; it existed, exerted itself economically and socially, and influenced Jewish toponymics (Lazerzon 1942) as well as the structure of Jewish households (in the eighteenth century; see Plakans and Halpern 1981). In summary, it may be said that CourlYid of the early twentieth century exhibits a “pull” from three directions: first, from the broader EY speech community; second, from Baltic German; third, though probably weakest, from general Baltic areal influences.
Yiddish in the Baltic region 305
4.6 Yiddish in Estonia Historically, Yiddish in Estonia is seen as part of CourlYid, which in turn is seen as a sub-branch of ZY. Lemkhen (1995: 23) cites Ariste’s (1970: 250–252) list of Lithuanianisms in Estonian Yiddish, further emphasizing the connection of Estonian Yiddish with northermost NEY. However, Lemkhen (1995: 23) notes that these Lithuanian loans were used primarily by the older generation of Estonian Jews. Verschik (1999) gives an overview of Yiddish in Estonia, and provides discussion of the current sociolinguistic situation for Yiddish speakers. Verschik (1999) outlines the historical, social, and educational background of Estonian Jewry. Historically, the Jewish migration to Estonia came from Courland and Lithuania. The Jewish community has been, and is, quite small; it was first organized in 1829; by 1918 it numbered 1523 Jews; 1929 Jews in 1922; 4389 Jews in 1934 (Verschik, 1999). Verschik emphasizes the multilingualism of Estonian Jews: “It is quite evident that Yiddish monolingualism has always been and is impossible in Estonia due to the small size of the Jewish community: multilingualism is an essential characteristic feature of Estonian Jews.” Verschik provides discussion and transcripts of interviews with informants. One possible contact feature which I noticed had to do with word order of subject and finite verb. Generally in Yiddish the subject appears before the finite verb, unless another sentence unit (e.g., direct object, indirect object, adverbial) is fronted, in which case the finite verb moves back to second position, and the subject now follows; thus, StY ikh leyen dos bukh ‘I read the book’, but dos bukh leyen ikh ‘It’s the book that I’m reading’ (and not, for example, the newspaper). In Verschik (ms), we find subject before finite verb, even with fronting of adverbial (transcription as found in Verschik): reidn reidn zei nit, ober in idiš me [Subject] ken [finite verb] zei nit farkeifn ‘as for speaking, they don’t speak, but one cannot sell them away in Yiddish’. Verschik also provides detailed discussion of the occurrence of loan vocabulary from Russian, German, and Estonian. Finally, Verschik mentions as a topic for future investigation a style of mixed speech — Russian, Estonian, and Yiddish — used by some Jews in Estonia.
5.
Conclusions
In this paper I have attempted to place the discussion of Yiddish in a Baltic context. The Baltic contact features in Yiddish were primarily lexical. In the Thomason & Kaufman (1988) framework, the overwhelmingly lexical nature of the contactinduced influences (Lithuanian Æ Yiddish) suggests a more superficial type of influence. The Slavic impact upon Yiddish has been much deeper, encompassing all levels of the grammar, as shown by U. Weinreich (1958). Yet the Baltic data in
306 Neil G. Jacobs
Yiddish can provide us with valuable data on phonological and morphological nativization processes, as well as shed light on the general synchronic description of Yiddish, for example, in the description of syllable and foot structure in Yiddish. We have seen repeated examples of the important role played by the macrosociolinguistic orientation of Yiddish-speaking Baltic Jews. While there are significant CB influences observable, the pull is strongest from the consolidation processes in EY speech and writing. At the same time, the CB presence has had its impact — beyond the merely lexical — on Yiddish in the Baltic region; this needs to be taken into account in linguistic descriptions of varieties of Yiddish in the region. This present paper is, of course, only one step in the investigation of Yiddish in its Baltic context. Some questions are raised. Why are there so few Latvian loans in CourlYid? The answer does not seem to rely solely on the number of years Jews lived in Courland, for Lithuanian loans were taken up after comparable time in Lithuania. M. Weinreich (1923) noted the presence of a dominant Baltic German; this certainly played a considerable role in the speech both of Jews and of Latvians. However, the levels of historical Jewish-Latvian contact would perhaps lead us to expect a larger Latvian-origin corpus in CourlYid. It would perhaps be most useful to look more deeply at the interplay of multiple and shifting “pulls” — linguistic orientations and orbits, over time and geography.
Notes * My original research on Yiddish in Courland, conducted at the YIVO Institute for Jewish Research, and at the archives of the Language and Culture Atlas of Ashkenazic Jewry, was funded by the Melton Center for Jewish Studies at The Ohio State University. I thank Prof. Mikhl Herzog, director of the LCAAJ for access to the Courland (and other) tapes, as well as for personal discussion about Courland Yiddish. I also thank the following individuals for discussion on relevant issues: Mr. Zachary Baker and Ms. Dina Abramowicz of the YIVO, Prof. Mark Louden, and Prof. Ilse Lehiste. For research support for the present paper, I thank the Center for Slavic and East European Studies at The Ohio State University. I also thank Ms. Anna Verschik for access to her unpublished paper. Any mistakes or shortcomings in the present article are the responsibility of the author. 1. The number of years of contact is not always the determining factor in the degree of influence in language contact; see U. Weinreich (1958: 40–41) on the differing degree of Slavic influence in eastern Europe upon Yiddish and Colonial German. 2. The ethnographic present is used, reflecting developments before the Nazi-German genocide in World War II. 3. On earlier use of the term topography in another sense, see by Chambers (1994). Topography is used here to refer to the gradualness, or Staffelung in variation across the Yiddish dialect map; see Jacobs (1993). 4. The term Baltic Yiddish is not, to my knowledge, used in Yiddish linguistic scholarship. Rather, varieties of Yiddish spoken around the Baltic rim are generally seen in the context of their
Yiddish in the Baltic region 307
respective dialects — generally, as subsets of Northeastern Yiddish in present-day Lithuania, Latvia, and Estonia, and, historically, of Western Yiddish in Scandinavia. In the present paper the term Baltic Yiddish is used both as a shorthand for easy reference, and as a (straw?) construct for examining to what extent Yiddish along the eastern Baltic rim differs significantly from its sister subdialects within the given dialect region. Use of the term Baltic Yiddish is not meant to imply sufficient cohesiveness or distinctness that would force a special classification within the field of Yiddish dialectology. 5. As a toponym, the term was used in Medieval Hebrew to refer to German lands. Later, the term came to refer to the home territory of Ashkenazic Jewry. The first known use of the name Ashkenaz was as an anthroponym, referring to a son of Gomer (Genesis 10: 3). 6. The term Ashkenazic Jewry refers to that Yiddish-speaking Jewry which arose in Central and Eastern Europe, and their progeny. The Ashkenazic branch of the Jewish people is thus distinguished from other branches, which developed their own civilizations, specific Jewish languages, etc., as for example Sephardic Jewry, which arose on the Iberian Peninsula and created a distinct Judeo-Spanish language — Judezmo. 7. Both Katz and Wexler view the Jewish settlements in Loter as giving rise to a separate Jewish Germanic-based idiom, which Wexler (1981) terms Ashkenazic German. 8. For detailed discussion of Yiddish dialect classification, the reader is referred to Herzog (1965), Katz (1983), Herzog et al. (1992). 9. On the problematic nature of the Stammbaum classification of SEY, see M. Weinreich (1980), U. Weinreich (1969). 10. Though the specific Slavic-origin features need not be identical in NEY and SEY. 11. Language names are generally neuter in non-NEY varieties of Yiddish. In NEY, which lost the historical neuter gender, language names are generally masculine. In the present paper the standard romanization system employed by the YIVO Institute for Jewish Research will be used when referring to Yiddish names (e.g., personal names, toponyms), general concepts (e.g., sabesdiker losn), where phonetic detail is not at issue. Linguistic examples will be given in phonetic symbols. However, automatic processes such as the assimilation of /n/ to place of articulation of an adjacent consonant will not be indicated unless germane to the point being made. Likewise, the symbol 6 represents a set of variants across the NEY area. Unless otherwise indicated, stress in Yiddish words is on the penultimate syllable. 12. In Yiddish, the term Litviner refers to a non-Jewish Lithuanian. 13. The suggestion could be raised that the very recession in NEY of sabesdiker losn, and restoration of the husher-hisser distinction, might have been encouraged by the distinctions made in coterritorial Lithuanian and Latvian. However, this is not borne out by the data. Lithuanian loanwords in NEY often show the hushing-hissing confusion. Of more importance in the recession of sabesdiker losn in NEY is the influence of StY (U. Weinreich 1952). In CourlYid, however, contact with Baltic German did play a role as well (see M. Weinreich 1923: 202, and discussion, below). 14. Shapiro (1939, cited in Herzog 1965: 156, Note 4) suggested a Lithuanian substrate Jewish language from which Lithuanian Jews later shifted to Yiddish; this suggestion is refuted, among other things, by the findings of Yiddish historical phonology. See also Herzog (ibid.), citing Sapir (1915), Mark (1944, 1951: 446). 15. While Belarusian is a three-gender language (masculine, feminine, neuter), there has occurred some realignment of gender based on the phonetic similarity of fem -a nouns and neut -a (< -o) nouns; see also Jacobs 1990a.
308 Neil G. Jacobs
16. Yiddish came into direct contact with Estonian much more recently. Historical simplification of the NEY gender system likewise cannot be attributed to Estonian influence. 17. U. Weinreich (1961) proposed intermediate feminine category. The parallel category intermediate masculine was proposed in Jacobs (1990a). 18. The possibility of Yiddish influence upon Lithuanian as concerns the mass nouns should also be investigated. All too often, linguistic treatments of language contact have ignored the possibility of areal influences from Yiddish, while recognizing other languages in the “areal soup”; see Jacobs 1996:211–212 on contact features in Vienna Urban German. On Yiddish influences upon Colonial German, see U. Weinreich (1958: 15) on word order, and (ibid.: 16–17) on vocabulary. 19. The exact analysis of this form is more complicated. One could possibly analyze it as 1st/3rd p. plural + suffix dik, alternatively as present stem of verb + suffix (6)ndik, or as infinitive + dik. Any analysis chosen leaves problematic residue. See Mark 1978: 335. 20. Use of the vowel i reflects the Yiddish spelling; the vowel in the reflexive pronoun is usually realized as e, a, or schwa. 21. In Lithuanian, the reflexive element is placed at the end of the participle if the verb stem does not contain any prefixes and the whole participle is in the nominative case. Generally, the reflexive element in Lithuanian is placed between prefix and verb root (Maria Koptjevskaja-Tamm, p.c.). 22. NEY speakers use this circumlocution as well. 23. The source of the distinction possible in the literary language is thus a separate matter. 24. Mark (1951: 448) writes that Grodne was the seat of the Jewish Va’ad d’Lita ‘Council of Lithuania’, and had the Suvalker kehiles ‘Jewish communities’ under its jurisdiction; as well, there were strong business ties, similar networks of trade-fairs, rabbi selection, arranged marriages, etc. 25. On a different slicing of the perceptual geographic pie by Courland Jews, see below. 26. However, it has been pointed out to me by Maria Koptjevskaja-Tamm (p.c.) that the situation is somewhat more complex: “some Samogitian dialects have changed original diphthongs not to monophthongs, but to other diphthongs (e.g. /uo/ became /ou/ in North Samogitian, cf. Standard Lithuanian duona ‘bread’ vs. North Samogitian douna).” Koptjevskaja-Tamm also provides an example of an innovative diphthong in Samogitian Lithuanian kuoja ‘leg, foot’, vs. Standard Lithuanian koja. 27. Mark (1951: 447) does, however, provide some examples of limited parallelism in certain smaller phonetic details. Thus, he claims that in the Ezherener region the hard l sound is found in both the local Lithuanian and the local Yiddish. However, he points out the possibility that here both languages were influenced by contact with Russian. 28. The second-degree diminutive suffix represented in StY by -6l6 is reconstructed for ProtoYiddish as *-6rl6. The /r/ is evident in varieties of Austro-Bavarian German. On analogous r-coloring in some varieties of CY, see Jacobs 1990b: 82. The term r-coloring here refers to the realization of the 6 preceding the r as some sort of full vowel (a or o, depending on which variety of Yiddish). 29. However, the unrounding of the glide must be seen as fairly recent in many locations, as recently as the early part of the twentieth century; see Herzog 1965: 11. 30. U. Weinreich (1958:37) speaks of “relatively few recorded elements of Lithuanian… or Lettish origin…” citing Mark (1951) for Lithuanian, and M. Weinreich (1923) and Kalmanovitsh (1926) for Latvian loans. U. Weinreich is concerned here with the relatively higher impact of Slavic over other coterritorial influences in Eastern Europe.
Yiddish in the Baltic region 309
31. Maria Koptjevskaja-Tamm (p.c.) suggests that the Lithuanian words corresponding to som6n6s and saks6n6s seem to be restricted to the area of Palangà. 32. Lemkhen (1995: 71) writes that the Lithuanian loan gond(6)r6s~gandris~gandriš is a possible word for ‘stork’ in Yiddish in Lithuania, alongside German-origin štorx and Slavic-origin bušl; however, in mocking speech in Zamet, only the Lithuanian-origin word is used (in describing a tall person). 33. Lemkhen (1995: 118) notes the Lithuanian dialect variant smãgenys alongside sme4genys. 34. Lemkhen (1995) in Yiddish follows in large part, though not entirely, the author’s earlier publication in Lithuanian (Lemchenas 1970). 35. This conforms with a general tendency not limited to NEY; thus, e.g., Polish neuter b³oto ‘mud’ is regularly a feminine in Yid blot6, with final schwa. 36. Note also Swedish barkis (in Stockholm Swedish bergis) ‘a type of braided white bread’, a loan from WY berx6s~ barx6s. Here the Yiddish suffix -6s < Hebrew abstract noun suffix -u˜þ has likely been reinterpreted as the -is suffix under discussion. Whether there is a contact explanation linking Swedish -is and Lithuanian -ys, or whether this is merely coincidence, is unknown to me. 37. Note also augmentative adjectival suffix in NEY -6n6, as discussed in Mark (1951: 459): NEY grejs-6n6 ejgn ‘big eyes’. 38. See also Stankiewicz 1993 on Yiddish thematic verbs. 39. In Jacobs (1994a), I mistakenly referred to Weinreich (1923) as his doctoral dissertation. 40. M. Weinreich (1923: 207) notes that Jews from Zamet would make fun of Courland Jews for “speaking German” for using the verb mext rather than volt in subjunctive constructions. What is at issue here are perceptions, not necessarily the facts, since Weinreich points out the antiquity of the use of mext in Yiddish documents. 41. By the late nineteenth century there had occurred, for some Courland Jews, a shift from Yiddish to German as the home language. However, the present discussion is limited to consideration of CourlYid.
References Ariste, Paul. 1937. Tsu der hashpoe fun yidish oyf nit-yidishe shprakhn. [Includes a review of Kiparsky 1936]. Yivo bleter XI.1–2, 83–85. Ariste, Paul. 1970. Review of Lemchenas 1970. Baltistica 6.2. Chambers, Jack. 1994. “An introduction to dialect topography.” English world-wide 15, 35–53. Herzog, Marvin. 1964. “Channels of systematic extinction in Yiddishdialects.” In: For Max Weinreich, 93–107. Herzog, Marvin I. 1965. The YiddishLanguage in Northern Poland: Its geography and history. Bloomington: Indiana University Press. Herzog, Marvin, Uriel Weinreich z”l, and Vera Baviskar z”l. 1992. The language and culture atlas of Ashkenazic Jewry, Vol. 1. Tübingen: Max Niemeyer Verlag. Jacobs, Neil. 1990a. Northeastern Yiddish Gender-Switch: Abstracting dialect features regionally. Diachronica VII: 1, 69–100. Jacobs, N. 1990b. Economy in YiddishVocalism: A study in the interplay of Hebrew and nonHebrew components. Volume 7: Mediterranean Language and Culture Monograph Series, P. Wexler (ed.). Otto Harrassowitz Verlag: Wiesbaden. [Revised from Neil G. Jacobs. 1984.
310 Neil G. Jacobs
Unpublished Columbia University dissertation: “Economy in YiddishVocalism: The Case of Central Yiddish.”] Jacobs, Neil. 1993. Linguistic geography and phonological instability: Vowel length in East Central Yiddish. Paper delivered at Methods VIII Conference on Dialectology. University of Victoria, Victoria, B.C., Canada, August 3–7, 1993. Jacobs, Neil. 1994a. Structure, Standardization, and Diglossia: The Case of Courland Yiddish. In: Dagmar C. G. Lorenz & Gabriele Weinberger (eds.), Insiders and Outsiders. German-Jewish, Yiddishand German Literature and Culture in Contact, 89–99. Wayne State. Jacobs, Neil. 1994b. On a Structural ‘Fifth Column’ in Sociolinguistic Change: The Diffusion of a Standard Yiddish Feature in Yiddish Dialects. In: Howard I. Aronson (ed.), NSL. 7: Linguistic studies in the non-Slavic languages of the Commonwealth of Independent States and the Baltic Republics, 133–150. Chicago Linguistic Society: Chicago. Jacobs, Neil. 1996. On the investigation of 1920s Vienna Jewish Speech: Ideology and Linguistics. American Journal of Germanic Linguistics and Literatures 8: 2, 177–217. Kalmanovitsh, Zelik. 1926. Der yidisher dialekt in kurland. Landoy-bukh: D”r Alfred Landoy tsu zayn 75stn geboyrntog dem 25stn november 1925 fun zayne gute-fraynt un talmidim. Shriftn fun yidishn visnshaftlekhn institut, Band 1, Filologishe serye 1, 161–188. Katsenelnbogen, U. 1951. Litvakes. In: M. Sudarski (ed.), Lite. Katz, Dovid. 1983. Zur Dialektologie des Jiddischen. In: W. Besch et al. (eds.), Dialektologie: Ein Handbuch zur deutschen und allgemeinen Dialektforschung, 1018–1041. Berlin & New York: Walter de Gruyter. Katz, Dovid. 1987. The proto dialectology of Ashkenaz. In: Dovid Katz (ed.), Origins of the Yiddishlanguage. [= Winter Studies in Yiddish, Vol. 1]. Pergamon Press: Oxford. Kiparsky, V. 1936. Fremdes im Baltendeutsch. Mémoires de la Société Néophilologique de Helsingfors, XI. Helsinki. King, Robert. 1979. Evidence of the German component. Paper presented at First International Conference on Research in YiddishLanguage and Literature, August 6–9, 1979. Oxford Centre for Postgraduate Hebrew Studies. Lazerzon, M. 1942. A bisl baltish-yidishe geografye. Yidishe shprakh 2: 3, 84–87. Lemchenas, Ch. 1970. Lietuviu˛ kalbos ˛itaka Lietuvos Žydu˛ tarmei. Vilnius: Mintis. Lemkhen, Khatskl. 1995. Di hashpoe fun litvish oyfn yidishn dialekt in Lite. In: Oksforder Yidish 3, 6–130. Mark, Yudel. 1944. Di litvishe yidn un der neytraler min. Yidishe shprakh 4.83–94. Mark, Yudel. 1951. Undzer litvisher Yidish. In: Mendl Sudarski (ed.), Lite, 429–472. New York: Kulturgezelshaft fun livishe yidn. Mark, Yudl. 1978. Gramatik fun der yidisher klal-shprakh. Alveltlekher yidisher kultur-kongres. New York. Mathiassen, Terje. 1996. A short grammar of Lithuanian. Slavica Publishers: Columbus. Miller, David Neal. 1990. Ashkenaz: Paradigm and resistance. Paper delivered at conference: The role of geography in Jewish civilization — Perceptions of space, place, time, and location in Jewish life and thought. October 21–22, 1990. The Ohio State University. Plakans, Andrejs, and Halpern, Joel M. 1981. An historical perspective on eighteenth century Jewish family households in Eastern Europe. In: Paul Ritterband (ed.), Modern Jewish Fertility, 18–32. Brill: Leiden. Sapir, Edward. 1915–6. Notes on Judaeo-German Phonology. Jewish Quarterly. Review N. S. 6.231–236. (Reprinted in David G. Mandelbaum (ed.), Selected Writings, 1949, 252–272. Berkeley/Los Angeles: University of California Press.) Shapiro, Moyshe. 1939. Der gramatisher min in yidish. Afn Shprakhfront 3rd series, 3.111–163.
