247 108 22MB
English Pages 333 [336] Year 1969
The Field of Yiddish Third Collection
The Field of Yiddish Studies in Language, Folklore, Literature Third Collection Edited by
MARVIN I. HERZOG, WITA RAVID, and URIEL WEINREICH
1969 M O U T O N & CO. LONDON
. THE HAGUE •
PARIS
© Copyright 1969 by Mouton & Co., Publishers, The Hague, The Netherlands.
No part of this book may be translated or reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publishers.
Library of Congress Catalog Card Number 76-99758
Made and printed in Great Britain by William Clowes and Sons, Limited London and Beccles
PREFACE
This is the third collection entitled The Field of Yiddish-1 It is an outgrowth of the Conference on Yiddish Dialectology, jointly sponsored by the YIVO Institute for Jewish Research and the Language and Culture Atlas of Ashkenazic Jewry (LCAAJ), Columbia University, convened in New York in 1965. All but two2 of the papers read at the Conference are presented in this volume, some with major modifications. The articles fall into two large groups: one devoted to Western Yiddish, the other to Eastern Yiddish. To the first group belong the papers by Hutterer, Guggenheim-Grttnberg, Lowenstein, and Zuckerman. Of the seven essays in the second group, two are concerned with dialectal subregions—Herzog's with the Ukraine, and Weinreich's with Belorussia. The others focus on general problems: Wolf concentrates on the case and gender system. Green examines stress variations. Schwartz deals with two food terms. Onamastics is represented by Stankiewicz's outline of a program for a geographic approach to the study of personal names. Trends in the standardization of Yiddish are discussed by Schaechter. The two remaining papers bear on subjects somewhat less central to the field of Yiddish. Both share an interest in population statistics. KestenbergGladstein3 traces Jewish migrations in Bohemia, as reflected in census data. Ravid discusses the utilization of population maps in dialect geography. 1 The first collection The Field of Yiddish: Studies in Yiddish Language, Folklore, and Literature, Published on the Occasion of the Bicentennial of Columbia University (ed. U. Weinreich), New York, appeared in 1954. The second collection The Field of Yiddish: Studies in Language, Folklore, and Literature (ed. U. Weinreich), The Hague, was published in 1965. 2 These two papers appeared elsewhere: Mark, Y. Lithuanian Yiddish: a comparison of the Atlas data with other materials. Noble, S. The differentiation of dialects in a single community. 3 Her article is a somewhat modified version of the Hebrew original presented to the Fourth World Congress of Jewish Studies, Jerusalem, 1965. V
vi
PREFACE
The Bibliography gives full citations of literature referred to in the articles. Names of authors and titles of books and periodicals in languages not using the Latin alphabet have been transliterated, following the conventions currently accepted in linguistic publications. Titles of articles enclosed in square brackets are translations from the original. The publication of this volume was delayed by the untimely death of Uriel Weinreich on March 30, 1967. The fact that it has now appeared is due to the dedicated efforts of Wita Ravid. She joins me in dedicating the book to the memory of Uriel Weinreich, friend, colleague, and teacher, and Max Weinreich, his father and mentor, who passed away on January 28,1969. M.H.
CONTENTS
Preface Table of contents Theoretical and practical problems of Western Yiddish dialectology C. J. Hutterer Endinger Jiddisch Florence Guggenheim-Griinberg Results of Atlas investigations among Jews of Germany Steven Lowenstein Alsace: an outpost of Western Yiddish Richard Zuckerman Yiddish in the Ukraine: isoglosses and historical inferences Marvin I. Herzog The geographic makeup of Belorussian Yiddish Uriel Weinreich The geography of Yiddish case and gender variation Meyer Wolf On accentual variants in the Slavic component of Y i d d i s h . . . Eugene Green The geography of two food terms: a study in Yiddish lexical variation Rosaline B. Schwartz The derivational pattern of Yiddish personal (given) names Edward Stankiewicz The "hidden standard": a study of competing influences standardization Mordkhe Schaechterin The internal migration of Jews in 19th century Bohemia Ruth Kestenberg-Gladstein Introductory maps for the Language and Culture Atlas of Ashkenazic Jewry Wita Ravid The contributors Bibliography
vii
v vii 1 8 16 36 58 82 102 216 240 267 284 305 310 317 318
THEORETICAL A N D OF WESTERN
PRACTICAL
YIDDISH
PROBLEMS
DIALECTOLOGY
C. J. HUTTERER When we speak of research problems of Yiddish dialectology, we must never forget that language is a social phenomenon and that its fate cannot be considered apart from the fate of its creators and users. In particular, this applies to Western Yiddish dialectology. While I do not intend to cast doubt upon the status of linguistics as a self-contained science—on the contrary, I wish to emphasize the fact that in linguistic or dialectological analysis, the approach to all levels and styles of language ought to be purely linguistic—I should like to assert, nevertheless, that linguistic analysis cannot by itself reveal the nature of a given language as an operating human and social construction. To quote Ferdinand Wrede's apt observation: "The history of a language follows the fate of its speakers." In this light, it becomes obvious that in order to organize our practical work rigorously we require a model or, more exactly, a clearinghouse for the theoretical problems which determine both the field work itself and the presentation of the linguistic materials which it yields. Such a model would also suggest the general lines of a possible interpretation, which will often include references to strictly causal relations. This problem is of general theoretical importance and is valid for all languages and dialects. It is all the more important for languages or dialects whose formation and development was affected by such a multiplicity of social and cultural factors as have interacted in the case of Yiddish. It is by reference to such factors that some of the most deep-rooted differences between the EY and the WY areas are revealed. To be sure, Yiddish must be investigated as a whole. Still, some topics stand out so clearly in connection with WY that, they must be explicitly brought to the fore. First of all, the specific social and historical situation of WY has to be mentioned. In contrast to EY which developed in Slavic and other nonGerman environments, WY had never lost contact with German. Consequently, problems of the linguistic development of German as well as of 1* i
2
C. J. HUTTERER