Yiddish in the Baltic region
Stankiewicz, Edward. 1993. The Yiddish thematic verbs. In: David Goldberg (ed.), The field of Yiddish: Studies in YiddishLanguage, folklore, and literature. Fifth collection, 1–10. Northwestern University Press and the YIVO Institute for Jewish Research: Evanston. Thomason, Sarah G. & Terrence Kaufman. 1988. Language contact, creolization, and genetic linguistics. Berkeley, Los Angeles, London: University of California Press. Verschik, Anna. Ms. Some aspects of the multilingualism of Estonian Jews. Weinreich, Max. 1923. Dos kurlender yidish. Chapter 4 in: Shtaplen: fir etyudn tsu der yidisher shprakh-visnshaft un literatur-geshikhte, 193–240. Berlin: Wostok. Weinreich, Max. 1980. History of the Yiddishlanguage. Translated by Shlomo Noble, with the assistance of Joshua A. Fishman. Chicago and London: University of Chicago Press. [Partial translation of his: Geshikhte fun der yidisher shprakh: bagrifn, faktn, metodn. 1973. New York: YIVO Institute for Jewish Research.] Weinreich, Uriel. 1952. Sábesdiker losn in Yiddish: A problem of linguistic affinity. Word VIII, 360–377. Weinreich, Uriel. 1958. Yiddishand colonial German in Eastern Europe: the differential impact of Slavic. American contributions to the fourth international congress of Slavicists, 369–421. The Hague. Weinreich, Uriel. 1961. The Seven Genders of Yiddish. Unpublished paper read before the annual meeting of the Linguistic Society of America, December 29, 1961. Chicago. Weinreich, Uriel. 1964a. Four riddles in bilingual dialectology. American Contributions to the Fifth International Congress of Slavicists, 335–359. The Hague. Weinreich, Uriel. 1964b. Western traits in Transcarpathian Yiddish. In: For Max Weinreich, 245–264. Weinreich, Uriel. 1969. The geographic makeup of Belarusian Yiddish. In: Marvin I. Herzog, Wita Ravid & Uriel Weinreich (eds.), The Field of Yiddish Vol. 3, 82–101. London-The HagueParis: Mouton & Co. Wexler, Paul. 1981. Ashkenazic German. International Journal of the Sociology of Language 30, 119–130. Wexler, Paul. 1987. Explorations in Judeo-Slavic linguistics. Brill: Leiden. Wexler, Paul. 1991. Yiddish — The fifteenth Slavic language. A study of partial language shift from Judeo-Sorbian to German. International Journal of the Sociology of Language, v–150. Wolf, Meyer. 1969. The Geography of Yiddish Case and Gender Variation. In: Marvin I. Herzog, Wita Ravid, & Uriel Weinreich (eds.), The Field of Yiddish Vol. 3, 102–215. London/The Hague/Paris: Mouton & Co.
311
AUTHOR "Aleksandr Yu. Rusakov"
TITLE "The North Russian Romani dialect"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
The North Russian Romani dialect Interference and Code Switching* Aleksandr Yu. Rusakov
1.
Introduction
In addition to the traditional classification of Romani dialects in terms of genetic relationships,1 attempts have recently been made to create a new taxonomy based on on criteria having to with language contact phenomena. Besides the “paraRomani” dialects (in which a “switch” to an “alien” grammar has taken place, cf. Bakker & Courthiade 1991), we could distinguish different types of “inflected” Romani dialects, depending on the degree of interference from other languages (cf. Courthiade 1995: 97–106). In the present article,2 I shall try to show some characteristic features of a dialect belonging to one such “contact type” — an inflected dialect with strong interference, namely the North Russian Romani dialect (NRRD). Two major closely related points will be discussed. First, I shall look at some interference processes that affect the grammatical system, more precisely the system of verbal aspect. Then, I shall turn to a more general discussion of code-mixing and code-switching phenomena. NRRD, which belongs to the Baltic group of non-Vlach Romani dialects, is spread mainly in the northern part of Russia. One of the largest groups of NRRD native speakers live in the suburbs of St. Petersburg. The number of Roma (Gypsies) in St.Petersburg district (Leningradskaja oblast’) exceed 4000 according to the 1989 census, but not all of them are speakers of NRRD. The autonym for the ethnic group is russka roma, haladytka roma. Practically all speakers of NRRD are bilingual. NRRD is the first language of Roma children, but Russian seems to be the dominant language in the situation of Romani-Russian bilingualism. NRRD is a very typical example of strong linguistic interference (stage 4 or 5 of interference: moderate and heavy structural borrowing, according to Thomason & Kaufman 1988). Nearly all levels of its phonological and grammatical system resemble the Russian one. The phonological system of NRRD is practically identical to the Russian one (Eloeva & Rusakov 1990: 11–13). The interference in the grammatical system is manifested: (a) in meaning changes of the elements of
314 Aleksandr Yu. Rusakov
already existing morphological categories; (b) in the structural reshaping according to the Russian models of the forms of already existing morphological categories; (c) in the borrowing of the Russian grammatical markers. Such borrowings sometimes are and sometimes aren’t accompanied by the change in the system of grammatical categories. The word-order syntax of NRRD is practically identical to that of Russian.
2.
Interference in the NRRD verb system: The rendering of aspect
The original tense-aspect system common to the all dialects of Romani seems to be characterized by the aspectual opposition of aorist to imperfect and the single form of present-future, realizing aspectual meaning depending on verb lexical semantics and on the context: (1) bagand’a ‘he, she sang (perfective)’ bagala ‘he, she sings, will sing’
bagavas ‘he, she sang (imperfective)’
In the new verbal system arising under Russian influence, the opposition imperfect : aorist disappears, with the old aorist form serving as a general past, and the imperfect surviving as a special and rarely used aspectual form with iterative meaning, the aspectual opposition are expressed by borrowed prefixes, and the future is expressed by two new constructions: (2) bagand’a s-bagand’a ‘he, she sang (imperfect.)’ ‘he, she sang (perfect.)’ bagala ‘he, she sings’ lela te bagal: s-bagala ‘he, she will sing (imperfect.)’ ‘he, she will sing (perfect.)’
All the examples in schemes (2)–(3) are taken from the syntactic questionnaire. But these schemes reflect rather an idealized picture of the possibilities of express the aspectual meaning. The real picture is more complicated. 2.1 The invasion of Russian verb prefixes General. Verb forms formed with Russian prefixes are frequent in NRRD texts. In the texts recorded in 1984, the following distribution of prefixed forms is found:
The North Russian Romani dialect
Among them: Aorist Synthetic future Present
All finite verbal forms (without Russian borrowings)
Prefixed forms
Ca 500
107
160 072 168
056 019 018
The borrowing of alien prefixes is not an exclusive feature of NRRD: it is observed in some other Romani dialects in contact with languages with rich prefixal systems (cf. Boretzky, Igla 1991: 21–25). Practically each NRRD verb may be used with the borrowed prefixes on the Russian model: (3) a. b. c.
te ot-des ‘to give back’, cf. Russ. ot-dat’ te vy-des ‘to give out’, cf. Russ. vy-dat’ te roz-des ‘to distribute’, cf. Russ. roz-dat’
The Russian prefixes in NRRD preserve all their grammatical and lexical meanings: aspectuality, actionality, possibility to change the verb valency structure and so on. However, it has been noted in the literature that the copying of the Russian aspectual system in NRRD is not complete. N. Boretzky (1989: 368) writes correctly that in NRRD as well as in the other dialects with borrowed prefixes, imperfectivization as a derivational means is absent, cf. also: “Some dialects of Romani have developed a type of prefixal aktionsart similar to that which is found in Slavic or German, but this does not appear to constitute a true superordinate aspectual system” (Friedman 1985: 387). A careful analysis of the functioning of prefixed verbs in texts and in the answers to the syntactic questionnaires makes it possible to modify these statements. Equivalence of prefixed and simple forms. Prefixed and simplex forms of one and the same verb lexeme are often found in the texts and in the answers to the questionnaire without any obvious difference in meaning. Cf. the following pair of examples, drawn from the same text: (4) I avne roma, u-galyne so joj buty kerd’a ‘And the Roma came, (they) discovered that she worked’ (5) Nu dote gyne pal latyr te roden i vdrug galyne… ‘And then (they) went to look for her, and suddenly discovered…’
Commenting upon the questionnaire sentence “He asked money from me, but I
315
316 Aleksandr Yu. Rusakov
didn’t give it him”, the informant notes that it is possible to translate ‘I didn’t give’ either as na dyjom or as na ot-dyjom. On the basis of such examples, we may conclude that the use of verb prefixes for the expression of perfective meanings is not obligatory. On the other hand, prefixed verbs are used (though not very often) to express the meaning of the past imperfect tense. This should probably be seen as a calque of Russian imperfectivized prefixed verbs. (6) I vot, priježali ke joj but moly svatat’. Joj na gyja i daj la na ot-dyja ‘And so, (they) came to her many times trying to marry her. She didn’t go and her mother didn’t give (imperf) her (to them)’ (cf. Russian ne ot-davala “(she) didn’t give (imperf)”.)
Loan translations. N. Boretzky has pointed out that Romani verbs with Russian prefixes are modelled on individual Russian verbs. We are thus dealing with a kind of loan translation. Thus, differences in lexical or actional meaning between the members of Russian verb pairs are often transferred to Romani, as in: (7) phenava “I say, I’ll say” ros-phenava “I retell, I’ll retell”, cf. skažu (Rus) ras-skažu (Rus)
In the same way, Russian aspectual differences carry over to Romani verb pairs: (8) po-pychne “They asked (perf)”; u-cherde “have stolen (perf)” po-prosili (Rus) u-krali (Rus) (Boretzky 1989: 368, cf. also: Beskrovnyj 1971; Barannikov 1934).
Like in Russian, we find in NRRD combinations of “actional-lexical” and aspectual modification of verbs by prefixation: (9) A phen lakiri … vy-gyja pale rakleste pale rom. ‘And her sister … married non-Roma’ (lit ‘went out to the husband’)
An important conclusion can be drawn from the above: grammatical changes in the sphere of aspect representation are brought into NRRD “by lexical means”. The significance of the semantics of the verb. The usage of prefixed vs. unprefixed verb forms is greatly influenced by the semantics of the base verb. Thus, terminative verbs (f.ex. te aves “to come”) are as a rule used in the past and future tenses without prefixes (we mean only the aspectual prefixed forms), and have perfective aspectual meaning: (10) jov javja khere ‘He came home (perf)’
Such verbs may be used sometimes with the imperfective meaning also: (11) Me javja ke jov but molo ‘I came to him many times.’
The North Russian Romani dialect
Non-terminative verbs, on the other hand, may be used with perfective meaning with as well as without prefixes. (12) joj gyja ucˇit’s’a ‘She went (perf) to study’ joj u-gyja lesa ‘She went away (perf) with him’
Are prefixed verbs stored or constructed on-line? A very important, question arises, however. Are prefixed verbs part of the NRRD lexicon, or they are constructed online? It is nearly impossible to answer these questions without direct access to NRRD speakers’ linguistic competence. We may suppose, however, that NRRD speakers use both strategies. The case of prefixed verbs does not differ from the general situation of Russian borrowings in the dialect (see below) or from the processing of word-forms in languages with a highly developed morphology in general. The rather numerous cases of co-existence between Russian prefixed verbs and their Romani equivalents in NRRD texts may be considered as an indirect confirmation of the reality of the “on-line construction” strategy: (13) jov saro pro-igryvaet, graf Cˇornyj, … pro-khella saro ‘…he loses (Rus) everything, count Chornyj, (he) loses everything’
Similar examples are found among the answers to the syntactic questionnaire: (14) (the birds) pri-letajut, pri-urn’ana ‘fly in (Rus), fly in’
There are, however, some prefixed verbs in NRRD which may hardly be supposed to be constructed on-line. I am referring, first of all, to the prefixed derivatives of the polysemantic verb te les “to take, to begin; semi-auxiliary, forming the analytic future”. Consider text examples such as: (15) Daj la pri-lyja… ‘Mother received her’
where Russian: vz’ala “he, she took” corresponds to NRRD lyja and Russian prinjala “he, she received” to NRRD pri-lyja. From the synchronic point of view the Russian prefixed verbs are very obscure opaque derivatives with a non-independent root. On-line calquing is here almost excluded as a possibility. A more likely supposition is the lexical identification was done at an earlier stage of dialect development, perhaps, in the environment of an other Slavic language. Other examples found in our texts are the forms of the verb te ot-keres “to open (the door)”. The model for this verb was no doubt the Russian verb ot-tvorit’ (lit. from-do, work). It is highly unlikely for the interlanguage identification to take place on-line every time the forms are used — the “normal” counterpart of the Romani verb te keres is Russian verb delat’ “to do”, tvorit’ is a rather archaic form.
317
318
Aleksandr Yu. Rusakov
We may suppose, however, that the very correspondence between the Russian and the Romani verbs (here between keres and tvorit’) is “stored” in the lexicon, but the concrete prefixed forms may be constructed on-line according to existing models. Are Russian verb prefixes productive in NRRD? An other important question is whether the loan translation of certain Russian verbs is the only possible strategy for the NRRD speakers, or if they can have a strategy of forming new verbs (or of aspectual forms of existing verbs) by “adding” Russian prefixes to Romani verbs according to productive models, but without association to specific Russian lexemes. In other words, do the borrowed prefixes have their own derivational and/or inflectional potential in NRRD?3 It is very difficult to answer this question. We have, however (both in our texts and in the answers to the syntactic questionnaire), prefixed verb forms which may hardly be explained as calques of specific Russian verbs. See f.ex.: (16) te piri-pas’os ‘to spend a night, lit. through-lie’ (cf. Rus pere-nocˇevat’ ‘to spend a night’ lit. through - spend. a. night)
Consider also the set of prefixed forms based on the verb te chuves ‘to put’, which all denote actions applied to clothing (te ras-chuvespe “undress”, s-chuves ‘take off (one’s clothes)’ Such examples may be treated, on the one hand, as lexical residues of earlier calques of the certain words of other dialect or chronological variants of Russian which thus reflecting preceding stages of NRRD language contacts. It is possible, on the other hand, that NRRD speakers have the strategy of the constructing of new verbs (as well as, possible, of the aspect forms of the existed verbs) with the help of the borrowed prefixes according to productive models. It should be noted, that in the examples cited above the prefixes retain the transparent derivative meaning that they have in Russian. It is clear, of course, that strategies may be combined in different ways and play different roles in the competence of certain speakers. 2.2 The fate of the imperfect The imperfect is very rarely used to express temporal relations.4 In our texts there is only one fragment where imperfect forms were used. The subject is the past nomadic life of the Roma and the imperfect forms designate habitual actions, alternating freely with the forms of the preterite (aorist) and historical present: (17) Nu syr de šatry, nu dzhinde (pret) de šatry roma, rikirde (pret) gren, dadyves rat’kirnas (imp) de dava foro, tas’a tradenas (imp) de javir foro… Kernas (imp) o šatry, roma beshte (pret) po kham, tat’kirnape (pres), romn’a psirna (pres), piri gav dzhana (pres)… ‘How did the Roma live in tents, well, the Roma lived in tents, had the horses, today spent the night in this town, tomorrow went to the other town… Made
The North Russian Romani dialect 319
the tents, the Roma lived with the sun, spent nights, the Roma women go out, go through the village’
In the answers to the syntactic questionnaire, imperfect forms are used to express the combination of two elements — emphasis on the past character of the action and iterative meaning, also expressed by additional, usually lexical means: (18) Ran’še jov cˇasto bagavas ada gily ‘He sang this song often before’
In NRRD the imperfect seems to have the status of a special and rarely used aspectual form with iterative meaning. Thus, the former purely aspectual opposition aorist/imperfect has undergone a fundamental change in NRRD. 2.3 The status of the analytic future in NRRD The degree of grammatical adaptation of the elements which result from Russian interference is often rather difficult to determine. This holds for instance of a wellknown “Slavic” element of the NRRD verbal system such as the analytic future of the type lasa te bagas “we’ll sing”.5 This form seems to be very close to the corresponding Russian construction but is really very rare. In our texts we have only two examples of analytic future compared to 72 examples of future time reference which are not formally different from the present (19 of which have prefixes). We may suppose, however, that the model for this NRRD construction is the alternative half-grammaticalized Russian future model with the verb stat’ ‘become, begin’: (19) Russian ja stanu pet’ ‘I’ll sing’
If it is true, we might suppose that the NRRD analytic future should be treated rather as an inchoative (and, maybe, a not fully grammaticalized) construction, than a future one, cf. (20) Russian ja stal pet’ ‘I began to sing’ (21) NRRD me liya te bagav ‘I began (took) to sing’
2.4 Discussion The extremely heterogeneous system for expressing tense-aspect relations that has been forming in NRRD, is characterized by a high level of opacity and markedness.
320 Aleksandr Yu. Rusakov
Indeed, it is almost impossible to determine the aspect meaning of a verb form taken without a context. If we do not have a sufficient context and a corresponding Russian verbal pair for a given pair (prefixed and unprefixed verbs) of Romani verbs we cannot determine what kind of relations exist between the original and the derived words. On the other hand, as we have seen, in the dialect there is an impressive selection of different means, both pure grammatical and ranging from grammatical to lexical, which can express and in fact do express aspect meanings. There is no doubt, that the situation be found in NRRD is highly marked. Thus, the development of the means of expression of tense-aspect relations from those in “common-Romani language” up to those in NRRD,6 can be described in terms of a transition from transparency to opacity and high level of markedness. In principle, we could expect such a system to be unstable. However, if we look at textual materials from the middle of the 19th century, we find, that on the contrary, the expression of tense-aspect in the dialect have not changed very much since then. For instance, in the collection of Dobrovol’skij (1908), whose materials reflect a state of the dialect which is very close to that of modern NRRD, tense-aspect relations are expressed in a similar way. 65 out of 300 finite verb form taken from the Dobrovol’skij’s collection (Russian lexical loans were not included in the analysis) were prefixed. Like in the modern dialect, prefixed verbs are most often used in the forms of preterite (41 out of 145) and in the synthetic form of the future tense (this form formally does not differ from that of present tense; 12 out of 26). Both in our records and in Dobrovol’skij’s collection imperfect forms are extremely rare. There is also no important changes in the functional use of prefixes: they can modify both lexical and aspectual meanings of the verb. We could also point out many similarities in details.7 In other words, during the last century the means for expressing tense-aspect relations in the dialect have not substantially changed. The marked component of the language system demonstrates a high degree of stability. This is very well in accordance with the fact shown by Thomason and Kaufman. They found out, that in the situation of intensive language contacts, which can result in interference of the fourth or the fifth degree, interference phenomena can increase the degree of the markedness of the language system (Thomason & Kaufman 1988). In a situation where Russian becomes more and more dominating, and Romani language is preserved due to communicative-pragmatic needs, the Russian language system starts playing “a double role.” Russian elements penetrate the system of the dialect, promote the processes of language attrition and hamper the language potency of NRRD itself.8 At the same time, this penetration, paradoxically, increases the viability of the dialect. Russian language can be seen as an inexhaustible reservoir from which the dialect can “scoop” elements from different language levels. It is interesting, that, as we tried to show, the majority of interference phenomena penetrate the language system of NRRD via the lexical level. As for the borrowing
The North Russian Romani dialect
of Russian prefixes, the lexical character of the process can be postulated not only for the origin of this phenomenon, but also synchronically. At the same time (perhaps precisely for this reason), NRRD has not copied completely the Russian tense-aspect system. This is well in accordance with N. Boretzky’s opinion that complete copying or borrowing of a system is not obligatory in interference situations (cf. Boretzky 1989). Actually, in NRRD there is a possibility to express grammatical meanings almost in the same way as in Russian. However, we do not know to what degree the grammatical system of the dialect has incorporated these “russified” types of grammatical expression. We just do not have a method to determine “degree of grammaticalization.” Perhaps the relatively low degree of grammaticalization can be explained by the complexity of the Russian process of perfectivation, with tight interweaving of “lexical” and “grammatical” elements. Another explaining factors is the high level of Russian-Romani bilingualism in the situation where Russian is the dominating language. NRRD native speakers, having a high degree of language competence in Russian, can borrow Russian syntactic and word-formation models on-line. Thus, their “Romani” competence includes not the concrete models, borrowed from Russian, but the possibility of this kind of borrowing itself. A similar situation can be found not only in the system of expressing tenseaspect relations but also, for example, in the voice system. The original system of passive voice expression with -jov suffix now belongs to the domain of nonproductive word-formation. The expansion of the reflexive verbs formed with -pe (-pes) particle was no doubt strengthened by Russian influence. As in the case of prepositions, this refers primarily to loans of specific Russian verbs. Romani reflexive verbs exhibit a plethora of meanings carried over from the Russian sources: direct-reflexive, indirect-reflexive, passive, etc. (cf. Ventcel’ 1964: 74–75). At the same time, NRRD has a special analytical passive construction formed with a linking verb and a participle. which, although rarely used, are known from other Romani dialects and are also supported by Russian influence. Unfortunately, the insufficient number of attested passive forms does not allow us to reach a final conclusion concerning the relationship between these two “new” types of voice meaning expression, their possible connection to the category of aspect (as in Russian), etc. In the nominal domain, the question of direct calques vs. grammaticalized changes in the language system can be put for each case meaning change separately (cf. Ventcel’ 1964: 52–54).
321
322 Aleksandr Yu. Rusakov
3.
Code switching and code mixing phenomena in NRRD
3.1 The concepts of code switching and code mixing With regard to the use of NRRD in different communicative situations, we should note that the dialect shows a very high degree both of code switching and of code mixing phenomena. Whereas the term “code switching” is understood by scholars in more or less the same way, that is, in terms of a shift from one language to another during a monologue or a dialogue, there is no such unanimity about the meaning of the term “code mixing”. Some scholars use it of unrestricted insertion of lexical elements of one language into another language (Hock 1986: 480). Others define it as “the alternate use of two linguistic systems within a clause” (Bhatt 1997: 223). The latter approach treats code mixing as a variant of code switching and implies the possibility of speaking two languages at once. The issue is not merely terminological; its answer relies on the theoretical understanding of the phenomenon. Below we will understand the term code mixing as “the use of unadapted lexical elements of another language in the speech.” Thus, we admit that the possibility of an opposition between code switching and code mixing, though this has yet to be demonstrated. The phenomena of code switching and code mixing have been extensively discussed in recent literature on language contact. On one hand, many investigations have studied the specific factors that determine code switching from a sociolinguistic or pragmatic point of view: what the conversation is about, what real-world objects are referred to, who is spoken to, what is the internal structure of the text. Some attempts have been made to place the phenomenon of code switching in a general anthropological approach to linguistics. On the other hand, much attention is paid to different kinds of structural and functional constraints favoring or disfavoring code mixing (on any understanding of the term). Such constraints include: isomorphy/non-isomorphy of syntactical constructions in the contacting languages, matching/non-matching of grammatical categories, constraints in terms of the notions of Government and Binding theory, etc. (from the recent works on code switching and code mixing cf. Myers-Scotton 1987; Bhatt 1997; Auer 1998). Much less attention has been paid to another aspect of the problem: the possibility of a connection between degree of interference — first of all, grammatical interference — and the occurrence of code switching or code mixing, and if such as connection exists, what its nature is (cf. e.g. Matras 1998: 286).