several parts of the German language area assume a much more prominent role in investigations of W Y than in those of E Y. This is true of all linguistic levels in the entire WY area. On the phonological level, for example, let me refer only to the total system of affricates in Alsatian Yiddish; on the morphological level, to the models which differ from EY, but agree with German (WY ix tet/volt fregn ~ EY ix volt gefregt/fregn 'I would ask'); on the syntactic level to German-like word-order (WY ix vil dix a kase fregn ~ EY ix vil dixjdir fregn a kase 'I want to ask you a question'); on the lexical level, to Germanisms or to features characteristic of German dialects (WY kraml~ EY grlvn 'roast chicken skin'). The continuing affinity with German took the form of contact between the WY dialects and the so-called supraregional vernaculars of German. At the same time it must be remembered that WY was also—quite apart from the Haskalah movement—directly influenced by literary German. This circumstance is particularly apparent in the different frequency of Hebrew-Aramaic and German lexical elements common to both main wings of the Yiddish speech territory. All Yiddish speakers are acquainted, for example, with words like milxome 'war,' balegole 'wagonmaster,' on the one hand, and with krig and kutser (same meanings) on the other. However, in spontaneous speech in which no specific stylistic effect was intended, I have never heard the expression far der milxome instead of farem krig from WY informants; and I have rarely heard the word kutser instead of balegole used by an EY informant in reference to a wagonmaster. A further observation is highly significant from a sociolinguistic point of view: while increased eloquence is in EY associated with a rising frequency of the Hebrew-Aramaic component, in WY an analogous stylistic effect is achieved by an increase in the German component. In connection with the question, "What is the word for 'moon' in your dialect?", a very intelligent WY informant gave me the seemingly paradoxical answer that levöne is not a Jewish (jüdisch), but a Hebrew word. The remark reflected his firmly established attitude that to be Jewish (i.e. Yiddish), a word must be of German origin. At this point another question arises namely that of the social status of the WY dialects. Since the age of the Enlightenment, the pace of economic, social, and secular-cultural progress of Western Ashkenazic Jewry had been more rapid than that of East European Jewry. Its historical correlates were secularization, emancipation in various domains, and even assimilation. This was characteristic of Germany and Austria, but the case of Hungary was a good deal more ambiguous. At the end of the 18th and the beginning of the 19th century, within the framework of the Hungarian age of reform, a Jewish reform movement
PROBLEMS OF WESTERN YIDDISH DIALECTOLOGY
3
had been initiated by Emperor Joseph II of Hapsburg in religious and cultural life, the latter including language, of course. Its most outstanding leader, Eduard Einhorn (Horn), waged a relentless campaign against Yiddish. He considered it a corrupted form of German and its replacement by literary German became one of the central aims of his forum, Der ungarische Israelii, the very title of which reflected the trend toward national and linguistic assimilation in the middle of the last century. The connection which we see here between Hungarian national aspirations and German linguistic ambitions may seem strange today, but it contained no contradiction at the time—shortly after 1848—when the German population of the Hungarian towns comprised both the burghers and the intellectuals. They, too, were still German in their speech, but Hungarians in their social and national aspirations (Franz Liszt is a case in point). From there it was but one step to linguistic assimilation, for the Jewish as well as the German townspeople in Hungary. In the second half of the 19th century, Hungarian sermons became increasingly common in the synagogues. Yiddish was gradually losing status. Though still regarded by its speakers as an integral part of a Kultursprache (if only a "corrupted" variant of it), Yiddish was being swept away by the same wave that was eroding the Kultursprache, namely by the movement of Magyarization. As a result of these changes, Yiddish was inevitably reduced to the status of a jargon, a stain on the family coat-of-arms, as it were, for the proreform Jews of Hungary. The Khasam-Soyfer movement, a revival of nonHassidic orthodoxy in northern and western Hungary, pleaded against the abandonment of Yiddish, but to no avail. Its valiant efforts brought symptomatic and temporary relief, but did not stop the spread of destruction. The social and historical formation of WY became decisive for the contacts of WY dialects with the surrounding languages. The area lacks the specific traits of Yiddish-Slavic symbiosis which were so marked in EY (see U. Weinreich 1958a). From a purely linguistic point of view, the contact between WY and German was not unlike that between EY and Slavic. Literary German and the German vernacular of the environment came to be interwoven with the system of máme-losn. But while such interweaving produced a specific form of language synthesis in Slavic surroundings, it led to a step-by-step absorption of German norms by Yiddish in completely German environments. In the transitional area, especially in Hungary, the situation was more complex. When striving after literary German lost its appeal, Magyarization became the focal point of assimilation. What resulted was not a reshaping of Yiddish on Hungarian models, similar to the process that Yiddish underwent on the Slavic scene, but rather a total
4
C. J. HUTTERER
replacement of Yiddish by Hungarian. This process was, however, considerably restrained by the subsequent presence of German as an important lingua franca in the Carpathian region—as a consequence of social and cultural contacts, on the one hand, with the towns of Austria, Bohemia, and Germany, and on the other hand, with the German townspeople of Hungary who were themselves in the process of Magyarization. Thus the situation in which Transcarpathian WY (declared to be a jargon at the time) found itself was quite different from the situation of EY, and at the same time quite similar to the situation in which the German dialects in the Hungarian diaspora found themselves. The way to success in the middle class led through Magyarization. Thus, since the turn of the century and especially since 1918, both Yiddish and the Yiddish-influenced German used by Jews in Hungary {Jüdisch-Deutsch) yielded to the encroachments of Hungarian, especially in the realm of written language. The only area in which Yiddish remained intact was the narrow circle of friends and relatives and in the synagogue, and even there, only within Orthodox Jewry. In western and central Europe the WY dialects must have died out within a short time during the period of reforms following the Enlightenment. That is what took place in most parts of the compact Germanspeaking territory during the last century. In border zones and outlying areas, the demise of Hungarian Yiddish was retarded by factors previously mentioned, but above all by the westward movement of several groups of EY speakers. Hungary was not the only area where WY and EY were in contact, but nowhere was their contact more important to linguistic development than in the Carpathian zone in the second half of the 19th century. Polish and German Ashkenazic Jewry interacted here with extreme intensity ever since the establishment of Jewish communities had become possible in the central counties of Hungary. In this interaction the more adaptable and more dynamic were the EY speakers. They had brought along with them stable popular and cultural traditions, well established patterns of living. They were, moreover, less influenced by a direct penetration of German than their WY counterparts in Hungary. The centers of Orthodox religious education were removed partially to the east (Galicia); at the same time, the W Y synagogues and schools more and more engaged rabbis, cantors, and teachers from the EY area. From the start the confrontation between WY and EY Ashkenazim was productive of disharmony. The stand-offish WY Oberländer never ceased to look down on the EY Unterländer, often sarcastically referring to them as pöjlise jidn or tstilntige jidn. Nevertheless, the active, adaptable EY immigrants, more deeply rooted in their culture and way of life, inevitably and
PROBLEMS OF WESTERN YIDDISH DIALECTOLOGY
5
unmistakably left their mark on Western Jewish Orthodoxy, in regard not only to religious and popular customs (including folklore), but also to language. On the other hand, the socially more advanced WY speakers did not fail to influence the representatives of EY. The Easterners developed ramified cultural and linguistic responses. Moreover, representatives of WY in Hungary contributed to the disintegration of the communities of EY newcomers in pursuit of Magyarization. In the light of the situation just described, the priorities and the practical research problems in the field of WY follow directly from the basic postulates of a sociologically sophisticated dialectology. Let me here enumerate them briefly, rounding out, so to speak, the observations made by U. Weinreich (1964b) and others in their recent field work. (1) Concerning WY, the best practical method is a continuation of the Atlas work now in progress, in conformity with the insights of linguistic geography. We know that in several areas of the Yiddish speech territory, linguistic development proceeded along different lines. Furthermore, there were marked differences in each area in the extent of contact between Yiddish and other languages. Moreover, in Hungary, renewed contacts with EY led to the development of a transitional zone between EY and WY. All these factors produced different regional formations, and it would seem desirable to set up regional atlases of the most important areas, besides the general Language and Culture Atlas of Ashkenazic Jewry, to put problems specific to the various areas into proper perspective. (2) In order to reveal the dynamics of W Y, both in structure and in diachronic interpretation, we require a stricter coordination than is customary in the investigation of EY between Yiddish dialectology and parallel work in the field of German—in dialectology, language, history, and descriptive linguistics. (3) The dialects of WY which can still be collected have acquired their present-day form in close contact with other dialects and languages. Without a many-sided investigation of these contacts it will not be possible to reveal or to analyse the present-day situation. This statement does not concern territorial problems alone. One can attack "horizontally," i.e. in their geographical extension, the problems of interference between WY and, say, Hungarian, Slovak, and Rumanian; but the statement also refers to questions of sociolinguistics, i.e. to "vertical" structures and diasystems, for example to the relation between WY and Rotwelsch (thieves' cant), as well as the several slangs of non-German surroundings. It will also, I think, be very useful and important for Yiddish dialectology to clarify the role of Yiddish as a source: Yiddish elements in these conditions
6
C. J. HUTTERER
were transmitted to non-German argot speakers in southeastern Europe, mainly by users of German Rotwelsch. (4) One of the most important problems of our further work concerning the inner growth and structure of Yiddish is an all-around clarification of the stimuli, the present forms, and the results of contact between WY and EY in the transitional zones. If possible, this should be investigated in the framework of parallel ethnographic-linguistic study. (5) An adequate solution of the problems mentioned in paragraphs (2)-(4) can only be assured by a complex treatment of all materials that can be discovered. Such a complex method requires the utilization of all branches of linguistic, historic, ethnographic, and other social sciences for a consideration of synchrony in harmony with diachrony. *
A detailed account of the general methods of our fieldwork, which follow from the above considerations, is beyond the scope of this paper. Besides, these are widely known in general dialectology as well as in Yiddish studies. Here worth pointing out are some specific characteristics of WY dialectology, as carried on by myself and an occasional associate in Budapest. In the organization of fieldwork, the most difficult problem is to obtain suitable informants. Theoretically, we are bound to regard every traditional Jewish community in Hungary as a potential interview point, all the more so because it is the only possible way to get an adequate corpus from the different cultural and dialectal subunits. Tape recordings are of utmost importance: many native speakers of WY are old people, and in some cases postponement would be tantamount to forfeiting the opportunity of direct recording. In interviewing the informants by questionnaire, we try to avoid putting questions in colloquial Yiddish or German, bearing in mind the linguistic interference and affinity discussed above. Within Hungarian linguistic territory we generally pose the questions in Hungarian, thus reducing the possibility of other language or dialect infiltration to a minimum. If there is no chance of collecting current dialect texts, we try to reconstruct the outlines of the system of the extinct local dialect on the basis of Yiddish survivals or the phonological characteristics of the losn-hakojdes of the assimilated Jewish population. Such data must not, of course, be considered as equivalent to the recent materials without reservation. Given the two kinds of procedure, we have to perform a twofold task. Primarily, we collect data and describe the linguistic situation as it is, even if it retains only vestigial elements of its former dialect standard. At the
PROBLEMS OF WESTERN YIDDISH DIALECTOLOGY
7
same time, by enlarging and systematizing the basic corpus we try to screen out the genuine elements. This phase of our work is made more difficult by the infiltration of EY and literary and colloquial German forms. A further method which we can apply in order to reconstruct an older stage is the linguistic analysis of those written data which can be geographically identified. The specific techniques used in this connection need not be enumerated here, since they are well known to linguists. I have purposely left one problem until the very end, namely the urgent need for maps combining detailed synchronic and diachronic information. The most recent situation would thus be displayed graphically alongside the historical reconstruction which is connected with it. Such maps would be helpful to a specialist working in WY even more than to a student of EY. *
We are now witnessing the end of an era, an era when some parents are making heroic efforts, against what seem to be insurmountable odds, to preserve the WY and, ironically enough, also the EY tradition, which they encountered in the concentration camps in 1944. It was at that juncture that the immense wealth of the culture which they had once considered inferior revealed itself to them. Similar encounters in kibbutzim in Israel and in diaspora settlements elsewhere have led to the same realization. We hope that the crisis for Yiddish as a whole has now passed. For WY, of course, this means that it will be absorbed by the more vigorous EY. It is our scholarly task and moral duty to rescue all that can be rescued of what remains of WY wherever it is spoken or used as a medium of communication.
E N D I N G E R
JIDDISCH
FLORENCE G U G G E N H E I M - G R Ü N B E R G
Die nachstehende Aufzeichnung ist ein Bruchstück des Textes, den ich aus meinen Tonbandaufnahmen vom aussterbenden westjiddischen Dialekt der Surbtaler Juden (Guggenheim 1964) zusammengestellt habe für die Anfertigung von zwei Langspielplatten. Diese werden, zusammen mit einem linguistischen Kommentar, als Heft 4 der Reihe "Schweizer Dialekte in Text und Ton" herausgegeben werden. Die Aufnahme stammt aus dem Jahre 1957; Aufnahmen von 1950 sind auf Sprechplatte ZA 150/151 des Phonogrammarchives der Universität Zürich aufgezeichnet. Der Sprecher des vorliegenden Textes, I.G., ist 1875 in Endingen geboren und starb 1961 in Zürich.