The North Russian Romani dialect 323
3.2 Code mixing in NRRD 3.2.1General In the NRRD materials, we find, in addition to inter-sential code switching, where the shift from one language to another takes place on the border of two sentences, there are also some examples of intra-sential code switching. One of the most distinct features of NRRD, which sets it apart from many other Romani dialects, is the extensive use of grammatically unadapted Russian lexical elements: (22) Da nat, me prosto na dumind’om, so me tut date vstrechu ‘No, I didn’t just think that I meet you here’
This is exactly what we called code mixing above. In order to reach an understanding of this phenomenon in NRRD we have to answer two closely related questions which have already been touched upon: 1. Can a NRRD text which includes includes unadapted Russian lexical elements really be seen as a text in one language? 2. Can a strict border-line be drawn between phenomena of this kind and classic code switching? Before answering these questions we have to discuss some characteristics of the use of Russian lexemes in NRRD. 3.2.2Adapted and unadapted forms First of all, we should note that in all Roma dialects there are very strong mechanisms of adapting borrowed words. The use of special morphemes inserted between the stem and the inflections in borrowings motivates treating them as a special component of the lexicon which differs markedly from the inherited vocabulary (for the most detailed account of this problem see Boretzky & Igla 1991). In the nominative singular, masculine and feminine borrowed nouns take the endings -o and -a, respectively. (Neuter Russian nouns are assimilated to the masculine gender when adapted.) Since in borrowed nouns the last syllable cannot be stressed, the stress is moved to the first syllable of the stem whenever necessary. In the oblique cases of the singular, borrowed masculine and feminine nouns take the endings -os and -a (without palatalization of the previous consonant), respectively. In the plural, all nouns take -y/-i in the direct case and -en (without palatalization of the previous consonant) in the oblique cases. Borrowed adjectives form a special inflexional class in NRRD as well as in other Romani dialects. The words which belong to this class are characterized by with a possible shift of a stress and lack of differentiation in gender. In the oblique cases, a suffix -on- is inserted between the stem and the ending. In verbs, a suffix -in- is inserted between the Russian verb stem and the Romani inflection.
324 Aleksandr Yu. Rusakov
Thus, the general adaptional mechanism in NRRD are the same as in the other Roma dialects. However, there is one very peculiar difference. Russian adjectives and nouns are practically always used in NRRD in adapted form: (23) tu dzhasa de foro… peskire podrugenca uzhe ‘you’ll go to the city… with his/her female friends already’
Russian verbs, on the other hand, are mostly used in an unadapted form and are conjugated according to the Russian model: (24) tu poedeš de foro ‘You’ll go to the city’ pišem po-romanes ‘We write Romani’
Thus, in the syntactic questionnaire there are 103 unadapted and 17 adapted occurrences of Russian verbs. In one of the texts, there were 35 cases of unadapted verb forms and not a single adapted one. 3.2.3Why the differences between the parts of speech The question thus arises in what manner we should interpret the difference in use of Russian verbs and of adapted nouns and adjectives. Common sense and the necessity to build a simple and psycholinguistically feasible model makes it preferable to regard all the cases of use of Russian nouns, adjectives and verbs as linguistically similar and generated by the same mechanism allowing native speakers of NRRD to use in their Romani speech practically all Russian words (for some exceptions to this rule, see below). But for the nouns an additional rule exists, which is probably active on the sentence production level rather than in the lexicon. Support for this assumption is found in the cases of where the singular and plural forms of nouns are independently adapted: (25) drugo (nom.sg), cf. Russ. drug ‘friend’ druzji (nom.pl), cf. Russ. druz’ja ‘friends’
A separate question is that of the cause of this difference between nouns and verbs. A possible explanation is the simpler nature of noun adaptation mechanisms. Indeed, for the masculine nouns the adaptation mechanism consists only of the addition of the -o suffix to the original noun form, for the feminine nouns in the application of a simple stress rule. The Romani noun has a clear agglutinative structure, thus the formation of the indirect case forms consists of the adding of case affixes to the adapted original form. Adjectives are adapted also by replacing the Russian ending by a Romani one. The adaptation mechanism for verbs is more
The North Russian Romani dialect 325
complicated, it includes the separation of the Russian verb stem, a complicated operation due to the complex character of the Russian verb morphonology. The Roma verb paradigm also contains rather difficult morphonological rules. A further point can be made. Because of the high degree of interference we can nearly always establish a simple correlation between the Romani and the Russian grammatical forms. The absence of the infinitive in the Romani grammatical system constitutes an exception to this generalization which might provoke the intrusion of some unadapted Russian infinitive forms at an early stage. This supposition is supported indirectly by the Hungarian Romani dialects, where Hungarian finite verb forms are borrowed in an adapted form, while the infinitives are unadapted (Leksa Manuš 1979: 252). Then the use of unadapted Russian infinitives might trigger the unadapted use of finite verb forms. 3.2.4A grammatical model for code mixing in NRRD There has been an intensive discussion in the literature of the questions whether similar “mixed codes” have one or two different underlying grammatical structures, and whether there exist some kind of transformational rules, which help to make a transition from one structure to another one (cf. Sankoff & Poplack 1981; Woolford 1983; Bokamba 1988). In cases of strong interference, such as NRRD, this question seems somehow irrelevant, though. The following grammatical model may be postulated for NRRD: The surface syntactical structure of the dialect is identical to that of Russian one. The sentence in NRRD represents by itself the prosodic, syntactical and semantic frame, which coincides with the corresponding Russian sentence9 and into which both Romani and foreign (Russian) lexical elements can be put. The Russian elements may be used in a Russian grammatical setting. The assumption that the examples which were define as “mixed code” can be interpreted as a Romani speech with unlimited insertion of Russian lexemes, receives at least indirect support from metalinguistic comments by native speakers. In the answers to the syntactic questionnaire, which implies the usage of Romani language par excellence, we can find many unadapted Russian lexical elements: (26) me tas’a poedu de foro ‘tomorrow I’ll go to the city’
Interestingly enough, answers of this kind are often followed by comments such as: (27) u nas tak i govorjat: poedu (Russ) ‘we often say it just like this: “pojedu”’
proving, on one hand, that the examples of code mixing do not change the Romani character of the text, and on the other, that speakers are nevertheless able to identify the inserted elements as Russian. The monitoring of the translation provoked by the
326 Aleksandr Yu. Rusakov
request to translate a text from Russian into Romani makes the speaker look for a translation which would not contain borrowed lexemes: (28) atasya me poedu de foro, atasya me dzhava de foro, možno skazat’, so tradava de foro atasya. ‘I’ll go (Russ) to the city tomorrow, I’ll go (Rom) to the city tomorrow, it is possible to say I’ll go (Rom, “archaic” word) to the city tomorrow’
Very often this search of purist translation variants is accompanied by comments in Russian: (29) A kak bystrej, tak i govorjat ‘It’s faster, that’s why we say like this’ (30) My russkie cygane, u nas mnogo russkix slov ‘We’re Russian Roma, we have many Russian words’ (31) Redko, kto znaet cˇisto cyganskij jazyk, potomu cˇto my živ’om sredi russkix i uzhe kak-to samo po sebe slova pereputalis’ ‘It is very rare, that someone knows pure Romani language, because we live among Russians and somehow all the words got mixed up’
It is worth mentioning that not only unadapted but adapted Russian elements also are “introspected” as Russian inclusions: (32) psirlas, — hodine, èto po-russkomu. ‘psirlas, — hodine, that’s in Russian.’
This fact gives indirect support to the hypothesis put forward above of unadapted verb and adapted nouns being elements of the same language level in NRRD. 3.2.5Clear cases of code switching in NRRD Maybe we can see even more clearly the common nature of adapted and unadapted Russian lexemes if we compare these cases with undisputed examples of code switching. In our recordings, there are numerous examples of switches from Russian to Romani and back. We can single out several factors influencing such shifts. On the textual level, to begin with, such shifts can be found when the narrator makes a transition from his own narration to the direct speech of one of the characters of the story:10 (33) Yake ev avya, stadina nahmureno leste, nadvinuto po shero i do zalo dorik, kogda ev de peskiri komnato, o ray pal: nu cˇto, cˇto Pjotr, cˇto ty takoj? Kak cˇto, ved’ ona menja vygnala, von otlupili menja pljotkami… ‘And he came. The hat is on his head, and (goes) into the hall. When he (moves) to his room, the count (goes) after him: “What has happened to you, Petr?” – “What do you mean, she chased me away, they beat me up with a whip.”’
This kind of shift occurs very often in narrative texts (short stories, fairy tales, etc.). In the stories recorded by us, wherever such a transition from the narration to the
The North Russian Romani dialect 327
direct speech takes place, we find in the vast majority of cases typical examples of classic code switching. An analysis of these cases shows us that they most often involve a transition to the direct speech of non-Roma characters. Thus, this transition can be explained purely by pragmatic needs. However, the transition to the direct speech of non-Roma characters is not always followed by code switching. We think that this phenomenon plays a very specific stylistic role, marking the contrast between the direct speech of the characters (including those of fairy tales) and the narration itself. Interestingly, this kind of transition shows up in the most important and culmination parts of a story. For instance, in one short story about a man who was buried alive (the plot is quite widespread), we can find many examples of direct speech but they are all in Romani, no matter which nationality the character has. The only exception is found in the most culminational moment of the story, when the man who was buried alive addresses the robbers who excavated him. In this connection, it is interesting to point out that the end of a story is also marked very often by the transition to Russian: (34) na ètom i koncˇilos’ ‘that’s how it ended’ etc.
An additional proof of the fact that code switching in a story marks the border of stylistically different fragments of the text, can be found in our records: often the transition to Russian language takes place after the end of the direct speech. Another question, which is quite interesting, is to what degree a narrator realizes that he makes the shift from one language to another. It is very typical when a speaker continues to tell the the story in Russian after the direct speech is over. Then he suddenly recollects this fact and after a little pause switches back to Romani. In general, after listening to the records of the Romani texts, one gets an impression that, for some degree, it does not really matter for the narrator in which language to speak, and that the transitions from one language to another are made because of the need to make the text more “picturesque” stylistically, rather than to tell some specific fragments of the text in one language or another because of their genre.11 This supposition does not deny, of course, the fact that the transition is made in some particular parts of the texts (direct speech, commentaries, etc.). At the same time, at every single moment the narrator can realize which language he speaks, though sometimes he seems to forget to speak Romani, as in the above mentioned examples, when he keeps on talking in Russian, though the original setting was to tell this part in Romani.12 A slightly different situation can be found in the texts dealing with everyday life. In this case the transition to Russian language seems to be caused not by stylistic but by thematic reasons. A certain topic can make the speaker to use many Russian words. That, in its turn, causes the speaker to make a transition to Russian language, because it is easier to talk about this in Russian, the situation quite typical for code switching.
328 Aleksandr Yu. Rusakov
3.2.6The borderline between code switching and code mixing Let us return to the second question put above: Is there any strict boundary between code-mixing and code-switching? This question has been answered in the negative by quite a number of specialists (e.g. Myers-Scotton 1997: 23–24). Nevertheless we can try to distinguish between the two phenomena from a formal point of view. As code-mixing is typical of the Romani speaking only (when they speak Russian they do not use Romani elements in their Russian speech), we should distinguish between the cases of Romani speaking that includes Russian lexical elements and the cases of speaking Russian. The following test criterion of the Russian character of the text and, thus, the switch to Russian could be proposed. In the speech of NRRD-speakers there are the elements which mark speaking Russian. These elements are never used in the texts we have the reasons to suppose to be Romani. From the “operational” point of view we can postulate that the word directly preceding and the word immediately following such an element inside one sentence could not be both the Romani ones Among these elements first of all are Russian nouns in oblique cases in Russian grammatical form. We have already mentioned above that Russian nouns, when used in NRRD, are nearly always used with Romani grammatical setting. The exceptions are those of the nouns which are used in the idiomatic adverbial and prepositional phrases. (35) so vremenem ‘as the time passes by’ rano utrom ‘early in the morning’; the use of nouns in oblique cases in the meaning of adverbs (36) rundja golosom ‘yelled, lit. cried with the voice’
also the cases when nouns are used with Russian numerals: (37) ev dodzhindya do sto let ‘He lived till he was one hundred years old’.
For nouns in the direct case it is often difficult to define whether they are adapted or not, for example for Russian feminine gender nouns with non-final stress (e.g. doróga) “adaptation” flexion coincides with the Russian one and there is no stress shift. Thus, if we come across the form of Russian nouns inflected according to Russian grammatical system, we can conclude that this fragment is pronounced in Russian. Indeed, in our texts it is almost impossible to find the Russian oblique case forms in the environment of Romani lexical elements. Russian personal pronouns as well as adjectives in oblique forms inflected according to the Russian grammatical system could be also considered as the indicators of the Russian character of the text. The texts are lacking Romani forms of the Russian verb byt’ ‘to be’.
The North Russian Romani dialect 329
There are, however, Russian elements which being used in speech of Roma do not indicate code-switching; such elements are either in the relations of free variation with their Romanified variants or simply do not have Romani equivalents. Among these elements are 1. some verb forms, already mentioned above. As we have pointed out earlier, Russian verbs can be used in an adapted form, though the cases of unadapted use are much more frequent. It is interesting to note, that our oldest informant, a 90-year-old Roma woman L. M. Stepanova, used more adapted verb forms than other members of her family (below we discuss how to distinguish different styles or registers in Romani speech). On one hand, it is possible that that the use of adapted/unadapted forms depends on some individual lexical characteristics of the verbs. Every single verb can in principle be used in its adapted as well as in unadapted form. There are some verbs however, which are used in their adapted form much more frequently than in unadapted form. 2. adverbs formed from adjectives, which occur both in adapted and unadapted forms. Cf.: bystro ‘quickly’ — bystres. We could not find any regularity in the distribution of these forms. 3. function words: a very complicated relationship exists in NRRD between Romani and Russian function words. So, for example, we have in our texts both Romani kai and Russian gde ‘what’. Of course, this method of differentiation between Romani and Russian “pieces” of a NRRD text is preliminary and rather relative. Sometimes this method does not allow to achieve an unambiguous result. Interestingly, the code mixing situation in modern NRRD is very much like that found in the collection of Dobrovol’skij (1908). (38) a my dumali, so rashaj, azhno kozlo (Dobrovol’skij 1908: 5) ‘we thought, that a priest, but it’s a goat’
However some differences can be found. For instance, in the texts Dobrovol’skij (1908), practically all Russian nouns and adjectives are used in an adapted form. As for verbs, the ratio between adapted and unadapted forms in the texts collected by Dobrovol’skij is a little bit different than that in the modern NRRD texts. In the prose text of the collection the above mentioned ratio is around 80 to 70. The same verb forms could be used both in their adapted and unadapted forms. Nevertheless, some verbs tend to be used in adapted forms, when another ones almost always are used in their unadapted variant (e.g. moch’ ‘to be able’, dolzhen ‘to have to’, and other modal and half-auxiliary verbs). It is interesting that in the prose texts of the collection we found only two examples of code switching, despite the fact that many of the texts are quite similar with the texts recorded by us.
330 Aleksandr Yu. Rusakov
Another curious fact is that these two cases of code switching in the Dobrovol’skij’s texts are found on the border between plain narration and direct speech: (39) Lyja te schupine. “Schupine sygydyr”, kricˇit na zasekax sedja, “Da u buje da gazhes” (Dobrovol’skij 1908: 64) ‘Started to touch. “Touch faster”, he yells, sitting near an abattis, “you damned man!”’
In this example the narration is marked by the transition to Russian language. Here too, as in the case with the borrowing of Russian prefixes, the dialect shows very high degree of stability. Thus, in NRRD speech Russian lexical elements can be used with practically no restrictions. The number of Russian words can reach such a level that there is an illusion of a “mixed Russian-Romani text.” However, as we tried to show, there is a possibility to differentiate Romani texts and the cases of code switching. Most probably, this differentiation has some kind of psycholinguistic basis. We think, that both code mixing and grammatical interference (at least, some of the phenomena, which are traditionally considered to be grammatical interference) have a similar nature. Both code mixing and grammatical interference are possible because of the existence of some common processes, which allow the speaker to use the resources of Russian language right in the moment of locution. In the first case, the speaker uses “material” lexical elements (some of them, especially nouns and adjectives, are adapted according to special rules). In the second case we deal with translation loan. It can be a word-formation model of individual Russian words which is borrowed, or it can be a whole syntactic model. It is important to note that this kind of phenomena are very wide-spread and are used practically by all the native speakers of the dialect, thus being an integral part of their language competence (cf. for example some speculations on code mixing as a discourse mode in Pyemi 1993). We should also point out that this approach erases the distinction between interference processes in language and interference processes in speech, which was noticed by Weinreich (1953). The observed phenomena being realized in the moment of speech surely influence the language system. Word-formation and syntactical loan translations gradually replace the previous modes of grammatical expression, and code-mixing-like phenomena lead to the erosion of the lexicon. Undoubtedly interference phenomena and the code-mixing influence each other. Maximal approximation of NRRD syntactical structure to that of Russian results in the abolishment of the restrictions to use code-mixing, and code-mixing in its turn provides growing similarity of two language systems (by inclusion of Russian grammatical elements into Romani texts first of all).
The North Russian Romani dialect
3.2.7The domain of validity of the NRRD situation In connection with the situation described above two questions arise: 1. To what extent is the code switching and above all the code mixing situation with code mixing in NRRD typical also of other Romani dialects and other languages with intensive language contacts? 2. Can we assert that this situation is typical for NRRD in general, and not for some specific social register? When answering the first question we have to note that code mixing phenomena are found in some other Romani dialects such as the Ukrainian and Southern Russian dialects analyzed by A. P. Barannikov (1934: 139 and elsewhere) and in non-Romani languages of the previous Soviet Union such as some German dialects (cf. Verešcˇagin 1965; also Smirnitskaja & Barotov 1997: 83–96). At the same time code mixing of such a type is not common in a large number of Romani dialects,13 although “classic” code switching is found almost everywhere. A very interesting situation can be observed in the Romani dialects of Kabudzhi in Albania and the Agia Barbara dialect in Greece. Some time ago the native speakers of these dialects were in a very close contact with the native speakers of Turkish. At the moment only some members of the elder generation of these two dialects (Kabudzhi and Agia Barbara) are somewhat familiar with Turkish (cf.: Cortiade 1992: 8; Messing 1988: 27–28). Nevertheless, in both dialects we can find many originally Turkish verb with the Turkish grammatical setting. Thus, we can assume, that earlier, while these dialects were still in close contact with Turkish, code mixing type of phenomena were typical for them. Then, after the contact was over, some kind of grammaticalization process took place and Turkish verbs, which became real borrowings, formed a separate paradigm. It is interesting, that the modern state of both dialects is not characterized by code mixing. We cannot find any unadapted loans either from Albanian, or from Greek. It is also difficult to say if code mixing is typical only of some specific (social of stylistic) variant of NRRD or it is a feature of NRRD in general. Practically all our informants use code mixing in their speech. The situation in the dialect described by Dobrovol’skij proves that this phenomenon has undergone some kind of “institutionalising”. Nevertheless, we cannot deny that there may exist a variant of NRRD without use of unadapted Russian loans.14 Researchers postulate that for some Romani dialects there are two “styles” or registers of speech: “higher” and “lower”. For instance, such a situation is reported to exist in Finnish dialect of Romani language (cf. Valtonen 1972). In this case “the higher style” is similar to “an older, more original form of the language, while “the lower style” refers to a more recent, grammatically simplified Romani, which is heavily influenced by Finnish” (Vuorela & Borin 1994: 10–11). However, as the authors report, these language forms are better seen as a continuum, and the degree
331
332 Aleksandr Yu. Rusakov
of interference depends on individual speaker characteristics, above all age (ibid.). Possibly NRRD also represents such a continuum, as suggested, for instance, by the following. In spite of the absence of texts without code mixing in our material, the number of unadapted Russian lexemes varies between individual informants15 and depending on text type. Unfortunately, the material is not large enough to allow a statistical analysis, but presumably the situation in question can be described by Labov’s variable rule model. It would be interesting to carry out a sociolinguistic investigation of NRRD, analyzing how linguistic variables (including usage of unadapted Russian lexemes/adapted borrowings/genuine Romani words) are used depending on social characteristics of informants and type of communicational situation. Such an investigation could help to answer in a more precise way the question of the psycholinguistic reality of on one hand code switching-models, implying bilingualism in the speaker and on the other the Labovian model, which involves a monolingual speaker using variable rules (cf. Labov 1970), in consonance with the code mixing notion. Presumably, a concrete speaker can use both strategies in his language performance. Very interesting is the question of the relationship between the NRRD type of dialects on one hand, and mixed Romani dialects, on the other. Dialects such as Finnish Romani are evidence for a gradual transition from a dialect with strong interference to a dialect which has lost its original, “genuine” grammatical component. 19th century British Romani may be another case in point (cf. Yeloyeva & Rusakov 1990: 39–40). There is an apparent contradiction here with the rather widespread viewpoint according to which para-Romani dialects are artificial or semi-artificial formations like argot (a review of the origins of this type of dialects can be found in Bakker & van der Voort 1991; cf. also Boretzky & Igla 1994). We must here make a clear distinction between the history of the “para-Romani dialects” and their current synchronic state. Indeed, at least some of the paraRomani dialects are very close to argot both functionally and structurally, as they – – –
are used as secret languages; are not the first language to be acquired by children; the ways of the creation of new words in these dialects resemble argot (periphrastic constructions are widely used, some word-formation suffixes are used in a specific way, existence of “cryptolalic” formations, etc.) (cf. Eloeva & Rusakov 1990: 41–43, Bakker 1995: 134).