PHONEMISCHE UND PHONETISCHE HINWEISE
Da wir keine phonetische Umschrift haben, ist unsere Uebertragung nicht rein phonemisch, sondern berücksichtigt auch die wichtigsten phonetischen Erscheinungen, vor allem bei den Vokalen. Die Grundlagen für unsere Arbeit haben wir in dem Aufsatz "Zur Phonologie des Surbtaler Jiddischen" (1958; im Nachfolgenden als "Phonologie" zitiert) zusammengestellt. Das Phoneminventar des Surbtaler Jiddischen umfasst: VOKALE
lang
y
ii
uu ee
kurz
ee oo aa
/
od
üü
ej
ou ai
u ü » od a Als wichtigere Varianten sind anzuführen: Das offene lange ii steht nur bei Wörtern, in welchen es, etymologisch, durch Entrundung von ü entstanden ist. Das offene lange ee und das offene kurze e stehen nur vor r, velarem ch und j. e
e
8
9
ENDINGER JIDDISCH
Offenes / neigt leicht zu geschlossenem e, geschlossenes e neigt vor Nasalen zu offenem i. aa und a ist ein mittleres a. Bei Nasalierung wird aa jedoch stark verdumpft. Das offene lange öd und das offene kurze o stehen nur vor r und uvularem ch. Geschlossenes oo und o neigen leicht zu üü resp. ü. Das geschlossene o tönt nach r oder ch etwas offener als in anderer Stellung. Der Reduktionslaut a ist immer unbetont. Die Aussprache von ou schwankt zwischen du und au. KONSONANTEN
p b m
ff f iv
t d
ss s
f
f
gg g
Ich
ch h
j
n ng nng ngg I r Pf ztz tf p und t sind unbehauchte Fortes, die, besonders im Anlaut, eine starke Tendenz zur Lenisierung zeigen. Im Auslaut wird t nach Vokalen meist als Fortis realisiert, nach Konsonanten, ausgenommen r, /, m und n, jedoch stark lenisiert. Der besseren Lesbarkeit wegen schreiben wir im Auslaut immer die Fortis. Die Lenes, b, d, s, g sind stimmlos.—Bei vielen Sprechern werden Fortis und Lenis in der Aussprache kaum unterschieden, sie erscheinen dann meist als schwächere oder stärkere Halbfortes (s. "Phonologie", §§3-1, 3-4).—Auch bei starker Lenisierung von t und p schreiben wir immer ft und fp. Der alveolo-palatale Reibelaut p (nur in Wörtern deutschen Ursprunges vorkommend und dem deutschen Ich-Laut entsprechend) lässt für das Ohr nur einen Intensitätsunterschied gegenüber / h ö r e n . Er kommt nach Ii, y, ee, e, ai, ej und nach Konsonanten im Inlaut und im Auslaut vor. kh steht vor Vokalen in betonter Silbe, gg in allen übrigen Positionen, wobei letzteres in den Stellungen vor und nach den Konsonanten /, n und r stark zur Lenisierung neigt. ch ist im Anlaut vor den Tiefzungenvokalen a, aa, ö und öd, im Inlaut und Auslaut nach denselben sowie nach r uvularer Reibelaut, in allen übrigen Fällen velar. Es sind kombinatorische Varianten desselben Phonems (nicht als zwei verschiedene Phoneme aufzufassen, wie ich es in meiner "Phonologie" getan habe). nng, ngg, tz benützen wir um der besseren Lesbarkeit willen.
10
FLORENCE GUGGENHEIM-GRÜNBERG
Die Doppelkonsonanten t t , f f , II, mm, nn werden nur geschrieben, wenn sie entweder ausgesprochen werden, oder wenn sie zur besseren Lesbarkeit nötig sind. Von Wörtern hebräischen Ursprunges wird der Sg. nur angeführt, wenn er gebräuchlich war. Der Akzent wird nur bezeichnet in Wörtern mit mehr als einer Silbe ohne a. Für weitere Einzelheiten siehe "Phonologie". TEXT ERINNERUNGEN EINES ALTEN ENDINGERS
(.Berufe der Juden.— Vorbereitungen auf den Sabbat.)
5
10
15
20
1
2 3
4 5
Endingan is (1) zu mainar, ¡also for [achzig jöor, noch 3 ganz refpegg[dabli (2) khils (3) gawee. For ahm di grouss sina[goog, wü ms fon [ali saita gseja hot, daas zu sebar (4) zait fejn [fuulhous, wü drai leerar in sibd gglassa dsr 'untarift erHaalt hen, an 'arbaitsleeraran, wü di1maadlig (5) ftrigga un neea gal'ernt hot, dsr chasan (6), bai deem hemmar loofan (7) 3xkhoudif (7) ibar'setza und1 fraiba galernt. Es sin noch dou gawee a ^jidifar roufa (8), fyr fpezahaihendlar, zwai1zchöörahendlar (9),fimf2 khazoufam (10), zwai 'fifhendlar, zwai fnaidar, a 'zuggarbegg un 3 1buuchbindsr. E paar hen aach mit nachUooas (11) gamacht, aber dar 1 waitous greeft taal sin bs'hejmashendlar (12) gawee. Si hen maiftsns3 Hri madynas (13) *usswerts4 ghet. Am lsuntig odar 'meentig frii sind5 sifört und1 menngar hot noch a paar fee (14) mess laafs, bis ar bai saim ftall gawee is. Di ganz wuch hen si fghandelt und1 sin aach uf di merggt gganga. Si hen sif 1groussi zdöras (15) laangadoun, hen ganz *fpaarsam galeebt un nid aanar hot trejfas (16) ggessa. Si hen sif di ganz wuch gaplougt un gaplangt uf da fraitig zoubad und1 uf fabas (17).
D i e N o r m lautet un\ und ist schwzd. Einfluss.
Norm ist finaf. D i e alte Form
maanftans wird
Norm ist 1ouswerts. D i e Norm lautet
sin; sind ist
noch zuweilen gehört.
schwzd. Einfluss.
ENDINGER JIDDISCH
25
30
35
40
11
Da fabas (17) is z 1 Endinga for ganz khaal (18) a joum (19) manuucha (19) und1 a joum (20) simcha (20) gewee. Am 1 dd(r)fiig hen mir jaloodim (21) in di myl ousse zum miliar.6 Frai mesa un hen meel gholt7 mit em 1khindarweegala. E seggla 1hiinarfutter hemmar bachinam (22) dazuu bakhuma. Wü mar 1haamkhoma sin, hot di mamma (23) ghebalt und1 znaacht gaggnet. Am fraitig amorga foumar (24) labouggar (24) is si 1uffgftanda un hot dar ofa ,aanghaazt, un am fimfa8 zmbrgas hob ig, bafbr ig nach ] Zdrzi( in di baxzirggsfuul galoffa9 bin, fou 1haassi 1 herdepfal- un 1zibaldina ghet. 'Nödchheer hots in dar ganza 1wounig nach frifam leecham (25) gfmeggt. Dar etta (26) hot aach maiftans3 a huun odar zwai [haamgabringt, si hend10 jou 1sebmoul (27) nid fyl g/uggt (28), a frangga odar a frangga 1zwanzig pro ftigg. Aach doogam (29) hen nid feela derfa, mar hot eba da fabas (17) aach mit toffar (30) alchyla (30) gfaijart. foun am fraitig 'förmetaag hen sif di 1manslait bal'byra losa, un
am 1nöchmetaag hot ma sig fou \fabasdigg (31) 'aangadoun. Di memma (23) hot da fulchan (32) mit ara Jejnna zweel (33) gadeggt, zwai lfabaslaigdar mit kherza, odar dy, woo a 'fabaslamp (34) ghet hen, a fabaslamp [aangazunda. Un 45 in dar mitt11 zwai 1chalabrejtlig (35), mit ama fejnna 1 khidafdeggla (36) 1zuugadeggt, dazwifa galejgt. Wüs zait is gawee zum fuulagej71 (37) hot si ent*zunda (38) und1 di brbocha (39) gsaagt. Dy 1manslait sin fuula (37) un nooch fuul (37) hot mar uns khindar, zeerf dar etta (26) 50 un xnbochheer di memma (23), gabenft (40), un myr khindar hend10 ina a guut fabas (17) gawunfa. Mar is am tif1anagsessa, hot da füülam a]leecham (41) gsunga un barychas goodal (42) znaacht ggessa. DEUTSCHE UEBERSETZUNG
Endingen ist zu meiner Zeit, also vor achtzig Jahren, noch eine ganz respektable Gemeinde gewesen. Vor allem die grosse Synagoge, die man a
Prot, mülhr, schwzd. Norm ist ghoult. e Prot, fümfa, schwzd. 9 Man hört auch noch die ältere Form galaafa. 10 Die Norm lautet hen; hend ist schwzd. Einfluss. 11 Prot, 'mini, schwzd. 7
12
FLORENCE GUGGENHEIM-GRÜNBERG
von allen Seiten gesehen hat, das zu jener Zeit schöne Schulhaus, wo drei Lehrer in sieben Klassen den Unterricht erteilt haben, eine Arbeitslehrerin, welche die Mädchen stricken und nähen gelehrt hat, der Vorbeter, bei dem haben wir Hebräisch übersetzen und schreiben gelernt. Es sind noch da gewesen: ein jüdischer Arzt, vier Spezereihändler, zwei Schnittwarenhändler, fünf Metzger, zwei Fischhändler, zwei Schneider, ein Zuckerbäcker und ein Buchbinder. Ein paar haben auch mit Liegenschaften gemacht, aber der weitaus grösste Teil sind Viehhändlergewesen. Sie haben meistens ihre Gegend (wo sie ihre Kundschaft hatten) auswärts gehabt. Am Sonntag oder Montag früh sind sie fort, und mancher hat noch ein paar Stunden zu Fuss gehen müssen, bis er bei seinem Stall gewesen ist. Die ganze Woche haben sie gehandelt und sind auch auf die Märkte gegangen. Sie haben grosse Mühsale auf sich genommen, haben ganz sparsam gelebt, und nicht einer hat verbotene Speisen gegessen. Sie haben sich die ganze Woche hindurch geplagt und sich gesehnt nach dem Freitagabend und auf den Sabbat. Der Sabbat ist in Endingen für die ganze Gemeinde ein Tag der Ruhe und der Freude gewesen. Am Donnerstag haben wir Kinder in die Mühle hinaus müssen zum Müller Frei und haben Mehl geholt mit dem Kinderwägelchen. Ein Säcklein Hühnerfutter haben wir umsonst dazu bekommen. Wie wir heimgekommen sind, hat die Mutter Hefeteig angemacht und abends geknetet. Am Freitag Morgen, bei Tagesanbruch, ist sie aufgestanden und hat den Ofen angeheizt, und um fünf Uhr morgens habe ich, bevor ich nach Zurzach in die Bezirksschule gelaufen bin, schon heisse Kartoffel- und Zwiebelkuchen gehabt. Nachher hat es in der ganzen Wohnung nach frischem Brot gerochen. Der Vater hat auch meistens ein Huhn oder zwei heimgebracht, sie haben ja damals nicht viel gekostet, einen Franken oder einen Franken zwanzig pro Stück. Auch Fische haben nicht fehlen dürfen, man hat eben Sabbat auch mit gutem Essen gefeiert. Schon am Freitag Vormittag haben sich die Mannsleute halbieren lassen, und am Nachmittag hat man sich schon sabbatmässig angezogen. Die Mutter hat den Tisch mit einem schönen Tuch gedeckt, zwei Sabbatleuchter mit Kerzen, oder diejenigen, welche eine Sabbatlampe gehabt haben, eine Sabbatlampe angezündet. Und in der Mitte zwei Brötchen für den Brotsegen, mit einem schönen Weihe-Deckchen zugedeckt, dazwischen gelegt. Wie es Zeit gewesen ist, in die Synagoge zu gehen, hat sie (die Lichter) angezündet und den Segensspruch darüber gesagt. Die Mannsleute sind in die Synagoge, und nach der Gottesdienst hat man uns Kinder, zuerst der Vater und nachher die Mutter, gesegnet, und wir Kinder haben ihnen einen guten Sabbat ge-
ENDINGER JIDDISCH
13
wünscht. Man ist an den Tisch gesessen, hat das Begrüssungslied für den Sabbat gesungen und in grosser Behaglichkeit zu Nacht gegessen. MORPHOLOGISCHE, SYNTAKTISCHE UND LEXIKALISCH-SACHLICHE ANMERKUNGEN Abkürzungen: dU. = deutschen Ursprunges; hU.=hebräisch-aramäischen Ursprunges; hebr. = hebräisch-aramäisch; schwrd. = schweizerdeutsch(e, -er, -en); Dim. = Diminutiv; Akk. = Akkusativ; Z. = Zeile.