Another question is how the current state of these dialects came about. We think that the most “attractive” view is that of D. Kenrick, according to which present-day British Romani is a result of a gradual development from a Romani dialect, during which a great number of English grammatical constructions penetrated into the language system (Kenrick 1979; a similar view is expressed in Thomason & Kaufman 1998: 103–04).
The North Russian Romani dialect 333
Can we consider the NRRD type of dialects to be a stage in the development of para-Romani dialects? In other words, can a wide use of code mixing stimulate the borrowing of the grammatical system of a neighbouring language? We can imagine that the paradigms which were borrowed into NRRD along with unadapted Russian lexical elements were later on extended to genuine Romani words. However, it seems that both code mixing phenomena and “the borrowing of another grammatical system” should be regarded in the frame of one process, that is a convergence of a Romani dialect and a neighbouring language. The NRRD language system is becoming more and more similar to that of Russian. (In reality this process realizes in the competition between the elements of the original system and the elements inducted by interference.) Grammatical meanings can be expressed both by Romani and Russian elements: At some point, the system of the dialect allows Russian lexical elements, which have the same syntactic functions as Romani ones, to be used freely. It is possible that in other Romani dialects the equivalence between Romani and non-Romani grammatical elements, at some stage of development of a dialect, could result in their variational usage in Romani texts. Later the original Romani ways of expression of grammatical meaning would be replaced by non-Romani ones (cf. also Myers-Scotton 1998).16 In our opinion, one of the advantages of the suggested model is that the language changes are seen and treated as rather homogenous. It can be seen, for example, from the priority of the lexical changes in the interference processes. Some details of this process, however, are still to be uncovered. It is unclear, for instance, exactly in which way the original grammatical system is replaced. In this connection, we would like to point out the following: 1. From the psycholinguistic point of view, para-Romani dialects are extreme points on the way to the type of language contact situation what in the beginning of the century Šcˇerba called “one language with two terms”.17 In this case it is not clear if we can speak about a bilingual situation (in the broad sense) or if we are dealing with one language system, which has lexical language variables (as it was understood by Labov). In any case, it is quite evident that some kind of psycholinguistic research of Romani dialects should be carried out. 2. From a sociolinguistic point of view, there is no doubt that those dialects which have a high degree of code mixing as well as para-Romani dialects are two stages on the way to what is called with an elegant term “language death”. Indeed, out of 13 symptoms of “language death”, listed in Dressler (1996), the dialects of NRRD type share 11. Those 11 symptoms are: – – – – –
massive lexical loans from the dominant language; borrowed lexical items that are not integrated but are used like citation words; borrowed words that are not integrated and are used like regular words; indigenous words replaced by borrowed words; speakers inability to use the indigenous lexicon appropriately;18
334 Aleksandr Yu. Rusakov
– – – – – –
loss of use of proper names of the recessive language; loss of productivity of word formation rules; phonological alignment with the dominant language; decreased use of subordinate clauses; replacement of synthetic by analytic constructions; use of only a casual, conversational style of the language.
As for the remaining symptoms on Dressler’s list, it is questionable if Romani dialects of this kind are characterized by “decay of inflectional morphology.” Also, the speakers of these dialects definitely do not have “negative attitudes … toward their language”. The absence of the last feature can be explained by the role of the Romani language as an “ethno-preserving” factor. Because of the heterogeneous nature of para-Romani dialects, the application of the symptoms to them is more problematic, in particular those which deal with lexical borrowings. We cannot deny, however, that they represent probably highly striking examples of “language attrition”. At the same time, both NRRD type of dialects and para-Romani dialects show a relatively high degree of language stability. The reasons for this stability are evident. There is a need to preserve the Romani language as a special, secret language, which cannot be understood by outsiders (cf. Thomason & Kaufman 1998: 103–04). That is why this stability exists despite the dominance of the language of the neighbouring population and the consequent very high degree of interference. In this connection we find very interesting N. B. Vaxtin’s ideas about the possibility of preservation of a language as an “ethnical identifier”, notwithstanding drastic changes of the language. It would very interesting to investigate the role of intentional efforts of language community members in the process of preserving the language (cf. Vaxtin 1998: 122). For example, all the original grammatical elements in para-Romani dialects have been replaced, but that is not in contradiction with the goal of preserving the Romani character of these dialects.
Notes * This work was supported by the Research Support Scheme of the OSI/HESP, grant No: 616/1996. 1. According to this traditional division, the Romani dialects in the circum-Baltic area, like those in the rest of Europe, are divided into two main branches, on one hand, non-Vlach dialects, whose speakers migrated to Central and Western Europe in the 15th century, and Vlach dialects, which spread from Romania to other parts of Europe in the second part of the 19th century and later. Here, we shall speak only about the non-Vlach dialects, which are divided into several groups: – – –
the Baltic group: the dialects of Central Poland, Lithuania, Latvia and Estonia, the North Russian Romani dialect and the very close related dialect of Belarusian Roms; the Sinto dialect; the dialects of Finnish Roms.
The North Russian Romani dialect 335
In the area, we also find Swedish and Norwegian “para-Romani” varieties (see below). All these dialects could according to some linguists be united in the North branch of Romani (non-Vlach) dialects (Boretzky, Igla 1991: 7–8). 2. The materials for this article (syntactic questionnaire and texts of different genres) were mainly collected through fieldwork in 1984–88 and 1996 in the Vyrica and Mixajlovka villages near St. Petersburg, in collaboration with F. Eloeva, E. Perexval’skaja and O. Abramenko. 3. Verb derivation by prefixation is treated in traditional Russian grammar as a word-formation process, regardless of whether the prefixes modify only verbal aspect, or the lexical meaning. 4. Imperfect is used rather regular as the conditional in NRRD. 5. It is possible that besides the Russian model budu pet’ ‘I shall sing’ the NRRD model was influenced by ” the Ukrainian one as in xot-im-u want-take-1sg.pres. ‘I’ll want’ (Boretzky, 1989: 369). If it is true, we have here a very interesting example of a “contamination” of the influence of the two synonymous models, one of which influenced the syntactic structure while the other the “lexical” semantic structure of the future construction in NRRD. It is interesting that the other model of analytical future more close to the Russian one exists in NRRD: me avava te bagav ‘I’ll sing’ (lit. ‘I’ll be sing’). This form is videly represented in Dobrovol’skij’s texts (1908) and rarely in our materials. 6. Of course, we should clearly realize, that this “common-Romani language” is, firstly, more a linguistic construct, and, secondly, is a product of interference influence of Balkan languages (first of all, of Greek language) too. 7. For example, in the texts collected by Dobrovol’skij as wel as in our texts we can find quite a popular verb te hodinas ‘to walk’, which was borrowed from Russian and then adapted. It is interesting to note, that this verb forms a quasi-aspectual pair with a Romani verb te dzhas. This pair is parallel in all respects to the Russian verb pair khodit’ – idti ‘to walk’ (in general) – ‘to walk’ (purposefully). In both pairs, the verbs are opposed to each other by the distinctive feature of unidirectionality. In the dialect we can find a few more pairs which would contain two verbs of movement, where the first verb is a “genuine” Romani one, the second one is an adapted Russian verb. For example: te prastas – te begines ‘to run’, te urjes – te letinas ‘to fly’. Unfortunately, we do not have enough material to decide if an opposition similar to the above mentioned one exists between the members of these pairs. 8. In that connection, note that NRRD is characterized by loss of productivity of word formation rules, the feature which was mentioned by Dressler (1996) as one of most symptomatic ones for a situation of language death. Unfortunately, we cannot examine this issue in detail here. 9. Pfaff (1979) expressed similar views on the factors favorable for code switching (cf. also Thomason & Kaufman 1988: 154). 10. We should note, however, that the majority of the texts analyzed are quite homogenous genrewise — fairy tales, tales, “true stories”, etc. 11. Of course, we should take in account that the researcher is Russian, and this factor can provoke the informant to start speaking Russian. At the same time the informant is “controlled” by the original directions (also given by the researcher) to speak Romani. 12. Sometimes, however, it seems that it is easier for the narrator to speak Russian and he forgets to return to Romani. 13. It should be noted, that very often there is no reliable evidence to decide whether code mixing exists in the dialect. 14. A perfect example of such a variety is the artificial “standard Rom language” created in the Soviet Union in the 1920s on the basis of NRRD. It is interesting, however, that in the plays (that
336 Aleksandr Yu. Rusakov
is the genre of literature which is oriented mostly towards spoken language) written in this language by Rom-Lebedev, there are cases of code mixing: Gadzhengire buty kerna, kos’at, zhnut. (Rom-Lebedev 1931: 29) They do gajos’ work, mow, reap Unadapted verb forms are used here in the place of rarely used Romani verbs. (On the problems of standard language based on NRRD cf. Rusakov 1997.) 15. For example, there are fewer cases of code mixing in the speech of our oldest informant L. M. Lobanova, whom we have mentioned already. 16. We have mentioned above extensive overlapping of the Romani grammatical system (in the NRRD type of dialects) and the grammatical system of the dominating language. We can suppose, that in para-Romani dialects grammatical system coincides with that of the dominating language even on the surface level of morphology, that is on the level of “material” signifiers. 17. In modern psycholinguistics the notion of compound bilingualism can be considered as an equivalent to the Šcˇerba’s term. 18. This can be also found in the fact that very often speakers of NRRD (as the speakers of other Romani dialects) use the borrowed words instead of Romani words, familiar to them, even in the answers to the questions of the questionnaire.
References Auer, Peter (ed.). 1998. Code-Switching in Conversation: Language, Interaction and Identity. London: Routledge. Bakker, Peter, Cortiade, Marcel (eds.). 1991. In the margin of Romani. Gypsy languages in contact. Amsterdam: Institute for General Linguistics. Bakker, Peter, van der Voort, Hein. 1991. Para-Romani languages: an overview and some speculations on their genesis. In: Bakker & Cortiade, 16–44. Bakker, Peter. 1995. Notes on the genesis of Calo and other Iberian Para-Romani varieties. In: Matras, Yaron (ed.), Romani in Contact. Amsterdam: Benjamins, 125–150. Barannikov, A. P. 1934. The Ukrainian and South Russian Gypsy Dialects. Leningrad. Beskrovnyj, V. M. 1971. O vtoricˇnoj prefiksacii glagola v indoarijskom jazyke (na materiale cyganskogo jazyka. In: Vostocˇnaja filologija, Moscow, 14–20. Bhatt, Rakesh M. 1997. Code-switching, constraints, and optimal grammars. Lingua 102: 223–251. Bokamba, Eyamba G. 1988. Code-Mixing, Language Variation and Linguistic Theory: Evidence from Bantu Languages. Lingua 76: 21–62. Boretzky, Norbert. 1989. Zum Interferenzverhalten des Romani. Zeitschrift fur Phonetik, Sprachwissenschaft und Kommunikationsforschung. 42: 357–374. Boretzky, Norbert, Igla, Birgit. 1991. Morphologische Entlehnung in den Romani- Dialekten (Arbeitspapiere des Projektes “Prinzipien des Sprachwandels”, Nr.4). Essen. Boretzky, Norbert, Igla, Birgit. 1994. Romani Mixed Dialects. In: Peter Bakker & Maarten Mous (eds.), Mixed Languages. 15 Case Studies in Language Interwining. Amsterdam: IFOTT, 35–68. Cortiade, Marcel. 1992. I Rom in Albania. Un profilo storico e sociale. Lacio Drom, 28: 3–14. Courthiade, Marcel. 1995. Phonologie des parlers rom et le diasystème graphique de la langue Romani. Thèse pour le doctorat. Paris. Dobrovol’skij, V. N. 1908. Kiselevskie cygane. St. Petersburg.
The North Russian Romani dialect 337
Dressler, Wolfgang. 1996. Language Death. In: Singh, Rajendra (ed.), Towards a Critical Sociolinguistics, 195–210. Amsterdam: Benjamins. Eloeva, Fatima A. & Rusakov, A.Ju. 1990. Problemy jazykovoj interferencii. Cyganskije dialekty Evropy. Leningrad. Friedman, Viktor A. 1985. Balkan Romani Modality and Other Balkan Languages. Folia Slavica 7: 381–389. Hancock, Ian F. 1988. The development of Romani linguistics. In: Jazayery, Mohammad Ali & Winter, Werner (eds.), Languages and Cultures: Studies in Honor of Edgar C. Polome, 183–223. Berlin: Mouton de Gruyter. Hock, Hans H. 1986. Principles of Historical Linguistics. Berlin: Mouton de Gruyter. Kenrick, Donald. 1979. Romani English. In: Hancock, Ian F. (ed.), Romani Sociolinguistics (International Journal of the Sociology of Language 19). The Hague: Mouton. Labov, William. 1970. The study of language in its social context. Studium Generale. 23: 30–87. Manuš, Leksa. 1979. O kontakte cyganskogo i vengerskogo jazykov. Sovetskoe finnougrovedenie. 15: 248–255. Matras, Yaron. 1998. Utterance modifiers and universals of grammatical borrowing. Linguistics 36: 281–331. Messing, Gordon M. 1988. A Glossary of Greek Romany as Spoken in Agia Varvara (Athens). Ohio: Slavica Publishers, Inc. Myers-Scotton, Carol. 1997. Duelling languages: grammatical strucure in codeswitching. Oxford: oup. Myers-Scotton, Carol. 1998. A way to a dusty death: the Matrix Language turnover hypothesis. In: Grenoble, Lenore A. & Lindsay J. Whaley (eds.) Endangered Languages. Cambridge: cup, 289–316. Pfaff, Carol W. 1979. Constraints on language mixing: intrasentential code-switching and borrowing in Spanish/English. Language 55: 291–318. Rom-Lebedev, Ivan. 1931. Kham dro blato. Moscow. Rusakov, Aleksandr. 1992. Imperativ v severnorusskom dialekte cyganskogo jazyka. In: Xrakovskij, V. S. (ed.), Tipologija imperativnyx konstrukcij. S.-Petersburg: Nauka. Rusakov, Alexandre. 1997. The education of Russian Gypsies: The problem of dialects. In: Vassiliadou, Maria (ed.), Education of Gypsies. Development of Teaching Material. Athens. Sankoff, Gillian & Poplack, Shana. 1981. A formal grammar for code-switching. Papers in Linguistics, 14: 3–46. Smirnitskaya, Svetlana V. & Barotov, M. A. 1997. Nemeckije govory Severnogo Tadžikistana. S.Petersburg: Institut lingvisticˇeskix issledovanij RAN. Thomason, Sarah Grey, Kaufman, Terrence. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Valtonen, P. 1977. Trends and special traits in Finnish Romani. Studia Orientalia, 47: 281–286. Vaxtin, Nikolaj B. 1998. Iscˇeznovenie jazyka i jazykovaja transformacija. In: Kozinceva N. A. & Ogloblin, A. K. (eds.), Tipologija — Grammatika — Semantika. K 65-letiju Viktora Samuilovicˇa Xrakovskogo. S.-Petersburg. van der Voort, Hein. 1991. The Romani dialect(s) of the Finnish Gypsies. In: Bakker & Cortiade, 132–151. Ventcel, Tatiana V. 1964. Cyganskij jazyk (severnorusskij dialekt). Ìoscow. Verešcˇagin, E. M. 1965. K voprosu raznojazycˇnogo oformlenija ispol’zuemoj leksiki v recˇevyx proizvedenijax bilingvov. In Voprosy obšcˇego i romano-germanskogo jazykoznanija. Tezisy dokladov. Vyp. 1. Ufa. Vuorela, Katri, Borin, Lars. 1994. The Finnish Gypsies and Finnish Romani (ms.). Weinreich Uriel. 1953. Languages in Contact. The Hague: Mouton. Woolford, E. 1983. Bilingual code-switching and syntactic theory. Linguistic Inquiry 14: 520–36.
AUTHOR "Valeriy Cˇekmonas"
TITLE "On some Circum-Baltic features of the Pskov-Novgorod (Northwestern Central Russian) dialect"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
On some Circum-Baltic features of the Pskov-Novgorod (Northwestern Central Russian) dialect Valeriy Cˇekmonas
1.
Introductory remarks
The Baltic and Finnic languages are flanked geographically on the North-East by the dialects of North-Western Russia, that is those of the Pskov, Novgorod and St. Petersburg regions. In Russian dialectology, the Novgorod and Pskov dialects are regarded as belonging to the north-western part of the Central Russian dialect area (sredne-russkie govory). The Novgorod dialect is characterised by okan’e, that is retention of o in the first pre-tonic syllable, and the Pskov dialect has (strong) nondissimilative akan’e-jakan’e (or neutralisation of o and a in the first pre-tonic syllable into a, regardless of the quality of the vowel in the next stressed syllable). Genetically, they are closely related, being derived from the old East Slavic dialects of the Krivichi and Slovene. It is commonly assumed that those East Slavic groups that primarily colonised the areas around Lake Ilmen’, along the eastern banks of Lake Cˇudskoe and the banks of the Velikaja river were in close contact with the local Finno-Ugric population. The East Slavs could hardly have reached these areas before the end of the 8th century AD1 (Goehrke 1992: 31). From this time on there was intimate interaction between the Slavic and the predominantly Finnic cultural traditions up until the 10th century AD. Considering that the Slavs “were settled here among the Finnic population, there are, naturally, some features in their [archaeological — V.Cˇ.] culture, which are comparable to those of the western Finnic cultures” (Sedov 1994: 9). The culture of the long kurgans, which are often considered to have been left by Krivichi, has obvious Finnic features (Sedov 1970b: 96, 1994: 3–17). Moreover, some scholars suggest that this culture was developed by both local Finns and Slavs. The so-called kruglye sopki (“the round knolls”) in the central part of the Novgorod Land dating from the 6–8th century AD most likely belonged to the Slavic population mixed with the Finnic autochthones as well (Sedov 1970a: 1–8). For a critical analysis of the problem see Bjørnflaten (1995, 1998).
340 Valeriy Cˇekmonas
The northwestern part of the Novgorod Land (Novgorodskaja Zemlja) belonged to the Vod’, one of the Baltic Finno-Ugric tribes. The ancient population did not leave their place of residence — the Slavs on the Ižora plateau and along the shores of Lake Cˇudskoe settled among the Vod’ villages. The result of territorial confusion between the Slavs and Vod’ was cultural accommodation and the gradual assimilation of the local Finnic population” (Sedov 1982: 174; 183–184). Old Pskov (presumably earlier Izborsk < *Isa-borg)2 was founded near or on the site of the ancient Finno-Ugric settlement. (Beleckij 1996: 78–86)
The propagation of Slavic speech within the limits of the contemporary Novgorod and Pskov dialects was due not only to the diffusion of the Slavic population from the most ancient places of its residence in this region but also to assimilation of the original Finno-Ugric population. This process has continued to the present time in the Pecˇory district, where the remnants of the so-called Rus. poluvercy ‘Halfbelievers’ or “Orthodox Estonian” or Setu population still preserve their language (a variety of the southeastern Estonian dialect) and ethnic identity. This is also true of the northern part of the Gdov district, where “the assimilation of the remnants of the local Vod’ is continuing now (or has recently been completed) in a Russian milieu” (Gluskina 1973: 50). A strip of Russian accents on the Gulf of Finland in the St. Petersburg region is regarded as a new (and mixed) variety that developed on Votian and Ingrian substrata and was strongly influenced by St. Petersburg’s koine¯. “These facts allow the territory under review to be eliminated from the area of the main old East Slavic colonisation” (Stroganova 1970: 395). The above summary allows us to assume the existence of traces of the Finnic language substratum in the Pskov and Novgorod dialects. These could be comparable to those in the archaeological culture of the ancestors of the present Russianspeaking population and, consequently, one may speculate that these dialects may share some common linguistic features with the southern Finnic languages. The main consequences of the long tradition of treating some peculiarities of the Russian language and its dialects as having been developed under the influence of a Finno-Ugric substratum were summarised in Veenker (1967), although it is often claimed that “empirical evidence is hard to get at” (Dahl & KoptjevskajaTamm 1992: 8). This is because most of the phenomena attributable to the influence of a substratum can also be explained as the results of immanent (internal) evolution. Besides, the idea of substrata in Russian and its dialects is least admissible in the East Slavic linguistic tradition, especially when it comes to possible Finno-Ugric substratum influence. Such ideas are usually treated as being of less importance and peripheral with respect to the main stream of Russian historical linguistics. One explicit attitude to the Finno-Ugric substratum could be expressed by the words: “yes, such a substratum may be possible in most of central Russia, it is absolutely obvious in Northern Russia from a historical point of view, but in spite of that it has
On some Circum-Baltic features of the Pskov-Novgorod dialect 341
almost nothing to do with the rise of Russian dialects”. The same can be said of the Pskov and Novgorod dialects. The history of studies of the Pskov-Novgorod cokan’e provides one of the best examples to illustrate it.3
2.