(1) Die Mehrheit unserer Surbtaler Sprecher braucht noch fast durchwegs die alte Form is, eine Minderheit spricht fast immer i f , in Angleichung an schwzd. Dagegen haben alle elsässischen Sprecher das if der nichtjüdischen Umgebung angenommen. (2) refpegg'dabli: aus schwzd., resp. als Lehnwort Übernomen, viell. aus dem Elsässer Jiddischen ? (3) khih hU. = Gemeinde nbnp. (4) seber: aus schwzd. übernommen, die Norm ist sellar. (5) Der Plural der Verkleinerungsformen auf -la und -ala lautet immer -//f.
(6) chasan hU.=Vorsinger, Vorbeter ]tn. (7) loofan alkhoudif hU. = die Sprache des Heiligen, d.h. hebräisch p®1? trnpii. (8) roufa hU.=Arzt, auch Tierarzt N S n . (9) zchböra hU.=Ware, hier speziell Textilien m w o . (10) khazoufam hU. = Metzger, PI.; Sg. khazafnxp. (11) nach[looas hU. = Güter, Liegenschaften msVni. (12) bahejmas hU. = Grossvieh, PL; Sg. bahejma niöP2. (13) madynas hU. = Gegenden, Bereich, wo sie handelten; Sg. madyna m n » . Die jüdischen Viehhändler und Hausierer (und auch die Bettler!) pflegten die verschiedenen Landstriche, wo sie handeln durften, unter sich aufzuteilen und vom Vater auf den Sohn zu vererben. In der madyna eines andern zu handeln galt als unkorrekt. (14) fee hU. = Stunde, wird für Sg. und PL gebraucht ns?B>. (15) zööras hU. = Sorgen, Leiden; Sg. zööra ITHS. (16) trejfas hU. = verbotene Speisen, eig. zerrissene Tiere, d.h. nicht rituell Geschlachtetes; Sg. trejfa mants. (17) fabas hU. = Sabbat, Ruhetag n a » . (18) khaal hU. = Gemeinde, im engeren Sinne die Gemeindeversammlung, während khih mehr allgemein die Gemeinschaft bezeichnet, khaah = die Gemeindeversammlung abhalten; "?np. khaal geht (wie auch z.B. chafar = Freund i a n ) mit der Aussprache des Kamez als a, nicht als o
14
FLORENCE GUGGENHEIM-GRÜNBERG
wie in der heutigen hebr. Gebetsprache, auf die vor der Mitte des 13. Jahrhunderts bei den Juden Deutschlands übliche Aussprache des Hebräischen zurück, welche später die sefardische genannt wurde (s. M. Weinreich 1954). (19) joum manuucha hU. = Ruhetag nnilö m \ (20) joum simcha hU. = Freudentag nn»B> DV. (21) jaHoodim hU.=(kleine) Kinder; Sg. jebd, Dim. jehdh 0 , T 1 ?\ (22) bachinam hU. = umsonst, unentgeltlich Dina. (23) mamma: schwzd., ist heute bei den Surbtaler Juden die geläufigere Bezeichnung für die Mutter als das alte memma. (24) fotimar labouggar hU. = frühmorgens, bei Tagesanbruch npa 1 ? "löW. (25) leecham hU. = Brot an1?. (26) etta: diese alte Bezeichnung für 'Vater' (nicht für 'Grossvater' wie im schwzd.) wird heute weniger gebraucht als schwzd. bappa. (27) 1sebmoul: schwzd. Einfl., die Norm ist 1selmoul. (28) gfuggt hU. = gekostet. Von dem Subst. fugg = Markt (auch für die deutsche Münze 'Mark' gebraucht) wurde ein Verb fugga = 'kosten' gebildet p w . (29) doogam hU. = Fische c m . (30) toffar a'chyla hU. = (mit) guter Speise. Das Adjektiv toff=lguC wird dekliniert wie das deutsche 'gut' fiVoK, 310. (31) \fabasdigg hU. = sabbatlich. Die Adjektivendung -tig wird an das Substantiv hU. angefügt na®. (32) fulchan hU.=Tisch in1?®. (33) zweel: das Jiddische hat das veraltete deutsche Wort Zwehle = 'Tuch' bewahrt, auch als Hifzweel='Tischtuch' und als [handzweel='Handtuch'. (34) 1 fabaslamp: für den Sabbat benützten wohlhabendere Juden anstelle von Kerzenleuchtern eine Hängelampe aus Messing, lümp genannt, mit sechs sternartig angeordneten Behältern für Dochten und Oel. (35) 1chalabrejtlic: ein Ausdruck für die Sabbatbrote, den wir sonst bei keinem unserer Gewährsleute gehört haben; zusammengesetzt aus chah hU. = 'Teighebe' nh>n und *brejtli( dU. = Brötchen, PI. (36) khidafhU.=Weihesegen für Sabbat und Festtage, bei einem Becher Wein tfnp. Während dieser Zeremonie werden die Sabbatbrote zugedeckt. (37) fuul= Synagoge, auch: Gottesdienst in der Synagoge. Vom lateinischen 'schola judaeorum'. Als fuul bezeichneten die Juden schon im Mittelalter ihr Bethaus, in welchem auch gelehrt und gelernt wurde. fuula=it\s Bethaus, im Bethaus, ifuulagejn = 'm die Synagoge gehen; ursprüngliche Form ist fuulan geen, wahrscheinlich Akk. beim Verb der Bewegung.
ENDINGER JIDDISCH
15
(38) ent^zindd, Partizip ent[zunda, bei jüngeren Sprechern ent'zindt, wird nur für die Zeremonie des Lichtersegens gebraucht, für profanes Lichteranzünden sagt man ^aanzinda. (39) brbochd hU. = Segensspruch ¡1313. (40) ben/3, vom lateinischen 'benedicere'= segnen, auch: das Tischgebet nach der Mahlzeit beten. (41) füübm aHeechdm (auch foohm aHejcham) hU. ='Friede mit Euch' OlVtP. Anfang eines hebräischen Liedes, mit welchem man den Sabbat begrüsst, unter Umhergehen in der Stube. Im Alltag ist das verstümmelte fümHeecham der Willkommensgruss unter Männern, besonders bei einem Wiedersehen nach längerer Zeit (vgl.Guggenheim 1954a). (42) barychas goodal h U . = i n grosser Behaglichkeit m a n s a . (Laut freundlicher Mitteilung von Prof. Max Weinreich sind die Wörter auf -us > -3s im Jiddischen älterer Zeit maskulin.)