Cokan’e
The term cokan’e is used to refer to the merger of *c (from the 2nd or *kai and the 3rd palatalisation) and *cˇ (of different provenance) into c, that is a pronunciation of (stand.) Rus. pecˇ(ka) ‘stove’, ˇcort ‘devil’, ˇcistyj ‘clean’ etc. as pec(ka), cort, cystyj, where c is of the same quality as in cena ‘prise’ (< *kaina), cerkov’ ‘(Orthodox) church’ (< *kirk-), ovca ‘sheep’ (< *ovika), malec ‘lad’ (with -ec- < *-iki-) etc. The c (< *c and *cˇ) of the old Pskov dialect is supposed to have been previously a soft and “slightly hushing” (šepelevatyj) front-palatal *c” (Karinskij 1916: 237; Orlova 1959: 140) as in some areas of the Northern Russian dialect at present (DARJa, 1986: Maps 45–47). Cokan’e was one of the most characteristic features of the old Pskov (Karinskij 1909: 178) and old Novgorod dialects (Zaliznjak 1995: 34). According to the (DARJa 1986) maps mentioned above, hard cokan’e (that is, a pronunciation of c < *c and *cˇ as (hard) c is common all over the Pskov region nowadays, but see Cˇekmonas (1997). Šaxmatov hesitated in his opinion of what kind of substratum — Lechitic (West Slavic) or Finnic — could be reflected in cokan’e (1915: 317–318). The idea that a Finnic substratum was responsible for the rise of this phenomenon was proposed by Cˇernyšov (1902) and resolutely argued by Selišcˇev (1931), who suggested that all kinds of mixing together of the sibilants and hushing sounds in Slavic languages “everywhere… were provoked by the action of another [non-Slavic. — V.Cˇ.] sound system” (1931: 738). He reconstructed the general prerequisites of the Finnic substratum influence on Russian as follows: Due to a variety of circumstances in economic, social and public life, many of the Finnish groups were assimilated in the Russian milieu and began to speak Russian. They used Finnish along with Russian during a period, and after that it went out of use. The existence of other neighbouring Finnish-speaking groups, the non-intensive influence of the Russian language and a defective command of Russian by those assimilated were reflected in their Russian which had acquired some features of Finnish. (1931: 737)
The substratum nature of cokan’e was suggested by Lytkin (1961: 86) and Kiparsky (1968: 126), who said of cokan’e that “as the whole phenomenon is ultimately of foreign origin, it cannot be connected with any particular Old Russian tribe…” The phenomenon was territorially restricted, and originally possible where there were russified Balto-Finns. Veenker thinks that “the substratum explanation” of the origin of cokan’e is highly probable (“sicher”) (Veenker 1967: 158).
342 Valeriy Cˇekmonas
Avanesov (1947:125) gave a new turn to this idea: “The Slovene and the majority of the Krivichi experienced an important innovation: the sounds ˇc and c were merged to one sound (a soft, sibilant, possibly hissing affricate), that is cokan’e. This process was completed in the context of mixing in a Finnic element, but it has also some foundations in the phonetic system” [of Slavic speech — V.Cˇ.]. First of all, as Avanesov claimed in 1949 (226–227), “The opposition between ˇc and c was semasiologically insignificant” (protivopostavlenie…lišeno …nagruzki), it was “a weak link in the Slavic phonetic system”. In addition, ˇc and c had closely related properties, c being the soft one; they differed only “by the place of their explosion” (ibid.). The supposition that the Finnic substratum could influence the rise of cokan’e was quite resolutely rejected by Orlova (1959: 122), however. She criticised Selišcˇev for using the characteristics of Russian dialects simply to illustrate some general observation without any pertinent analysis of them. On the basis of a large amount of data collected for DARJa, she finds it more appropriate to treat cokan’e as a result of the natural merging together of the two phonetically closely related (as suggested by Avanesov) sounds. This merging, she assumes, has nothing to do with a Finnic substratum (1959: 140). The very phonetic history of the phenomenon — a merge of the “highly palatalised” *c and *cˇ (= *c”, *cˇ”) to *c”, then hardening of *c” > c — testifies to its complicated immanent internal nature, according to Orlova. As to the phonetic value Orlova postulates for the soft c (= *c”) representing cokan’e, it must be added that as early as 1915, Šaxmatov suggested that *c’ was reflected in the ancient Pskov-Novgorod written records (Šaxmatov 1915: 317). It was well known to Karinskij (1909 and especially 1916: 237). Orlova also gives an overview of the main findings about the history of East Slavic — Finno — Ugric ethnic contacts and concludes that there was little likelihood that “the close contacts between the Slavic-speaking and the Finnicspeaking populations would result in bilingualism, the russification of those groups or the rise of cokan’e in their language. Its later spread in a milieu of constantly increasing numbers of primordial native speakers of Russian whose public and economic positions became more and more firmly established was also held to by unlikely. A flattening of the features adopted from the substratum language is usually observed in such situations” (1959: 137). Gluskina agrees with the idea of the Finnic impact on the development of cokan’e in the Russian dialects with the exception of the Pskov and Novgorod dialect (1968: 42). In this dialect, cokan’e developed, she supposes, after the hypothetical ancient peripheral East Slavic varieties which lacked soft *c from the 2nd and 3rd palatalisations (preserving *k in the place of it as in dial. kevina ‘rod, pole’, kefka ‘(textile) bobbin’, kvet ‘flower, blossom’, matika/matica ‘tie-beam’, m’arkat’ = stand. Rus. mercat’ ‘twinkle’) had mixed together with varieties which had both *cˇ and *c (ibid.: 40–41). She observes:
On some Circum-Baltic features of the Pskov-Novgorod dialect 343
Since the 2nd palatalisation had not been completed in the [original — V. Cˇ.] Pskov-Novgorod dialect, it possessed only one affricate ˇc, not two as the other Slavic dialects. The interaction between dialects and the adoption of words with c resulted in confusion of the two affricates. The phonological weakness of the opposition of the affricates in the contributing dialects promoted their blending together. (ibid.: 42)
The origin and history of the Pskov-Novgorod cokan’e need not be explained in another way than those of other Russian dialects, What matters is that all the immanent theories supposing the natural (internal) character of the confusion between *c and *cˇ, including Gluskina’s theory, could not explain why it widely resulted in c-type sounds and not in ˇc- type. We shall return to this problem a little later after making some general remarks on the nature of substratum influence. A substratum — especially in phonetics — is not simply a “borrowing” of a phonetic item or certain qualities. Nor is it a straightforward introduction of these items or qualities into in a variety. Commonly, it is only a stimulus to trigger changes in a variety. The more radical changes in the structure of a language are the more grounds there are to suppose their substratum origin. As to the peculiarities of the changes, they are not usually “a photographic reproduction” of phonetic peculiarities of substratum languages. They can be treated only as a result of the “interpretation” of the sound form of an assimilated language from “the point of view” of an assimilating one. The end effects of the substratum influence started in this way depend on the peculiarities of the interacting languages and on many sociolinguistic factors, in the broad sense of the word, as well. Because of their nonmechanical nature, these effects are unpredictable in many respects. A certain logic can be observed in the modifications of language determined by the logic of the structures of the interacting languages, but the former does not immediately follow from the latter. In the final analysis, the only thing that can be described by means of internal linguistic methods in a narrow sense is the phonetic nature and logic of language change. An inner logic can be discovered in any case because any one linguistic change develops in the structured space of language and is, therefore, inevitably structured. But a description of the logic of the changes does not imply that the changes were generated because of the existence of a logic in them, that is, they are only of immanent or internal nature. The study of the internal history (or inner logic) of language changes is of greater significance and importance for theoretical historical linguistics as well as the study of specific languages. However, the internal history of a language is always incomplete without some explanation as to why a specific change has taken place in a certain area and at a certain time. The answer to this question obviously lies hidden in the history of language contacts. When studying an ethnic history, the hierarchy of things is reversed. In this case, it is of prime importance to take into account all possible areal connections of
344 Valeriy Cˇekmonas
language changes and try to understand in what way they could be connected with the peculiarities of the neighbouring languages. A broader areal perspective of the changes may, and usually does, provide some very important hints to understanding the nature of their sources in the individual languages and help to discover the peculiarities of the ethnic history of the people who speak these languages. Returning to the Russian cokan’e, it has been established that all its variations are characteristic of three large separate areas: the Pskov-Novgorod dialect, the Northeastern part of the Northern Russian dialect and a dialect found in eastern Rjazan’ on both sides of the Oka River (DARJa, map 47). These large areas are isolated, but a number of small and very small islands are distributed between them. The very geographical characteristics of Russian cokan’e suggest that it is widely known in the territory previously inhabited by a Finno-Ugric population that was colonised by East Slavs (Vasmer 1936: 271–344). It is also obvious that cokan’e was developed independently at least in the third (Southeastern) area (which was colonised much later than the Pskov-Novgorod Lands). This could also have been the case in the first two areas with the possibility that cokan’e could have propagated from the main Novgorod area (the Il’men’ region) in all directions as the Novgorodians colonised the present Pskov area and the Russian North. It is a well-known fact that the affricate c is the only one common in the Finnic languages and it is usually restricted to the internal and final positions in words. “In the Southern Estonian dialect, it is also found in the initial position in indigenous words” (Osnovy 1975: 28). The affricate ´c occurs in some Finnic languages as a consequence of the palatalisation of c (ibid.). “The affricate ˇc is found in borrowings and onomatopoeia, yet is palatalised only in the Karelian and Ingrian languages as well as the Onega dialect of Veps. … The hard ˇc of the Votian language is a result of the k > ˇc change before front vowels” (ibid.). In the Veps language, the phonetic distribution of c and ˇc is described by Bubrix as follows: “Cˇ is found in the initial position of the words, and both ˇc and c occur in other positions. C is not presented after i and j…., while ˇc shows up after i and j” (Bubrix 1928: 130). Kettunen suggests that c (ts) occurs in Standard Estonian in a limited number of words, mostly in “descriptive-onomatopoetic ones”, and the very emergence and propagation of it could be a result of Russian influence (Kettunen 1929: 93). It is of great interest to determine whether the Estonian c is a real affricate or may be treated as a ts cluster (ibid.).4 It is obvious that peculiarities of the function and phonetic distribution of c and ˇc in the substratum Finnic languages could stimulate a coalescence of the c- and ˇc -type affricates in Russian dialects into consonants of c-type. When studying how cokan’e arose in the Southeastern Russian dialects with substratum influence from the extinct Merja, Mešcˇora and Muroma languages, some difficulties arise because nothing is known about them. To treat cokan’e as the result of substratum influence here it must be assumed that they had a different system of affricates than the
On some Circum-Baltic features of the Pskov-Novgorod dialect 345
Mordvin dialects had (Bubrix 1953: 15–20; Orlova 1959: 129; Ravila 1973: 306–307) When considering the problem of the areal connections of Pskov-Novgorod cokan’e it makes sense to mention the “Latvian historical cokan’e” or changes *š, *ž, *cˇ > s, z, c; *k, *g > c, dz before front vowels and *kj, *gj > c, dz (Endzelı¯ns 1981 [1938]: 367). Balt *tj, *dj (which are represented as ˇc, dž in Lithuanian) and *stj, *zdj (corresponding to the Lithuanian šcˇ, ž dž) became hushing sibilants š, ž in Latvian. Generally, it looks as if an invisible hand has done everything to prevent the emergence and propagation of the sounds ˇc and dž in this language. Cˇ and dž came into existence here later, mainly from *cj, *dzj, and arose in some Eastern dialects from *k¸, *g´ (ibid.: 366). Cˇ is possible in word initial position in loan-words and onomatopoeia only in Latvian (similar to the distribution of c and ˇc in Estonian). All the historical changes in Latvian mentioned above must have been made after the 4th century AD or so, when the assimilation of the local Finns began. In any case, they are several centuries older than the most ancient East Slavic-Finnic contacts and may therefore be evidence of Finnic influence on the development of the sibilant affricates. Taking all these facts into account, the area of the languages and dialects with “a functionally weak affricate ˇc” embraces the Estonian dialects as well as most of the Latvian and Russian Pskov-Novgorod dialects.5 The main characteristics of the present distribution and historical changes referring to c and ˇc in this area may be presented as follows: Latvian
Estonian
Russian (Pskov-Novgorod)
c and ˇc are distinguished presently as separate phonemes
c and ˇc are distinguished presently as separate phonemes
c and ˇc are distinguished presently as separate phonemes
initial ˇc is found in loan words and onomatopoeia only
ˇc is common in all positions ˇc and c are found in loan in native words words and onomatopoeia only; there is presumed to be an initial c in some native words of the Southern Estonian dialect
palatalisation of *k, *g before the front vowels and of *t, *d before *j resulted into the sibilants c, dz
historically, c and dz (occur in the Southern Estonian dialect in mid-words) are not native consonants; c is suspected to be rather a group of consonants ts
a difference between c and ˇc is a new phenomenon; hard c only (inconsistent hard cokan’e) occurs only in traditional speech of aged persons; cokan’e was common in the past as a result of coalescence of *c’ and *cˇ’ > *c’ (or *c” = soft cokan’e)
346 Valeriy Cˇekmonas
We should emphasise once again that cokan’e, at least in the Pskov dialect, is now a relic feature, being a characteristic of the oldest isolated illiterate individuals. There are several villages along the eastern bank of Lake Cˇudskoe in the Gdov district (such as Ostrovcy, Cˇudskaja Rudnica, Podoleš’e and part of Podborov’e) where cokan’e is commonly heard in everyday speech (especially in the absence of strangers). It is possible to meet individuals with almost consistent cokan’e in their speech everywhere in the Pskov region, but they are extremely rare. We found only four of them during our expeditions in 1992–1996. Because of this, map 47 of (DARJa) showing the all-round cokan’e in most of the Pskov region, refers to the situation at the beginning of the 20th century or even earlier.
3.
Š’okan’e and “s/š -cases”
Clearly the role of substratum influence would be disputable in the case of PskovNovgorod cokan’e if only the facts presented above are considered. The idea can be simply rejected on the premise that a phenomenon like cokan’e does not exist in the neighbouring languages. In other words, a coalescence of soft c and ˇc into a sort of the soft *c followed by a hardening of *c and restoration of the phonological opposition of c–cˇ is not found anywhere except for the Russian dialects. Being unique from this point of view, it can be argued that cokan’e should be studied only as a feature of the history of the Russian language. Once again, a substratum is almost evident, but “empirical evidence is hard to get at”. In this situation it is of great importance to find any other traces of the same substratum and to prove that cokan’e is not the only phenomenon to be developed under foreign influence. The so-called confusion of the sibilants s, z and hushing sounds š, ž should be considered from this point of view first of all. Here we are, in fact, dealing with two phenomena, the first being a merge of the soft *s, *z (of different origin) and soft š, ž into *s”, *z” and the second a free (and sporadic) alternation of the hard *s, *z with (the hard?) *š, *ž. When analysing the peculiarities of the ancient Pskov written recordings, Karinskij noted: “It is obvious that at least some of the scribes did not differentiate between c and ˇc, z and ž, s and š at all” (1909: 178). It follows from the examples in his work that he treated all the cases of such non-differentiation as the same phenomenon. Šaxmatov suggested that the development of the hissing sounds *s”, *z” < *s’, *z’ caused the native *š and *s”, *ž and *z” to merge along with the mixing together of the š and s, ž and z. Both of these innovations have made up “the main difference between the western and eastern [Old Russian — V. Cˇ.] dialects” (1915: 327). The following is one of his theories. In the Eastern Slavic dialects the palatalised *s’, *z’ were found before the front vowels and *’a < *’ (for instance, *sinij ‘blue’, *zeml’a ‘land’, *s’adu ‘I shall sit
On some Circum-Baltic features of the Pskov-Novgorod dialect 347
down’, *z’at’ ‘son-/brother-in-law’); they were changed to *s”, *z” in “the western dialect”. The hushing sounds were “soft” at the time. If they were of *s”, *z” quality, the previous change meant simply a coalescence of the newly arisen *s’’, *z’’ (< *s’, *z’) with the *š, *ž (== *s”, *z”). If *š and *ž were the palatalised *š’, *ž’ or palatal *s´, *z´ (which was possible) they would naturally have merged in the course of time with the new “half-hushing” sound *s”, *z”. It is easy to understand that the change under discussion is parallel to the merge of *c and *cˇ to *c”, that is, to the development of the ancient cokan’e. Compare the following: cokan’e (c”okan’e as a matter of fact): soft *c and soft *cˇ > *c”; s/š -okan’e (s”/z”-okan’e properly): soft *s, *z and soft *š, *ž > *s”, *z”. Filin was the only researcher after Šaxmatov who emphasised this parallelism and criticised Orlova for neglecting it (1972: 267–272). Zaliznjak considers the first phenomenon (cokan’e) to be characteristic of the entire Pskov-Novgorod dialectal zone (1995: 34) while the second (šokan’e in his terms) is more typical of the old Pskov dialect. According to him, šokan’e “occurs in the old Novgorod koine, but it is rare” (ibid.: 43). This confirms the previous observations that we find in Filin (1972: 267–268), Gorškova, Xabugrajev (1981: 64–65) etc. I think the striking parallelism in the development of the cokan’e and šokan’e (actually š’okan’e because of a mixing together of the soft consonants) is sufficient in itself to permit definite conclusions about their common nature to be drawn. Consequently, all that was said about the possible substratum influence on the rise of cokan’e holds for š’okan’e as well. It is not easy, however, to use all the arguments of the immanent hypotheses of cokan’e in reference to š’okan’e, because the oppositions /*s’ – *š *z’ – *ž/ were by no means the “weak links” in the East Slavic phonological system. On the contrary, they differentiated a great many native words. Clearly, we observe a strong tendency to eliminate the hushing sounds, which is not found in the other East Slavic dialects. Filin wrote: We suppose that a coalescence of the sibilants and hushing sounds (or a narrowing-down of the difference of the gap between them) was a result of the further development of the common Slavic heritage that was supported by the foreign influence in the Baltic region in the 8th and 9th centuries. The Baltic linguistic union was probably beginning to emerge at that time. (1972: 272)
This observation may well be correct, but it is remarkable, however, that so much effort has been made to avoid any references to the role of the Finnic substratum in the emergence of that union. Nowadays, a realisation of the soft s, z as s”, z” is common to all the Pskov and, supposedly, Novgorod dialects (DARJa, map 64; corrections and refinements to this
348 Valeriy Cˇekmonas
map are made in Cˇekmonas (1997). Small islands of the soft š, ž (characterised as the “soft” (half-soft) sounds š’, ž’ šs’, žz’ etc.) are scattered throughout the area where cokan’e was recorded (DARJa, map 63). But no areas with consistent šokan’e were found in the Russian dialects, even in the Pskov-Novgorod area. This indicates that, during the last four or five centuries, the ancient Pskov-Novgorod š’okan’e was completely eliminated. The process must have been faster than that of the elimination of cokan’e, because no significant traces of š’okan’e were noted at the beginning of the 20th century. Š’okan’e should be distinguished from the mixing together of the hard s, z with š, ž; these will be called “s/š-cases” for the sake of brevity (and to avoid using the term šokan’e for the present). “S/š-cases” are found in the written records of the Old Pskov (Karinskij 1909). As follows from data in Karinskij’s book, the “s/š-cases” are much more rare than š’okan’e. Šaxmatov’s observations support this fact (1915: 327–328). In more recent times, a 14th-century Pskov manuscript with š’okan’e but without any traces of “s/š-cases” has been described (Kandaurova 1957). According to Zaliznjak (1995: 43), “s/š-cases” are not characteristic of the old Novgorod dialect. A phonetic theory of the “s/š-cases” is lacking. One thing is obvious: š, ž in place of the hard s, z could not be soft (as š, ž from the other source are, see above). Otherwise, a merging of the hard s, z with the soft *s”, *z” (representing the etymological *s’, *z’, *š, *ž, see above) was inevitable and this meant a significant change in the shape of the corresponding words, which is permissible only in the Slavic speech of foreigners. On the other hand, the native hard š, ž were lacking in the Pskov dialect at that time. Consequently, these š, ž of the “s/š-cases” would have to be the new kind of consonants in old Pskovian. The confusion of the hard sibilants s, z and hushing sounds š, ž in the Russian dialects is recorded as being realised by two variants. Sometimes dialectologists note the “normal” s, z instead of the “normal” š, ž; in this case, it is rather a sporadic mixing of these consonants, a kind of free alternation between them with neither being a phonetic change in the proper sense of the word. Sometimes a peculiar articulation of š, ž is reported when they sound intermediate between s, z and š, ž (sš, zž or šs, žz) (Avanesov & Orlova 1989: 67). With respect to the hard š, ž of the Old Pskov dialect (or dialects) in “s/š-cases”, there is a high probability that these newly arisen hard š, ž could have been like sš, zž or šs, žz. The “s/š-cases” are recorded everywhere in the Northern Russian dialect, but because of their sporadic nature no isogloss can be drawn. Single instances of a “lisping” pronunciation of the hard s-z as /ž/najte, /š/yn (=syn), /sš/abaka, navo/zž/ in the Russian dialects are mentioned in Avanesov & Orlova (1965: 86). Filin generally agrees with this statement, adding some more facts to confirm it (Filin 1972: 270). For the Pskov dialect, the situation is quite different. Gluskina, summarising her observations about the “s/š-cases”, suggested that they are common here,
On some Circum-Baltic features of the Pskov-Novgorod dialect 349
especially in the northern part of the Pskov region (Gluskina 1962: 33–35 and the attached map). Moreover, in contrast to the authors mentioned above, she reasoned that “such cases, which were recorded in the various villages in the old Pskov Land, suggest that, in the past, there was no difference between the sibilants and hushing consonants in the Pskov dialect, whether they were soft or hard” (ibid., 34). But let us emphasise that this boldly designed situation is impossible in principle, because the native hard and soft hushing sounds never coexisted in Russian dialects. In our tape recordings made in 1992–1996 in some one hundred villages throughout the Pskov region, “s/š-cases” are observed rarely, but almost everywhere. They are characteristic of the speech of the illiterate (or almost illiterate) persons of the older generation who have always lived in the villages. When asked to repeat the words with irregular s or š, our informants were usually able to pronounce these words “correctly”, that is, as in the standard language. Quite often, “s/š-cases” can be treated as a result of distant dissimilation in current speech, for example: takím ž6 župcˇám (= zupcám) ‘with the same cogs’ (Pecˇ.) a patóm žacéšut’ ‘they will comb’ (Palk.) žapúšcˇana ‘it is neglegted’ (Gd.) psanícy (immediately after is pšanícy, Gd.) ‘wheat:gen.sg’ vžarváls”a (= vzarvals’a, Palk.) ‘it was blown’ ni addáje…žám6š (= zam6š, Seb.) ‘he does not marry (her) off’ sk6barí ža récˇk6j (za recˇkoj, Seb.) ‘the skobars (the Pskovians) are on the other side of the river’ etc. Clearly this factor may stimulate the “s/š-cases”, but it is not the only one as the following examples show: krýska (= krýška, Gd.) ‘lid’ šuval6f ‘Shuvalov (personal name)’ patóm et6t suvál6v utanúl (Psk.) ‘this Suvalov was drowned after that’ žýpki (= zybki) (Pecˇ.) ‘unsteady, shaking’ v maróž6vku ‘to Morozovka (name of the village)’ žagnécca (= zagncca) ‘will turn up-side-down’ ražmácˇ6vat’ (= razmacˇivat’) ‘soak’ kráška (= kraska) ‘paint, dye’ ž6gavar’ú (= zagovor’u) ‘I shall cast a spell over something.’ (all Seb.) žváli (= zvali) (Kun.) ‘we called somebody.’ etc. All research indicates that the “s/š-cases” have been observed more often in the past, but nobody has found the words with stabilised s, z instead of etymological š, ž, and vice versa. Some related place names of the type Sanevo (Pyt.) ~ Šanevo (Porx.), Saxnovo (Bež., Lok., Psk.) ~ Šaxnovo (Porx.), possibly Santalovo (Novsk.) ~ Šantilixa (Bež.)6 may be a result of such stabilisation. Generally, we should agree that the phenomenon under discussion is a relic in the present day Pskov dialect.
350 Valeriy Cˇekmonas
Šaxmatov suggested that a mixing or even coalescence of s, z and š, ž in the western part of the East Slavic dialects was a result of the Lechitic influence on them (1915: 318). He supposed the same about cokan’e as well (see above). But a consistent discrimination between s, z and š, ž in the Belarusian dialects and the comparatively recent origin of the Polish mazurzenie (that is, a non-discrimination of sibilants and hushing sounds) speak against this idea. Little is known about the Lithuanian parallel to the Pskov šokan’e and “s/š-cases”. These were first noted in a small area not far from the Latvian border in northeastern Lithuania (near Joniškis) by Jonaityte˙ (1960). A detailed description of the phenomenon (called “šlekiavimas”) is presented in Girdenis & Pabre˙ža (1978: 127–129). According to these authors, the phonemes /s – š/, /z – ž/, /s’ – š’/ and /z’ – ž’/ are represented in this dialect by the “intermediate hissing” sounds (s, z, s’, z’) (ibid.: 127). “Šlekiavimas” was more widespread in the recent past; it could not be attributed to the Latvian influence. Girdenis & Pabre˙ža suggest that the “hissing s, z” (“with a shade of š, ž”) are similar to the Finnic s and may represent an almost Indo-European archaism, as some scholars have suggested that the “apical-retroflex” /s/ was characteristic of an Indo-European parent language. But the areal distribution of the phenomenon under discussion rather suggests a local origin. Again, a Finnic influence is possible.7 A well known characteristic of the Finnic languages is that the phonological distinction between /s/ and /š/ is possible only in loan-words (Osnovy 1975: 27–28). The distribution and origin of /š/ differ not only in each language but sometimes in individual dialects of the same language. According to Bubrix, this peculiarity of Finnic may reflect ancient Sami influence on Finnic: The similarity between Samoyed’s intolerance for hushing consonants and the same intolerance of ancient Sami speech is striking (he wrote). The BalticFinns’ elimination of the hushing sounds is a phenomenon of the same kind; it is possible that a part of the Sami-speaking population was incorporated into the Finnic one. (Bubrix 1948: 516)
The lack of the /s/ –/š/ opposition in the Finnic languages is quite an ancient feature. A change from the Baltic *š, *ž (< i.-e. palat. *k, *g) into s, z in Latvian may have been completed due to Finnic influence (cf. Lith. šienas, žiema ‘hay, winter’ ~ Lat. siens, ziema ‘ibid.’ etc.). It seems very natural to suppose that the prerequisites for the mixing of sibilants and hushing sounds were created in the Russian dialects during and after the assimilation of the Finnic-speaking population. This would explain why š’okan’e and “s/š-cases” are observed not only in the Pskov-Novgorod dialect and varieties of Northern Russian dialects, but also in the eastern part of the Central Russian dialect. Finally, “šlekiavimas” in the Northern Lithuanian (Žemaitian) dialects could obviously reflect a Finnic substratum influence although not much is known about the prehistory of this territory.
On some Circum-Baltic features of the Pskov-Novgorod dialect
4. B/p-cases Yet another phenomenon of the Pskov and, supposedly, Novgorod dialects can be considered as circum-Baltic. The case in point is the peculiarities of the realisation of voiced and voiceless consonants. The most extensive material concerning this problem has been presented and exhaustively analysed by Gluskina (1973: 35–51). She distinguishes between motivated and unmotivated changes of voiceless to voiced and vice versa. The motivated changes can be treated as a result of hypercorrective morphophonological alternations. She established two main models of such alternations: (1) a type of sutki ‘day and night’ – sem sud6k ‘seven days and nights’; (2) a type of votka ‘vodka’ – votocˇka ‘ibid.:dim’. The unmotivated changes of voiced to voiceless are found: 1. in a word initial position (búza < puza ‘belly’; puxánka < buxanka ‘loaf of bread’); 2. in between vowels (výbucˇit’ (glazá) < výpucˇit’ ‘open (ones) eyes wide’; zatyxácca < zadyxacca ‘go out, die out’); 3. in a word initial position before the resonants (among them /j/ and /v/) as glubók < klubok ‘ball, clew’; 4. after the resonants (e.g. xlóbnut’ < xlopnut’ ‘flap’); 5. in between the resonants (for example, pram’al’gnút’ < pram’al’knut’ ‘flash, fly by’) and 6. after a resonant before a vowel (as in baldavn’á < baltavn’a ‘chatter’). While discussing the results of her observations, Gluskina mentions Kolesov’s article about an analogical phenomenon in the Northern Russian dialect (1963). Kolesov cites some examples from the Novgorod region, e.g. baket < paket ‘parcel’, sabogi < sapogi ‘high boots’, gl’ucˇ < kl’ucˇ ‘crutch:gen.sg’, glubnika < klubnika ‘straw-berries’, Glavdija < Klavdija ‘personal name’, zergalo < zerkalo ‘mirror’, ocˇudils’a < ocˇutils’a ‘found himself (where?)’, gil’ < kil’ ‘keel’, grugly < krugly ‘round’, glodok < glotok ‘mouthful’, dak < tak ‘so’, tred’ < tret’ ‘one third’ (ibid.: 105) which display alternations in the same positions. It is still not clear to what extent the alternation of voiced-voiceless occurs less often in the Novgorod dialect than in the Pskov dialect. Gluskina disagreed with Kolesov who explained the alternations of voiced and voiceless consonants as having developed naturally during and after the loss of the jers. After the latter change, Kolesov surmised that the opposition of syllables (groups of phonemes) ceased to exist and the opposition of separate phonemes began to develop. The half-voiced and half-voiceless consonants came into existence in the new preconsonantal positions and the half-voiceless appeared in word final position. This was the first step in the evolution of the phonological opposition
351
352 Valeriy Cˇekmonas
of the voiced-voiceless consonants which provoked a confusion of both kinds of consonants. Word initial position in proximity to resonants stimulated a mixing of the native voiced and voiceless consonants. The closeness of the Finnic languages could promote the conservation of some of the results of this confusion (ibid.: 105). The so-called fall of the jers was completed relatively early in all the Slavic languages, but it is obvious that a confusion of voiced and voiceless is a widely occurring phenomenon only in those Russian dialects which were in contact with the Finnic languages. This is a very strong argument in favour of its substratum nature. But by accepting substratum influence, we do not deny the possibility of structural logic or inner regularities in its development. Moreover, we assume that any language influence results in some kind of regular, not chaotic, changes. The peculiarities of this logic depend on the special features of both the substratum and recipient languages. In the case of the Pskov and Northern Russian dialects, it is of great interest to understand, for example, why word initial and word internal position (especially between vowels and resonants) are sensitive to voice-voicelessness; why the proximity of resonants promotes a voiced-voiceless alternation etc. There is another geolinguistic argument corroborating the hypothesis of the substratum influence in this case, namely, a very similar phenomenon that Kiparsky (1968) called “b/p-Fälle” which he observed in Latvian and Lithuanian dialects. He found some three hundred Latvian duplicates with word-initial voiced/voiceless consonants, and thirty-two Lithuanian ones. Brief mention was made of such duplicates by Endzelı¯ns in 1922 who cautiously suggested that they might reflect Livonian substratum influence (see also 1951: 254). It is obvious to Kiparsky that “b/p-Fälle” came into existence in the Baltic languages during the assimilation of the Finnic population: At the beginning, the balticized Finns were, naturally, not able to pronounce the voiced consonants, because they were absent in Proto-Finno-Ugric. Later, after the voiced consonants had been mastered by them, they began to produce hypercorrect forms, pronouncing voiced consonants instead of voiceless ones. (1968: 96)
To support his idea, Kiparsky cited Ravila’s observations about voicing of the initially voiceless consonants of the Russian loan-words in Mordvin, Karelian and Veps (ibid.). Of great interest is the fact that the “b/p-cases” are noted in the Baltic dialects more often in the neighbourhood of resonants. A question arises as to whether there is a typological or areal peculiarity that the voiced and voiceless consonants are so prone to free alternation in these positions. Modern Slavic and Baltic languages and dialects show no tendency to neutralise the phonological opposition of voice-voicelessness. Moreover, this opposition is realised consistently at the beginning of a word and before resonants. But word initial position is “weak” in many respects in the Finnic languages. Most importantly,
On some Circum-Baltic features of the Pskov-Novgorod dialect 353
several characteristic restrictions are observed in the realisation of the consonant phonemes in this position in each of them; i.e. an absence (in native words) of clusters and voiced (“weak”) consonants (Finno-Ugorskie jazyki 1966: 27). Wordinitial voiced consonants are not admissible in Estonian and Finnic at all (ibid.). This peculiarity is obviously native to the Finnic languages, and not developed due to the influence of the Indo-European languages. These facts suggest a Finnic influence, not only on the rise of the “b/p cases” discussed above, but even on the ways they have developed. Geolinguisticically, the Estonian and Ingrian languages constitute the core of a large southern circum-Baltic area with no phonological opposition of voice-voicelessness. There are some voiced-voiceless phoneme pairs with restricted distribution in Livonian (Finno-Ugorskie jazyki 1966: 140–141), Votian (ibid.: 120) and Veps (ibid.: 82–83) which are nearest to the core genetically. The core is flanked on the southwest by the Latvian-Lithuanian area with a great number of the “b/p cases” (presumably more numerous in Latvian than in Lithuanian) and by the Russian dialects on the northeast with the “confusion of the voiced-voiceless consonants”. The voice-voicelessness isogloss clearly marks the area of the southern Circum-Baltic languages. To complete this topic, one should note that all the Pskov “b/p cases” discussed by Gluskina are lexicalized; they are usually cited as variants of the corresponding main entries in “Pskovskij oblastnoj slovar’”. A lexicalisation of that kind in the Northern Russian dialect indicates, as Kolesov suggests, that the process has been completed there. Sporadic alternations of these consonants are often found in the modern Pskov dialect as well. For example, while repeating the word pušýstyj ‘downy, fluffy’ our informant says: a xleb bušýsyj pušýstyj ‘the bread is very fluffy’ (Pust. Vosr.). The form pušýstyj is, obviously, normal for her; in the Pskov dialect, voiced consonants are realised as voiceless in word final position and there are, therefore, no grounds to suppose progressive assimilation in /a xleb bušýstyj/. On the contrary, the final b in xleb attests to the voicing of the next b. Careful transcription of recordings show that the sporadic alternation of the voiced and voiceless consonants occurs more often than could be supposed. It is unreasonable to claim that a phonetic change in the proper sense of the word is completed in such cases as: batóm (= patom) ‘after’ (Vasc.) byjaxalšy (= vyjaxalšy) ‘is gone’ mahaiš (= maxaiš)‘wave:2sg.pres’ (V.Krup.) sód6k (= sót6k) ‘hundredth part:gen.pl’ plóha (= ploxa) ‘badly’ kizél’ (= kisel’) ‘kissel’ (Cˇud.Rudn.) dúš6cˇka (= duž6cˇka) ‘arc’ pakavórka (= pagavorka) ‘proverb’ (Malin.) etc.
354 Valeriy Cˇekmonas
because the same words can also be pronounced “correctly” by the same person. The most difficult problem lies now not in looking for additional examples of the alternation under discussion, but in defining the phonetic reasons for it. One thing is certain — the voiced consonants are much weaker in most of the Pskov varieties than in the central or standard Russian ones. The weakness manifests itself primarily in their shortness (especially at the beginning of a word), which is a result of the comparatively “lazy” stop or wide gap between the organs of speech during their articulation. The opposing voiceless consonants are much “stronger” in this respect. This has been noted for the Northern Russian dialect (Kasatkin 1995). It is of great interest, that Kasatkin claims to prove the existence of a phonological opposition between tense and lax consonants in the Northern Russian dialect, rather than the voiced-voiceless opposition found in other Russian dialects (ibid.; Kasatkin, Kasatkina 1993). He supposes that this tenseness-laxness, or at least a tendency to develop it, reflects a Common Slavic heritage in Northern Russian (1989). As for the Pskov dialect, it seems to be more realistic to suppose that the phonological opposition of voice is realised by means of a quite specific articulation basis which reflects substratum influence.
5.
Conclusion
The study of the history of the Pskov-Novgorod dialect may be considered to be a kind of experimental study of substrata in general. A considerable body of historical, linguistic, and more properly archaeological and historical information is available for the making and testing of realistic hypotheses about the rise and development of these dialects. The main features of the assimilated substratum languages are known as well, and moreover, the process of the assimilation of the Finnic population can be observed nowadays, which makes it possible (in principle) to verify many of the linguistic reconstructions. The history of research on the main Pskov-Novgorod dialect’s phonetic characteristics suggests that internal historical methods cannot reveal a substratum influence on the language. It is quite probable that there is always a weak link in every phonological system and phonetic prerequisites for any phonetic-phonological change in any language. A history of every language can, therefore, be represented as a realisation of its internal potential. But no one internal theory could explain why the changes occur in a certain time and place. So, there is no contradiction between internal (or immanent) and “external” historical linguistics. The history of the Pskov-Novgorod cokan’e, š’okan’e, “s/š-cases” and sporadic non-distinguishing of the voiced and voiceless consonants (“b/p-cases”) is a fertile field to test these ideas. Substratum influence arises or becomes topical when the facts of language history are investigated from an areal perspective. To better understand the nature
On some Circum-Baltic features of the Pskov-Novgorod dialect
of purported language contacts, historical and archaeological data should be taken into account. The cross-linguistic areal isoglosses are of primary importance to show that a period of ethno-linguistic contacts existed in the history of an individual language. There is always the possibility of parallel development in neighbouring languages, of course. But cross-linguistic areal isoglosses are no less important or reliable for studies of language contact than diachronic schemes for internal historical linguistics. A bundle of cross-linguistic areal isoglosses is sufficient evidence of intensive language contacts at some point in time. The main goal of the study of the language or languages sharing such a bundle is to determine the nature and history of these contacts. One terminological problem should be briefly discussed here. The term and notion “cross-linguistic isogloss(es)” are unwieldy. It makes sense, I think, to introduce the term heterogloss(es) as a generic and heterophone(s), -form(s), -lex(es) -pragm(s) as specific names instead. The geographical circum-Baltic area is marked by the famous heterophone known as the “tone (germ. Polytonie) languages isogloss” (Jakobson 1931). One heterophone, although obviously not accidental, is not sufficient to establish the relationship between the corresponding languages as a “Sprachbund” in the “Balkan” sense of the word (cf. Sandfeld 1930).8 Therefore, Jakobson’s statement “Ebenso bilden die Sprachen des Baltikums einen Sprachbund, den die Polytonie kennzeichnet”, made in 1931 (1971: 137), could be considered only as a starting point for further investigation of the relationship between the Circum-Baltic languages. Two models to describe the linguistic situation in the circum-Baltic area have been proposed since then. Décsy divided the Circum-Baltic languages into three Sprachbünde which he called “Viking-Bund”, “Peipus-Bund” and “Rokitno-Bund”. Sami, Finnish and the Continental Scandinavian languages, according to his theory, all belong to the large Viking-Bund, which embraces all the languages of Northwestern Europe. Lithuanian is assigned to the Rokitno-Bund, while four small languages — Latvian, Estonian, Livonian and Votian — constitute the Peipus-Bund (Décsy 1973: 43–87). This division implies a quite different kind of Sprachbund than is usually meant by the Balkan Sprachbund, but we cannot discuss this problem here. In a special investigation devoted to the relationship between the Circum-Baltic languages, Stolz (1991: 106–107) writes that Auf derzeitigen Stand der Dinge ist die Existenz eines Sprachbundes im Baltikum, in der hier gewählten Präzisierung als Kulminationszone einer sprachlichen Konvergenzlandschaft, zwar durchaus plausibel gemacht worden.
It is of interest that Stolz does not tend to specify what languages are or could be the members of the “Sprachbund im Baltikum”. His observations and data show that the two largest languages of the Peipus-Bund — Estonian and Latvian — could
355
356 Valeriy Cˇekmonas
form the core of this union. For the goals of this article, it is important to emphasise that Décsy and Stolz agree in treating the similarities between the Peipus-Bund languages as greater than those found in a real Sprachbund of the Balkan type. The (north)western limits of the Peipus-Bund may not be defined properly. The Lithuanian language, as it is commonly known, is represented by two main dialects — a southern one called Aukštaitian and a northern one called Žemaitian (Samogitian in the Middle Ages) (Zinkevicˇius 1966:13–17). This must be taken into consideration when analysing the cross-linguistic areal relationships. The Žemaitian dialect possesses some peculiarities that clearly distinguish it from the remainder of the Lithuanian dialects and bring it and the Peipus-Bund languages close together. Thus, Žemaitian shares with the Peipus-Bund languages a word-initial accent (a mobile accent is characteristic of the rest of the Lithuanian dialects). A kind of Stoßtone (lith. laužtine˙ priegaide˙) is peculiar to the Žemaitian area (see, for example, (Zinkevicˇius 1966: 515, Map 70; LKA 1982: Map 20). It is the largest area with this feature which is also characteristic of Livonian, some of the Latvian dialects and some Estonian varieties. Apocope is a feature of Žemaitian as well as of all the languages of the Peipus-Bund. It is my hope that this list will be continued further. The Russian data that have been analysed in this article suggest that, from an areal point of view, the Russian Pskov-Novgorod dialect is obviously in the sphere of influence of the Peipus-Bund languages. In other words, the Peipus-Bund language area is not confined by Estonian in the Northeast, but it includes some of the Russian dialects, which makes it more similar to the Balkan language union than was thought. The Žemaitian-Pskov-Novgorod heterophones seem to generally support this statement and their very existence reveals their substratum nature.
Notes 1. “Gegenwärtig lässt sich aber schon soviel sagen: Burgsiedlungen wie Pskov Ladoga, Izborsk, “Kamno” und “Rjurikovo gorodišcˇe” am Il’mensee weisen slavische Fundkomplexe seit dem 8.Jh. (eher erst seit der Mitte des 8.Jh.) auf…” [ibid]. This traditional dating is revised by Sedov (1994: 9) who admits the first wave of the East Slavs reached this area in the 5th–7th centuries AD. 2. The Russian name for the Velikaja River is a translation of the Finnic Isa ‘Great’; nowadays the Isa is considered to be a tributary of the Velikaja. 3. For a valuable overview (with an extensive bibliography) of the main findings of the study of the historical phonetics of the old Pskov dialect, reflected in the written records, see Galinskaja (1997). 4. This short overview gives the main facts needed for the purposes of the article without going into details. 5. We will disregard problems associated with this area of the Old Prussian language and Polish dialects with so-called mazurzenie. These are special problems that need to be considered separately. The southern Lithuanian dialect’s dzûkavimas may present a parallel to the Pskov cokan’e if c’ dz’ of these dialects are treated as having developed from ˇc, dÏ of the Aukštaitian Lithuanian
On some Circum-Baltic features of the Pskov-Novgorod dialect 357
dialects (which go back to *tj, *dj before non-front vowels). The possibility exists, however, that these c’ dz’ developed directly from *tj, *dj, see Zinkevièius (1980: 135); Girdenis (1979–1980). 6. The full names of the Pskov region’s districts our examples originated from are as follows: Bež(anickij), Gd(ovskij), Kun(jinskij), Lok(n’anskij), Novsk (Novosokol’skij), Palk(inskij), Pecˇ(kovskij), Porx(ovskij), Psk(ovskij), Pyt(alovskij), Seb(ežskij). 7. After Girdenis became acquainted with the Pskov dialect evidence he agreed with this idea. 8. In spite of the fact that Jakobson developed the idea of a Eurasian language union created by a single very complicated heterophone.