RESULTS OF ATLAS AMONG
JEWS OF
INVESTIGATIONS GERMANY
STEVEN LOWENSTEIN 1. INTRODUCTION
It is only in the last few years that Yiddish dialectology has made an effort at a systematic coverage of Western Yiddish (WY) dialects. 1 Whereever the language had remained (until recently) a current medium of discourse—viz. on the Swiss, Alsatian, and Hungarian fringes of WY territory—field work on behalf of the LCAAJ, as well as other investigations, could proceed in a fairly conventional manner, and, with appropriately modified questionnaires, satisfactory results were obtained. 2 In the heartland of WY, however, the problems of coverage were different and vastly more complicated. Yiddish in Germany and Austria had presumably died out almost completely in the 19th century and German cultural influence on German-speaking Jews is known to have been very strong. It was hardly to be expected that any but a few remnants of WY language and culture could be collected from emigrants from German-speaking Europe in the middle of the 20th century. 3 Our work on behalf of the LCAAJ, however, has shown that the extent of linguistic and cultural assimilation among German-speaking Jews has been overestimated. Not only can many facts about Jewish speech in German-speaking areas still be collected, but regional patterns can still be differentiated sharply enough to be readily recognizable on maps. [Since this report was first presented in June, 1965, Beranek's Westjiddischer Sprachatlas has appeared. Eight of his maps cover much the same material as some of our own and others bear indirectly on discussions in our text. While most of Beranek's maps seem to reflect many of our findings quite 1
For a comprehensive programmatic statement, see M. Weinreich (1953). Fieldwork for the LCAAJ was conducted among emigrant informants from Germany, Austria, Holland, Czechoslovakia, and Hungary now residing in the United States and in Israel. Alsatian and Swiss informants were interviewed in Alsace and Switzerland, respectively. 3 More or less the same applies to Yiddish in the Netherlands. 2
16
ATLAS INVESTIGATIONS AMONG JEWS OF GERMANY
17
accurately, several of them are irreconcilable with our data. Generally speaking comparison is difficult because of Beranek's stated emphasis on "simplicity" and his use of "kritisch durchgeführte Retuschen". Suppressing the data from individual locations, his maps display only isoglosses. It might have been expected that data gathered over at least four decades —decades of chaos and catastrophe for the communities under investigation—would be better documented. Yet Beranek reveals neither the dates of his interviews nor pertinent biographical data about his informants; he fails to distinguish oral from written responses and, though he admits to the use of literary sources, he nowhere indicates where or how they have been employed. Such lapses detract from the value of materials that might otherwise have proved priceless.] In German-speaking surroundings where Yiddish was subject to a process of dilution by German, the extent to which features have been conserved varies both according to the age and profession of each informant, and to the size and location of his place of birth. Except for a small minority in Alsace and Switzerland, and in those parts of southwest Germany immediately bordering these areas, the Jews of German-speaking Europe did not claim, and cannot be said objectively, to have spoken Yiddish in the generation of our informants. In many of the areas covered by our study, however—especially in southern and central Germany—our older informants were able to remember that their parents and grandparents had spoken a language diverging widely from the local German dialects.4 Our best informants were former cattle dealers from small communities in southern and central Germany. In their businesses they had made extensive use of a Jewish market language called losnakoudes 'the holy tongue'. This "language' was made up primarily of words of Hebrew origin in a German grammatical framework. Thus '200 marks' was bejs meje suk from bejs 'two', meje 'one hundred', and Suk 'market'. Since the German words Markt 'market' and Mark (the coin) were similar, suk 'market' came to mean Mark in losnakoudes. Other complex lexical items in this language consists of stems of Hebrew origin to which Germanorigin verbal prefixes and suffixes have been added; thus ferkinjenen 'to sell', from Ashkenazic Hebrew kinjen 'a purchase', modeled on German kaufen 'to buy', verkaufen 'to sell' (cf. Guggenheim 1954b). Many Central European Jews who did not make use of this market language could nevertheless give their German a "Jewish flavor" by using 4
This is in accord with the findings of Guggenheim (1964).
18
STEVEN LOWENSTEIN
specifically Jewish words or proverbs in their German speech. Such expressions were common in conversations among Jews especially in small, closely-knit, rural communities. Most were of Hebrew origin, although many stemmed from other sources. Some specifically WY words of Hebrew origin are axeln 'to eat', and louxejfec 'loafer, shirker, good-for-nothing'. 5 Some that are of German origin are either obsolescent in the non-Jewish dialects or are put to special use by Jews. Thus volfl (cf. EY volvl) means only 'cheap' (cf. MHG wol veile, wolveil with its many references to venality and corruptibility), while 'to go to the synagogue' is sul gejn. A considerable number of WY words are of Romance origin: oren 'to pray' (cf. Latin orare) and lajenen 'to read from the Torah' (cf. Latin legere) are but two examples. A few of the words are even derived from Slavic, e.g. nebif 'poor thing', koulec 'Shavuoth cake', and xoce 'although'. 6 2 . WESTERN Y I D D I S H
Yiddish in Germany, Austria, Holland, Alsace, and Switzerland, as well as in Hungary, has been classified as WY in contrast with the more familiar Eastern Yiddish (EY). The two differ considerably both in vocabulary and pronunciation. In distinguishing them, the determining criterion has usually been phonological (see e.g. M. Weinreich 1953): in WY, proto-Yiddish et and o4 (cf. MHG ei and ou) have merged and are rendered a; thus, a WY speaker would say (if gej) flas kafn '(I am going) to buy meat', in contrast to the EY variants kojfn flajs, kojfn flejs, or kejfn flejs. Unfortunately, this criterion—apart from the question of the legitimacy of a single phonetic distinction—can no longer be relied upon; Jewish pronunciation in many western areas has been assimilated to German and we must rely instead on lexical differences to differentiate 20th-century WY from EY. The following preliminary maps depict some of the contrasts between EY and WY as well as some differences within WY itself. They are based on the unedited transcripts of completed interviews. Many gaps still remain and a thorough study of the material has hardly begun. Nevertheless, patterns have already emerged that may be confirmed and clarified when all the material is in. Instances of penultimate stress in louxejfec were also recorded. Except in instances where a specific consonantal or vocalic alternation is the subject of discussion, our transcription usually ignores a number of differences among informants. Thus s may represent the alternation of s and z; ( of f, s, and x; g of g and x. Among the vowels, ou may reflect ou~au~d; ej may cover ej~e~aj-, o stands sometimes for d~it; a for a ~ an and a~aj; u for it i. 6