References Avanesov, R. I. 1947. Voprosy obrazovanija russkogo jazyka v ego govorax. Vestnik MGU 9: 109–158. Avanesov, R. I. 1949. Ocˇerki dialektologii rjazanskoj mešcˇery. Materialy i issledovanija po russkoj dialektologii I: 135–236. Avanesov, R. I., Orlova, V. G. (eds.). 1965. Russkaja dialektologija. Moskva: Nauka. Avanesov, R. I., Orlova, V. G. (eds.). 1989. Russkaja dialektologija. Moskva: Nauka. Beleckij, S. V. 1994. Nacˇalo Pskova. Sankt-Peterburg: Gosudarstvennyj Ermitaž. Bjørnflaten, Jan Ivar. 1995. Prehistory and formation of East Slavic: The case of Krivicˇi. In: Ambrosiani, Per, Nilsson, Barbro & Steensland, Lars (eds.), Podobaetu˘ pame˛tı˘ sotvoriti. Essays to the memory of Anders Sjöberg: 39–49. Stockholm: Almqvist & Wiksell International. Bjørnflaten, Jan Ivar. 1998. Proto-East Slavic: fact or fiction? Archaeological aspects of a linguistic problem. In: Bjørnflaten, Jan Ivar, Kjetsaa, Geir & Mathiassen, Terje (eds.), A centenary of Slavic studies in Norway: The Olaf Broch Symposium, 23–45. Oslo: The Norwegian Acad. of Science and Letters. Bubrix, D. V. 1928. Svistjašcˇie i šipjašcˇie soglasnye v Karel’skix dialektax. Ucˇenye zapiski LGU, Serija vostokovedcˇeskix nauk, vyp. 2, Sovetskoe finnougrovedenie: 129–159. Bubrix, D. V. 1948. K voprosu ob otnošenijax meždu samoedskimi i finno-ugorskimi jazykami. Izv. ORJaS AN SSSR, t. VII, vyp. 6: 512–517. Bubrix, D. V. 1953. Istoricˇeskaja grammatika erzjanskogo jazyka. Mordovskoe kniznoe izd. Cˇekmonas, Valeriy. 1977. Osobennosti realizacii soglasnyx /c/ i /cˇ/ v govorax Pskovskoj oblasti. In: Bjørnflaten, Jan Ivar (ed.), Pskovskie govory: Istorija i dialektologija russkogo jazyka, 182–208. Oslo: Solum Forlag. Cˇekmonas, Valeriy. 1997. Mjagkie “šepeljavye” soglasnye (s’’) i (z’’) v govorax Pskovskoj oblasti (lingvogeografija i osnovnye artikuljacionno-akustiheskie osobennosti). Russian Linguistics 21, 247–270. Cˇernyšov, V. 1902. Kak proizošla mena c i ˇc v russkix govorax? Russkij Filologicˇeskij Vestnik 47: 117–118. Dahl, Östen, Koptjevskaja-Tamm, Maria. 1992. Language typology around the Baltic Sea: a problem inventory. Stockholm: University of Stockholm, Department of Linguistics. Décsy, G. 1973. Die linguistische Struktur Europas. Vergagngenheit — Gegenwart — Zukunft. Wiesbaden: Harrassowitz. Dialektologicˇeskij Atlas Russkogo Jazyka (DARJa). 1986. Vypusk 1. Fonetika. Moskva. Endzelı¯ns, Ja¯nis. 1938 [1981]. Latvišu valodas skan¸as un formas. In: Ja¯nis Endzelı¯ns, Darbu izlase, IV. 1. Dal¸a. Rı¯ga¯: Zinatne.
358 Valeriy Cˇekmonas
Filin, F. F. 1972. Proisxoždenie russkogo, ukrainskogo i belorusskogo jazykov. Moskva-Leningrad: Nauka. Finno-ugorskie jazyki. 1966 Jazyki narodov SSSR. Tom III. Finno-ugorskie i samodijskie jazyki. 9–362. Moskva: Nauka. Galinskaja E. M. 1997. Iz istoricˇeskoj fonetiki pskovskix govorov. Voprosy russkogo jazykoznanija. Vypusk VII. Russkie dialekty: istorija i sovremennost’: 67–119. Moskva: Nauka. Girdenis, Aleksas. 1979–1980. Žemaicˇiu˛ dzu¯kavimas. Baltistica 15: 111–123; 16: 32–44. Girdenis, Aleksas, Pabre˙ža, Juozas. 1978. Nauji šlekiavimo stebe˙jimai Žagare˙s apylinke˙se. Baltistica 14: 127–129. Gluskina, S. M. 1962. Morfonologicˇeskie nabljudenija nad zvukom ˇc v pskovskix govorax. Pskovskie govory I: 28–57. Pskov. Gluskina, S. M. 1968. O vtoroj palatalizacii zadnejazycˇnyx soglasnyx v russkom jazyke (na materiale severo-zapadnyx govorov). Pskovskie govory II: 20–44. Pskov. Gluskina, S. M. 1970. Mena zvonkix i gluxix soglasnyx v Pskovskix govorax. Pskovskie govory III: 35–51. Pskov. Goerhke, Carsten. 1992. Frühzeit des Ostslaventums. Darmstadt: Wissenschaftliche Buchgesellschaft. Gorškova, K. V., Xaburgajev, G. A. 1981. Istoricˇeskaja grammatika russkogo jazyka. Moskva: Nauka. Jakobson, R. 1931. Über die phonologischen Sprachbünde. Reprinted in Jakobson R. 1971. Selected Writings. Vol. 1: 137–143. The Hague: Mouton. Jonaityte˙, Aldona. 1960. Nauji duomenys apie Skaistgirio tarme˙s konsonsntizma˛. Lietuviu˛ kalbotyros klausimai III: 79–86. Kandaurova, T. N. 1957. K istorii drevnepskovskogo dialekta. Trudy Instituta jazykoznanija AN SSSR, t. VIII: 178–286. Karinskij, Nikolaj Mixajlovicˇ. 1909. Jazyk Pskova i jego oblasti v XV veke. Zapiski Istorikofilologiceskago fakulteta Imperatorskago S.-Peterburgskago universiteta 93. Sankt-Peterburg. Karinskij, Nikolaj Mixajlovicˇ. 1916. Issledovanie jazyka pskovskogo Šestodneva 1937 g. Žurnal Ministerstva Narodnogo Prosvešcˇenija (Novaja serija) LXI, N 2: 199–251. Kasatkin, L. L. 1989. Odna i z tendencij razvitija fonetiki russkogo jazyka. Voprosy jazykoznanija 6: 39–45. Kasatkin, L. L. 1995. Nekotorye foneticˇeskie izmenenija v konsonantnyx socˇetanijax v russkom, drevnerusskom i praslavjanskom jazykax, svjazannye s protivopostavleniem soglasnyx po naprjažennosti/nenaprjažennosti. Voprosy jazykoznanija 2: 43–56. Kasatkin, L. L., Kasatkina, R. F. 1993. Protivopostavlenie soglasnyx po naprjažennosti/ nenaprjažennosti v severnorusskix govorax. In Rusistika segodnja. Funkcionirovanie jazyka: leksika i grammatika, 123–137. Moskva: Nauka. Kettunen, Lauri. 1929. Eestin kielen äännehistoria. Suomalaisen kirjallisuuden seuran toimituksia, 156. Helsinki: SKS. Kiparsky, Valentin. 1968. Slavische und baltische b/p-Fälle. Scando-Slavica 14: 73–97. Kiparsky, Valentin. 1979. Russian Historical Grammar. Vol. 1. The development of the sound system. Ann Arbor: Ardis. Kolesov, V. V. 1963. Severnovelikorusskie cˇeredovanija soglasnyx, parnyx po gluxosti-zvonkosti. Vestnik LGU 2, Serija jazyka i literatury 1: 103–112. Leningrad. Krys’ko, V. B. Pis’mena i kosti: k poiskam istokov drevnego novgorodsko-pskovskogo dialekta. In Pskovskie govory. Istorija i dialektologija russkogo jazyka. Pod red. Jana Ivara Bjornflatena, 110–132. Oslo. Lietuviu˛ kalbos atlasas (LKA). 1982. Vilnius: Mintis.
On some Circum-Baltic features of the Pskov-Novgorod dialect 359
Lytkin, V. I. 1961. Review of: Orlova V. G., Istorija affrikat v russkom jazyke v svjazi s obrazovaniem russkix govorov. Izvestija AN SSSR 20: 79–86. Nikolajev, S. L. 1990. K istorii plemennogo dialekta krivicˇej. Sovetskoe slavjanovedenie, 4: 54–63. Nikolajev, S. L. 1994. Rannee dialektnoe cˇlenenie i vnešnie svjazi vostocˇnoslavjanskix dialektov. Voprosy jazykoznanija 3: 3–16. Orlova, V. G. 1959. Istorija affrikat v russkom jazyke v svjazi s obrazovaniem russkix narodnyx govorov. Moskva: Nauka. Osnovy finno-ugorskogo jazykoznanija. Pribaltijsko-finskie, saamskie i mordovskie jazyki. 1975. Moskva: Nauka. Ravila, P. 1932. Zum Einfluss des Russischen auf das Mordwinische. Annales Academiae Scientarum Fennicae 27: 256–262. Ravila, P. 1973. Zur Kritik der mordwinischen Substrathypothese der russischen Akan’e. In Commentationes Fenno-Ugricae in honorem Erkki Itkonen, 296–310. Helsinki. Russkaja dialektologija. 1965. Moskva: Nauka. Šaxmatov, A. A. 1915. Ocˇerk drevnejšego perioda istorii russkogo jazyka. Petrograd, (Ènciklopedija slav’janskoj filologii. Vyp. 11. 1). Šaxmatov, A.A. 1915. Ocˇerk drevnejšego perioda istorii russkogo jazyka. Enciklopedija slavjanskoj filologii. Pod redakciej akad. I. V. Jagicˇa. Vypusk 11.1. Petrograd. Sedov, V. V. 1970. Novgorodskie sopki. Moskva: Nauka. Sedov, V. V. 1970a. Slavjane Verxnego Podneprov’ja i Podvin’ja. Moskva, Nauka. Sedov, V. V. 1982. Vostocˇnye slavjane v VI–XIII vv. Moskva: Nauka. Sedov, Valentin Vasil’evicˇ (ed.). 1987. Finno-ugry i balty v èpoxu srednevekov’ja. Moskva: Nauka. Sedov, V. V. 1994. Vostocˇnoslavjanskaja ètno-jazykovaja obšcˇnost’. Voprosy jazykoznanija 4: 3–17. Selišcˇev, A. M. 1931. Sokan’e i šokan’e v slavjanskix jazykax. Slavia 10: 718–741. Shevelov, George Y. 1964. A prehistory of Slavic: the historical phonology of common Slavic. Heidelberg: Carl Winter. Stolz, T. 1991. Sprachbund im Baltikum? Estnisch und Lettisch im Zentrum einer sprachlichen Konvergenzlandschaft. Bochum — Essener Beiträge zur Sprachwandelforschung. Band XIII. Bochum. Straxov, A. B. 1994. Kriticˇeskie zametki po povodu nekotoryx cˇert ‘krivicˇskogo’ dialektnogo cˇlenenija v interpretacii S. L. Nikolajeva. Paleoslavenica II, 249–311. Stroganova T.Ju. 1970. Zapadnye srednerusskie govory. Obrazovanie severnorusskogo narecˇija i srednerusskix govorov (po materialam lingvisticˇeskoj geografii). Avtorskij kollektiv: K. F. Zaxarova, V. G. Orlova, A. I. Sologub, T. Ju. Stroganova: 392–452. Moskva: Nauka. Toivonen, Y. H. 1927. Zur Geschichte der finnisch-ugrischen inlautenden Affrikaten. FinnischUgrische Forschungen 19: 1–270. Vasmer, Max. 1936. Die ehemalige Ausbreitung der Westfinnen in den heutigen slavischen Ländern. Beiträge zur historischen Völkerkunde Osteuropas. Sitzungsber. d. Preuß. Akad. d. Wiss., Philos.- hist. Klasse, 351–440. Veenker, Wolfgang. 1967. Die Frage des finnougrischen Substrats in der russischen Sprache. Bloomington: Indiana University. Zaliznjak, A. A. 1995. Drevnenovgorodskij dialekt. Moskva. Škola “Jazyki russkoj kultury”. Zinkevicˇius, Zigmas. 1966. Lietuviu˛ dialektologija. Vilnius: Mintis.
AUTHOR ""
TITLE "Name index"
SUBJECT "Studies in Language Companion Series, Volume 54"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Name index
Note: page numbers 361–761 refer to Volume 2. A Aalto, Pentti 701 Aavik, Johannes 202, 445, 455, 625 Abondolo, Daniel 179 Ågren, Per-Uno 164, 165 Ahlbäck, Olav 143, 148–151, 163, 166 Aitchison, Jean 454, 456 Aleksandravicˇius, Egidijus 82, 87 Alho, Irja H. 531, 533, 656, 660 Almqvist, Ingrid 408, 655 Alvre, Paul 685, 686, 717 Ambrazas, Vytautas 376, 378, 379, 391, 408–410, 484, 544, 663, 664, 666, 672, 719–726 Anderson, A. O. 585 Antonsen, Elmer H. 222, 223, 225 Anward, Jan 555 Ariste, Paul 201, 203, 302, 304, 305, 373, 381, 575, 576, 632 Arndt, W. W. 220, 222 Arumaa, P. 423, 425, 720, 725 Asher, R. E. 587 Auer, Peter 322 Avanesov, R. I. 342, 348 B Bacot, J. 589 Bakker, Peter 313, 332 Balin´ski, M. 82 Balode, Laimute˙ 642, 645, 657, 676 Barannikov, A. P. 316, 331 Barotov, M. A. 331 Barwise, Jon 528 Basanavicˇius, Jonas 712 Behaghel, Otto 527, 528, 554
Berg-Olsen, Sturla 657, 662 Bergman, Gösta 162, 164 Bergsland, Knut 704 Beskrovnyj, V. M. 316 Bhaskararao, P. 587, 588 Bhatt, Rakesh M. 322 Bhattacharya S. 588 Bielenstein, A. 16 Biezais, Haralds 623 Bisang, Walter 531, 562 Bjørnflaten, Jan Ivar 339 Boiko, Kersti 675 Bokamba, Eyamba G. 325 Boretzky, Norbert 315, 316, 323, 332 Borin, Lars 331, 701 Borkovskij, Viktor I. 384, 396 Bouzet, J. 584 Braun, Maximilian 634, 637 Braunmüller, Kurt 624 Brooks, M. Z. 579 Brugmann, Karl 409, 527, 545, 606, 667, 712 Bubrix, D. V. 344, 345, 350 Burrow, T. 588 Bušs, Oja¯rs 424 Bybee, Joan 419, 445, 453, 629, 719, 726 C Campbell, Lyle 408, 410, 627, 629, 630 Cˇekmonas, Valeriy 341, 348, 625, 626, 643, 681, 685, 688, 694, 709, 732 Chafe, Wallace 720 Christen, Simon 533, 544, 628, 693, 698, 709, 712, 713, 725 Claudi, Ulrike 445, 609
i 2 Name index
Collinder, Björn 238, 241, 244, 607 Comrie, Bernard 193, 584, 671 Cooper, Robin 528 Corbett, Greville 543, 699–701, 703, 704 Courthiade, Marcel 313, 331 Croft, William 531, 562 Csató, Eva 276, 624, 626, 714 D Dahl, Östen 340, 419, 443, 487, 622, 628, 629, 633, 655, 659, 669, 699, 700, 716, 719, 726 Dahlstedt, Karl-Hampus 164, 165 Danell, Gideon 147, 153–158, 160, 164 de Rijk, R. P. G. 670 de Sivers, Fanny 485 Décsy, Gyula 192, 355, 356, 627 Delbrück, Berthold 527, 528, 545, 606, 667 Delsing, Lars-Olof 137, 144, 159, 163, 548, 554, 713 Denison, Norman 536 Dini, Pietro U. 620 Dirr, A. 591 Disterheft, Dorothy 396, 398 Dottin, G. 586 Dressler, Wolfgang 333, 334 Driem, G. van 590 Dryer, Matthew 710–713, 715, 718 Dunn, J. A. 409, 666 E Ebert, Karen H. 731 Ehala, Martin 714, 715 Ejskaer, Inger 145 El’sberg, I.Ja. 111, 114 Elert, Claes-Christian 166, 168, 220, 232, 622 Eloeva, Fatima A. 313, 332 Endzelı¯ns (Endselin), Ja¯nis 345, 352, 369, 371, 397, 408, 409, 414–416, 423, 481, 488, 489, 518, 620, 630, 631, 686, 703 Erelt, Mati 381, 541 Ericsson, Torsten 168 Eschenbach, Carola 532, 550, 551
F Fairbanks, Gordon H. 530 Falkenhahn, V. 626 Fennell, J. 581 Fici Giusti, Francesca 691 Filin, F. F. 347, 348 Forsberg, Hannele 193 Fraenkel, Ernst 403, 577, 582, 592, 608 Frajzyngier, Zygmunt 365 Friedman, Viktor A. 315 Fries, Sigurd 147 Fromm, Hans 201, 574, 575 G Gårding, Eva 166, 167 Ga¯ters, Alfreds 409, 605 Geniušiene˙, Emma 695 Gimbutas, Marija 410, 621 Girdenis, Aleksas 52, 350, 417, 435 Givón, Talmy 573, 678, 683 Gluskina, S. M. 113, 340, 342, 343, 348, 349, 351, 353 Goldberg, B. A. 84 Gomonov, I. T. 120 Gorškova, K. V. 347 Göseken, Heinrich 456 Grappin, H. 578, 579 Greenberg, Joseph H. 276, 701, 704 Greene, D. 585 Grierson, G. A. 589, 590 Grünthal, Riho 179 Gun, Otton 91, 229 Gustavson, Herbert 158, 159, 162, 164, 165 H Haarmann, Harald 627, 719, 720 Hahmo, Sirkka-Liisa 198 Häkkinen, Kaisa 179 Hakulinen, Auli 661 Hakulinen, Lauri 372, 373, 382, 384, 385, 536 Hallberg, Göran 142 Halling, Tiina 198 Hansegård, Nils Erik 238 Harris, Alice C. 408, 410
Name index i 3
Haspelmath, Martin 409, 633, 673, 674, 686, 690, 694, 695, 716 Hasselblatt, Cornelius 453–455, 486, 495, 613 Hausenberg, Anu-Reet 202 Hauzenberga-Šturma, E. 417 Havránek, Bohuslav 408, 409 Heine, Bernd 445, 609, 681, 686, 687 Heinrici Chronicon 7, 623 Helasvuo, Marja-Liisa 196 Helimski, Eugene 184 Heltoft, Lars 549 Hentschel, Elke 550 Herslund, Michael 527, 555 Hertzen, Erik von 240 Hesselman, Bengt 139, 141, 143, 147, 159, 228, 229 Hinderling, Robert 624 Hint, Mati 187, 195, 470, 715 Hirt, H. 413 Hock, Hans H. 322 Hockett, Charles F. 699–701 Hoeksema, Jacob 528 Holm, Gösta 148, 155, 156 Holvoet, Axel 374, 383, 391, 392, 397, 410, 642, 643, 645, 657, 658, 664, 666, 673, 674, 676, 678, 689, 690, 692, 714, 715 Honti, László 188, 636, 704 Hoop, Helen de 528 Hopper, Paul J. 531, 652, 669 Huldén, Lars 162, 699 Hummelstedt, Eskil 159 Hünnemeyer, Frederike 445, 609 Hupel, August Wilhelm 456 Hurford, James 562, 704 Huss, Leena 238 Hyenstrand, Åke 218, 219, 232 Hyldgaard-Jensen, Karl 623 I Igla, Birgit 315, 323, 332 Ingo, Rune 640 Itkonen, Erkki 558, 559 Itkonen, Terho 184, 202–204, 242
Ivars, Ann-Marie 148, 166 Iwaniec, E. 102, 103 J Jablonskis, Jonas 408, 409, 666 Jacobs, Neil G. 624, 626, 699, 717, 732 Jacobsson, Gunnar 409, 410 Jakobson, Roman 355, 434, 626, 627, 641, 644 Janhunen, Juha 180, 182, 184, 186 Jansson, Valter 143, 148 Jaunius, Kazimieras 251 Jeffers, Robert J. 409 Johansen, Paul 413 Johanson Lars 272, 277, 278 Johnson, Mark 595, 596, 683, 684 Jokinen, Kristiina 506, 512–514, 575 Jonaityte˙, Aldona 350 Jörgensen, Nils 144–147, 155, 159, 161, 162–165 Jurginis, J. 82 K Kacˇiuškiene˙, Genovaite˙ 417, 435 Kagaine, Elga 416, 423, 424 Kalima, Jalo 237–239, 250 Kalinina, Elena 590 Kandaurova, T. N. 348 Kangasmaa-Minn, Eeva 195 Kangere, Baiba 408, 657, 672, 675 Kapterev, N. F. 101 Karinskij, Nikolaj Mixajlovicˇ 341, 342, 346, 348 Karjalainen, Merja 368, 614, 656 Karlsson, Fred 248, 607, 655, 661 Karma, Tõnu 676 Karsten, Torsten Evert 238 Karttunen, Lauri 382, 384, 385 Kasatkin, L. L. 354 Kasatkina, R. F. 354 Kasik, Reet 198, 454 Kask, Arnold 453 Kaufman, Terrence 87, 305, 313, 320, 332, 334, 410, 454, 676 Kauppinen, Anneli 192
i 4 Name index
Keenan, Edward L. 363, 380 Kemmer, Susanne 695 Kenrick, Donald 332 Kettunen, Lauri 21, 249, 250, 344, 385, 413 Kibrik, Aleksandr 590, 591, 634 King, Gareth 229, 286, 560, 670 Kiparsky, Valentin 304, 341, 352, 391, 408–410, 642, 666 Klaas, Birute 443, 720, 722 Koivisto, Vesa 195, 200 Koivulehto, Jorma 201, 203, 238 Kolesov, V. V. 351, 353 Kont, Karl 443, 564 Koponen, Eino 201 Koptjevskaja-Tamm, Maria 340, 408, 503, 508, 509, 512, 514, 516, 613 Koreckij, P. I. 102 Koreva, A. 83 Korhonen, Mikko 180, 183, 191, 194, 195, 241 Kornfilt, Jaklin 560, 561, 668 Koshal, S. 589 Krause, Wolfgang 223–225 Krull, Diana 574, 575, 645 Kruuse, E. 168 Kuhn, Hans 222 Kulakauskas, Antanas 82, 87, 88 Kulonen, Ulla-Maija 195 Kuršaitis, Aleksandras 422 Kuz’mina, Irina Borisovna 117, 119, 120, 122, 125, 659, 663, 666, 674, 694, 717 Kytömäki, Leena 199 L Laakso, Johanna 534, 622, 632, 673, 696, 714 Laalo, Klaus 186 Laanest, Arvo 179, 184, 185, 187, 192, 198, 250, 603 Labov, William 332, 333, 454 Lagman, Edvin 705 Lagman, Herbert 153, 155, 156, 163 Laitinen, Lea 670, 673 Lakoff, George 595, 596, 683, 684
Lalou, M. 589 LaPolla, Randy J. 670 Larin, Boris A. 398, 408–410 Larsson, Lars-Gunnar 160 Larsson, Seth 237, 239, 240, 242, 246, 247, 536, 564, 622, 650, 658, 668, 678 Lehiste, Ilse 187, 574, 576, 626, 645, 646 Lehmann, Winfried P. 712 Lehtinen, M. 192, 194, 199, 575 Leino, Pentti 534, 561 Leinonen, Maria 663 Lënngren, Tamara 106, 108, 109, 114, 115, 116, 118, 120–123, 125 Levanda, L. O. 84, 85 Levander, Lars 137, 143, 159 Levin, Beth 670 Levinson, Stephen 164 Lindqvist, Natan 141, 146 Lindström, Eva 701 Linell, Per 555 Loman, Bengt 148 Loorits, Oskar 421, 500 Lyons, Christopher 514 Lytkin, V. I. 341 M Mackinnon, R. 585 Mägiste, Julius 237, 445 Makaev, Ènver A. 222, 223 Maling, Joan 670, 674 Manuš, Leksa 325, 636 Markelov, G. V. 105 Markianova, L. F. 198 Markova, N. V. 694 Marold, Edith 222 Mathiassen, Terje 247, 291, 294, 295, 613, 627, 679, 680 Matras, Yaron 322, 720, 726 Matthews, W. K. 363 Meckelein, R. 578, 579 Meillet, A. 592 Merkys, V. 88 Messing, Gordon M. 331 Metslang, Helle 179, 487, 629, 656 Mikkola, Joz 632
Name index i 5
Milovidov, V. 104 Misius, Kazys 87 Mühlenbachs, Karlis 409, 488 Mullonen, Marija Ivanovna 381, 382, 501 Murav’ev, Mixail 82 M’urkxejn, Vera 125, 129 Murnikova, T. F. 103, 105, 106, 108, 124, 129 N Nau, Nicole 607, 613, 614, 628, 629, 630, 631, 672, 679, 683, 701 Nemceva, L. I. 111, 114, 128 Nemvalts, Peep 486, 492, 661, 662 Nepokupnyj, A. P. 430 Nesselmann, G. H. F. 620 Nesser, Anne 238 Nichols, Johanna 580–582, 590, 621, 720, 728, 730 Nielsen, Niels Åge 699 Niilus, V. 727 Nikiforovskij, N.Ja. 123 Nikkilä, Osmo 201 Nikula, Kristina 160 Nilsson, Torbjörn K. 250, 429, 430 Norde, Muriel 548, 713 Noreen, Adolf 138, 215 Novgorodov, M. A. 108, 112–114, 113–117, 120–122, 125 Nyström, Staffan 168 O Oinas, Felix J. 603 Ojanen, Muusa 238 Ojutkangas, Krista 198 Orlova, A. I. 341, 342, 345, 347, 348 Östman, Jan-Ola 628, 716 Ozols, Arturs 409 P Paasonen, Heikki 559 Pabre˙žña, Juozas 350 Pagliuca, William 445 Pajusalu, Karl 207 Palmaitis, Letas 410
Pamp, Bengt 161, 162, 165, 166 Paunonen, Heikki 189, 205 Paus, Charles 544 Payne, John R. 517 Pepicello, William J. 409 Perkins, Revere 445 Pisani, V. 379 Pitkänen, Antti J. 527 Plank, Frans 514, 554, 693 Plöger, Angela 238 Poplack, Shana 325 Postal, Paul 693 Posti, Lauri 182, 184, 204, 238, 243 Potebnja, Aleksandr A. 391, 408, 409 Pozdeeva I. 104 Prince, Ellen F. 614 Proxorova, S. M. 120–122 Prozorov, I. 101, 102 Prugavin, S. 104 Pugh, Stephan 626, 629 Pyöli, Raija 208, 261 Q Qvigstad, Just Knud 238 R Raag, Raimo and Virve 195, 205, 483, 486, 488, 491, 613 Rätsep, Huno 452, 454 Raukko, Jarno 628, 716, 717 Ravila, P. 345, 352 Reinhammar, Maj 161, 677 Reiter, Norbert 409 Rendahl, Anne-Charlotte. 624, 648, 677, 693, 699, 705, 713 Rhys Jones, T. J. 586 Ritter, Ralf-Peter 381 Rixter, E. V. 123 Robinson, Orrin W. 216, 220 Roos, Jaan 645 Rothstein, Robert A. 274 Rudzı¯te, Marta 414, 418, 500, 632, 648, 676, 699 Rusakov, Aleksandr 626, 629
i 6 Name index
S Saagpakk, Paul Friidrih 502 Sajantila, Antti et al. 241 Salminen, Tapani 179, 205 Salmons, Joe 643, 644 Saltarelli, M. 583, 584 Salys, A. 52 Sammallahti, Pekka 180, 182, 203 Sankoff, Gillian 325 Sarhimaa, Anneli 208, 627 Savijärvi, Ilkka 193 Sawyer, Peter 228 Šaxmatov, A. A. 341, 342, 346–348, 350 Schagerström, August 142 Schiffman, H. F. 587, 588 Schlachter, Wolfgang 696 Schmalstieg, William R. 377, 409, 608 Schöndorf, K. E. 623 Schuchardt, Hugo 583 Schwartz, Eugène 547, 548 Sedov, Valentin Vasil’evicˇ 339, 340 Sehwers, Johannes 624 Seilenthal, Tõnu 200, 445 Selišcˇev, A. M. 341, 342 Selkirk, E. 528 Semenova, M. F. 81, 85, 90–95, 106, 111–113, 122, 124, 131 Senn, Alfred 515, 577, 598, 608 Seppänen, Aimo 705 Seppänen, Ruth 536, 539, 540 Serebrennikov, B. A. 676 Setälä, E. N. 418, 425, 426 Sgall, Peter 396, 398 Shibatani, Masayoshi 365 Shore, Susanna 194 Short, David 26, 31, 46, 47, 154, 580, 581, 664 Siewierska, Anna 365, 633, 690, 691, 692, 708, 709 Sinica, A. I. 106, 108, 114, 122, 123, 128, 129 Sinor, Denis 179 Sirtautas, Vytautas 409 Sivers, Fanny de 418, 431 Sivickene, M. 109, 114, 125, 126
Sjöberg, Åke G. 158 Skardžius, Pranas 427 Sköld, Tryggve 238, 562 Smyth, Herbert Weir 527, 528 Söderman, Tiina 202, 207 Solove’v, S. M. 101 Sprincˇak, Jakov A. 409 Stang, Christian S. 243, 481, 622, 643 Stassen, Leon 536, 678, 679, 681, 683, 686–688, 698 Stepanov, Jurij S. 391, 409 Stolz, Thomas 355, 356, 494, 627, 630, 631, 632, 675, 679, 684, 685, 714, 717, 720, 726 Stone, Gerald 579, 664 Stroganova, T.Ju 340 Suhonen, Seppo 179, 238, 250 Sulkala, Helena 368, 614, 656 Svonni, Mikael 238 Swenning, Julius 162 Szemerényi, Oswald 249 T Tauli, Valter 368, 374, 384, 603, 604 Ternes, Elmar 649 Thomas, G. 560 Thomason, Sarah G. 305, 313, 320, 332, 334, 410, 454, 566, 627, 675 Thompson, Sandra A. 531, 652, 669 Thomsen, Vilhelm 238, 241, 242, 421 Tiberg, Nils 154–158, 161, 163 Tikka, Toivo 197 Timberlake, Allan 392, 398, 399, 404, 405, 406, 408, 409, 658, 664, 665, 666, 671, 674, 694 Tiselius, G. A. 145, 159 Tkacˇenko, Orest Borisovicˇ 631 Tommola, Hannu 194, 443 Tõnisson, Evald 413, 623 Toomsalu, Epp 455 Toporov, V. N. 622 Tret’jakov, P. N. 622 Trosterud, Trond 191, 194 Trubacˇev, O. N. 622 Tyla, Antanas 87
Name index i 7
U Udolph, Jürgen 220, 221 Ultan, Russel 717 Ureland, P. Sture 624, 696 V Vääri, E. 720 Vaba, Lembit 238, 250, 418, 687 Vähämäki, Börje K. 661 Vahros, I. 666 Vainikka, Anne 669 Valtonen, P. 331 van der Voort, Hein 332 Van Valin, Jr Robert D. 670 Vare, Silvi 200 Vasiliauskiencˇ, Virginija 712 Vasmer, Max 344 Vaxtin, Nikolaj B. 334 Veenker, Wolfgang 340, 341, 410, 454, 576, 592, 642, 658 Ventzel, Tatiana V. 321 Verešcˇagin, E. M. 331 Verkuyl, Henk J. 486, 487 Vesper, D. R. 588 Viitso, Tiit-Rein 204, 644 Vilkuna, Maria 651, 661, 670, 672, 673, 675, 697, 709 Vraciu, Ariton 633 Vuorela, Katri 331, 701 Vysotskij, S. S. 126 W Wälchli, Bernhard 414, 432, 536, 543, 567
Watkins, Calvert 680 Wessén, Elias 139–141, 143, 144, 146, 147, 158, 162, 163, 165, 166, 223, 226, 230, 231, 527, 547, 624 Westergaard, K-E. 623 Widmark, Gun 229, 230 Wiemer, Björn 675 Wierzbicka, Anna 491, 634 Wiik, Kalevi 621, 626, 644 Wiklund, Karl Bernhard 238 Willet, Thomas 719, 720, 726 Winkler, Eberhard 206 Woolford, E. 325 Wordick, F. J. F. 671 X Xejter, X. 108, 113, 124, 125, 129 Xodova, K. J. 396 Xolodovicˇ, A. A. 365 Z Zajceva, Marija Ivanovna 501 Zaliznjak, A. A. 341, 347, 348 Zavarina, A. A. 102–104, 123 Zen’kovskij, Sergej 101 Zeps, Valdis J. 484, 490–492 Žilinskiene˙, Vida 409 Zimmermann, Christiane 222 Zinkevicˇius, Zigmas 52, 86, 88, 243, 356, 410, 481 Zorina, Zoja Georgievna 559 Žuravlev, V. K. 406 Zverkovskaja, N. P. 114, 116
AUTHOR ""
TITLE "Language index"
SUBJECT "Studies in Language Companion Series, Volume 54"
In the STUDIES IN LANGUAGE COMPANION SERIES (SLCS) the following volumes have been published thus far or are scheduled for publication: 1. ABRAHAM, Werner (ed.): Valence, Semantic Case, and Grammatical Relations. Workshop studies prepared for the 12th Conference of Linguistics, Vienna, August 29th to September 3rd, 1977. Amsterdam, 1978. 2. ANWAR, Mohamed Sami: BE and Equational Sentences in Egyptian Colloquial Arabic. Amsterdam, 1979. 3. MALKIEL, Yakov: From Particular to General Linguistics. Selected Essays 1965-1978. With an introd. by the author + indices. Amsterdam, 1983. 4. LLOYD, Albert L.: Anatomy of the Verb: The Gothic Verb as a Model for a Unified Theory of Aspect, Actional Types, and Verbal Velocity. Amsterdam, 1979. 5. HAIMAN, John: Hua: A Papuan Language of the Eastern Highlands of New Guinea. Amsterdam, 1980. 6. VAGO, Robert (ed.): Issues in Vowel Harmony. Proceedings of the CUNY Linguistics Conference on Vowel Harmony (May 14, 1977). Amsterdam, 1980. 7. PARRET, H., J. VERSCHUEREN, M. SBISÀ (eds): Possibilities and Limitations of Pragmatics. Proceedings of the Conference on Pragmatics, Urbino, July 8-14, 1979. Amsterdam, 1981. 8. BARTH, E.M. & J.L. MARTENS (eds): Argumentation: Approaches to Theory Formation. Containing the Contributions to the Groningen Conference on the Theory of Argumentation, Groningen, October 1978. Amsterdam, 1982. 9. LANG, Ewald: The Semantics of Coordination. Amsterdam, 1984.(English transl. by John Pheby from the German orig. edition “Semantik der koordinativen Verknüpfung”, Berlin, 1977.) 10. DRESSLER, Wolfgang U., Willi MAYERTHALER, Oswald PANAGL & Wolfgang U. WURZEL: Leitmotifs in Natural Morphology. Amsterdam, 1987. 11. PANHUIS, Dirk G.J.: The Communicative Perspective in the Sentence: A Study of Latin Word Order. Amsterdam, 1982. 12. PINKSTER, Harm (ed.): Latin Linguistics and Linguistic Theory. Proceedings of the 1st Intern. Coll. on Latin Linguistics, Amsterdam, April 1981. Amsterdam, 1983. 13. REESINK, G.: Structures and their Functions in Usan. Amsterdam, 1987. 14. BENSON, Morton, Evelyn BENSON & Robert ILSON: Lexicographic Description of English. Amsterdam, 1986. 15. JUSTICE, David: The Semantics of Form in Arabic, in the mirror of European languages. Amsterdam, 1987. 16. CONTE, M.E., J.S. PETÖFI, and E. SÖZER (eds): Text and Discourse Connectedness. Amsterdam/Philadelphia, 1989. 17. CALBOLI, Gualtiero (ed.): Subordination and other Topics in Latin. Proceedings of the Third Colloquium on Latin Linguistics, Bologna, 1-5 April 1985. Amsterdam/Philadelphia, 1989. 18. WIERZBICKA, Anna: The Semantics of Grammar. Amsterdam/Philadelphia, 1988. 19. BLUST, Robert A.: Austronesian Root Theory. An Essay on the Limits of Morphology. Amsterdam/Philadelphia, 1988. 20. VERHAAR, John W.M. (ed.): Melanesian Pidgin and Tok Pisin. Proceedings of the First International Conference on Pidgins and Creoles on Melanesia. Amsterdam/Philadelphia, 1990.
21. COLEMAN, Robert (ed.): New Studies in Latin Linguistics. Proceedings of the 4th International Colloquium on Latin Linguistics, Cambridge, April 1987. Amsterdam/ Philadelphia, 1991. 22. McGREGOR, William: A Functional Grammar of Gooniyandi. Amsterdam/Philadelphia, 1990. 23. COMRIE, Bernard and Maria POLINSKY (eds): Causatives and Transitivity. Amsterdam/Philadelphia, 1993. 24. BHAT, D.N.S. The Adjectival Category. Criteria for differentiation and identification. Amsterdam/Philadelphia, 1994. 25. GODDARD, Cliff and Anna WIERZBICKA (eds): Semantics and Lexical Universals. Theory and empirical findings. Amsterdam/Philadelphia, 1994. 26. LIMA, Susan D., Roberta L. CORRIGAN and Gregory K. IVERSON (eds): The Reality of Linguistic Rules. Amsterdam/Philadelphia, 1994. 27. ABRAHAM, Werner, T. GIVÓN and Sandra A. THOMPSON (eds): Discourse Grammar and Typology. Amsterdam/Philadelphia, 1995. 28. HERMAN, József: Linguistic Studies on Latin: Selected papers from the 6th international colloquium on Latin linguistics, Budapest, 2-27 March, 1991. Amsterdam/Philadelphia, 1994. 29. ENGBERG-PEDERSEN, Elisabeth et al. (eds): Content, Expression and Structure. Studies in Danish functional grammar. Amsterdam/Philadelphia, 1996. 30. HUFFMAN, Alan: The Categories of Grammar. French lui and le. Amsterdam/Philadelphia, 1997. 31. WANNER, Leo (ed.): Lexical Functions in Lexicography and Natural Language Processing. Amsterdam/Philadelphia, 1996. 32. FRAJZYNGIER, Zygmunt: Grammaticalization of the Complex Sentence. A case study in Chadic. Amsterdam/Philadelphia, 1996. 33. VELAZQUEZ-CASTILLO, Maura: The Grammar of Possession. Inalienability, incorporation and possessor ascension in Guaraní. Amsterdam/Philadelphia, 1996. 34. HATAV, Galia: The Semantics of Aspect and Modality. Evidence from English and Biblical Hebrew. Amsterdam/Philadelphia, 1997. 35. MATSUMOTO, Yoshiko: Noun-Modifying Constructions in Japanese. A frame semantic approach. Amsterdam/Philadelphia, 1997. 36. KAMIO, Akio (ed.): Directions in Functional Linguistics. Amsterdam/Philadelphia, 1997. 37. HARVEY, Mark and Nicholas REID (eds): Nominal Classification in Aboriginal Australia. Amsterdam/Philadelphia, 1997. 38. HACKING, Jane F.: Coding the Hypothetical. A Comparative Typology of Conditionals in Russian and Macedonian. Amsterdam/Philadelphia, 1998. 39. WANNER, Leo (ed.): Recent Trends in Meaning-Text Theory. Amsterdam/Philadelphia, 1997. 40. BIRNER, Betty and Gregory WARD: Information Status and Noncanonical Word Order in English. Amsterdam/Philadelphia, 1998. 41. DARNELL, Michael, Edith MORAVSCIK, Michael NOONAN, Frederick NEWMEYER and Kathleen WHEATLY (eds): Functionalism and Formalism in Linguistics. Volume I: General papers. Amsterdam/Philadelphia, 1999.
42. DARNELL, Michael, Edith MORAVSCIK, Michael NOONAN, Frederick NEWMEYER and Kathleen WHEATLY (eds): Functionalism and Formalism in Linguistics. Volume II: Case studies. Amsterdam/Philadelphia, 1999. 43. OLBERTZ, Hella, Kees HENGEVELD and Jesús Sánchez GARCÍA (eds): The Structure of the Lexicon in Functional Grammar. Amsterdam/Philadelphia, 1998. 44. HANNAY, Mike and A. Machtelt BOLKESTEIN (eds): Functional Grammar and Verbal Interaction. 1998. 45. COLLINS, Peter and David LEE (eds): The Clause in English. In honour of Rodney Huddleston. 1999. 46. YAMAMOTO, Mutsumi: Animacy and Reference. A cognitive approach to corpus linguistics. 1999. 47. BRINTON, Laurel J. and Minoji AKIMOTO (eds): ollocational and Idiomatic Aspects of Composite Predicates in the History of English. 1999. 48. MANNEY, Linda Joyce: Middle Voice in Modern Greek. Meaning and function of an inflectional category. 2000. 49. BHAT, D.N.S.: The Prominence of Tense, Aspect and Mood. 1999. 50. ABRAHAM, Werner and Leonid KULIKOV (eds): Transitivity, Causativity, and TAM. In honour of Vladimir Nedjalkov. 1999. 51. ZIEGELER, Debra: Hypothetical Modality. Grammaticalisation in an L2 dialect. 2000. 52. TORRES CACOULLOS, Rena: Grammaticization, Synchronic Variation, and Language Contact.A study of Spanish progressive -ndo constructions. 2000. 53. FISCHER, Olga, Anette ROSENBACH and Dieter STEIN (eds.): Pathways of Change. Grammaticalization in English. 2000. 54. DAHL, Östen and Maria KOPTJEVSKAJA-TAMM (eds.): Circum-Baltic Languages. Volume 1: Past and Present. 2001. 55. DAHL, Östen and Maria KOPTJEVSKAJA-TAMM (eds.): Circum-Baltic Languages. Volume 2: Grammar and Typology. 2001. 56. FAARLUND, Jan Terje (ed.): Grammatical Relations in Change. 2001. 57. MEL’C UK, Igor: Communicative Organization in Natural Language. The semanticcommunicative structure of sentences. 2001. 58. MAYLOR, Brian Roger: Lexical Template Morphology. Change of state and the verbal prefixes in German. n.y.p.