6
20
STEVEN LOWENSTEIN
One of the best examples of lexical contrast between EY and WY is the word for 'Sabbath bread' (Map 1). In EY the usual term is xale; in large areas of WY it is berges/barges or dacer. The eastern limits of barges coincide with the border between WY and EY as earlier determined by the phonological criterion. A ceremony well known among all Ashkenazim is the search for leaven on the day before Passover. Its name varies, however, between EY and WY (Map 2). The WY term is xomec bat In 'to dispose of the leaven'; in EY, it is bojdek xomec/xumec zajn 'to search for leaven'. Note that in this case the border lies somewhat to the east of the isogloss in Map 1; western Poland as far as Lodz shares the WY form. Even more widespread is the WY word for the Hanukkah top (Map 3) —trend(er)l; this term is encountered as far to the east as Warsaw and even somewhat further east where it blends with EY drejdl to yield drejderl.1 Beranek (1965: Map 20) fails to record this extension of WY trendl on EY territory, perhaps because of the paucity of informants from Poland proper. 8 3 . CONTINUITY A N D DISCONTINUITY BETWEEN E Y A N D
WY
Not all Yiddish expressions are so sharply differentiated on an EY-WY basis. Frequently, when there is variation within the area historically occupied by WY, one of the WY variants agrees with an EY form. At other times, all of the WY forms may be distinctive. Quite commonly the specifically WY expression will occur only in Alsace, Switzerland, south and central Germany and parts of north Germany. Against this, Holland, eastern Germany, Austria, Bohemia, and Hungary often disagree. For example (Map 4), the typically WY term for 'dowry'—nedunje/ nedinje—is found only in south and northwest Germany, Alsace and Switzerland. In the rest of the WY area, the term is nedan which recurs again in NEY as nadan. In CY and SEY the word is nadn/nodn.9 Returning to the word for 'Sabbath bread' (Map 1), we find four different forms in the western area: in Holland and neighbouring areas in northwest Germany, xale (as in EY); elsewhere, berges/barges and dacer. A folk etymology connects berges and dacer with the Biblical verse birxas 7
Northeastern Yiddish preserves a totally different form: gor(n). In other instances, where Beranek ventures to draw internal EY isoglosses on the basis of such scant information, he is led more seriously astray. Compare, for instance, both the location of the EY isoglosses and the correspondence between them on his Maps 2 and 3 with those in Herzog (1965a: Fig. 5:74). 9 The picture drawn by U. Weinreich in 1959 (U. Weinreich 1965b: 13) is thus confirmed by new data. 8
ATLAS INVESTIGATIONS AMONG JEWS OF GERMANY
21
STEVEN LOWENSTEIN
ATLAS INVESTIGATIONS AMONG JEWS OF GERMANY
23
adonoj hi ta'asir 'God's blessing enriches' (Proverbs x: 22), said to have been inscribed on Sabbath bread-knives. The word cfacer, presumably from Ashkenazic Hebrew taasir, seems to be restricted to the area around Frankfurt and the Rhine Valley; berges and barges may similarly be derived from Ashkenazic Hebrew birxas by the shift of ir >er>ar (widely attested in Yiddish; cf. M H G kirsche 'cherry', Yid. kers, kars). The boundary between berges and barges seems to be similar to the boundary between nedan and nedlnje (cf. Map 4). The existence of eastern variants of W Y forms (cf. Map 1: barges); their co-occurrence with variants of EY forms in WY (cf. Map 4: nedan); and the extension of WY forms into western Poland (cf. Maps 2 and 3) compel us to recognize the presence of a transition area between the two major dialect regions—specifically, between CY and the Yiddish once spoken in Germany west of the river Elbe. This suggestion is borne out by the name given to the circumcision ceremony (Map 5). In most EY areas it is simply called bris 'covenant'. In Holland and most of Germany, the usual term is brismile 'covenant of circumcision'. In the transition area—in Hungary, Austria, sections of east Germany and in western Poland as far as Lodz, the prevalent word is sude (elsewhere simply 'a festive meal'). Thus, the transition area occupies both WY and EY territory. 10 It would be erroneous to assume, however, that resemblances between WY and EY occur only in the transition area. Sometimes the correspondence is between the geographical extremes—the WY word agrees with the easternmost Yiddish dialects rather than with neighboring CY. Thus, the fringed garment for men prescribed by religious law (Map 6) is usually called arbekanfes in Germany (regionally arfekansef) and in NEY (in Lithuania and Belorussia). Austria, Bohemia, and Hungary in WY territory share with southern Poland and the Ukraine in the southeast a variant of cedakl or la(j)bcede/akl (literally 'body covering'). In northeast Germany and in northern Poland, cicekanfes11 is the common term. 12 It appears then that arbekanfes, found only in the widely separated German and NEY areas, is probably an old form, while those in the center are innovations which, interestingly enough, all occur on both sides of the reconstructed WY-EY border. 10
On Transcarpathian Yiddish as a formation intermediate between WY and EY, cf. also U. Weinreich (1964b). 11 Apparently a blend of cices 'fringes' and arbekanfes, according to a well-known principle concerning marginal distributions. (For other examples cf. Herzog 1965a: 63ff.) 12 Still another term, taleskotn, is also widespread in EY (except Belorussia), but it is usually secondary and can be elicited only through prompting.
24
STEVEN LOWENSTEIN
ATLAS INVESTIGATIONS AMONG JEWS OF GERMANY
27
To some extent the old sharp differences between WY and EY have been blurred by eastern influence in the west. Mostly of very recent date, this is the result of the migration to Germany of many eastern Jews in the late 19th and early 20th centuries. According to the evidence of our maps, eastern influence was stronger in northern Germany than in the south. Its extent may be illustrated by the EY word dav(e)nen 'to pray' (Map 7) which was used alongside the indigenous WY oren both in north Germany and in Bohemia. Many western informants even said that dav(e)nen was the more common term. 13 Other examples of the same phenomenon are the use of xale alongisde bergesjbarges (cf. Map 1) and the use of a seemingly eastern pronunciation in brojges 'angry' and mojre 'fear' instead of western broges/brauges/brouges and more/maure/moure (cf. J. Fischer 1936:91). 4 . INTERNAL DIVERSITY IN W Y 4 . 1 . NORTH VS. SOUTH
Differences between WY in north and south Germany are to be found in vocabulary items as well as in areas of non-verbal culture. A word which seems to be native only to the northern and eastern parts of the WY area (Map 8) is getajlext/getejlext, past participle of *tajlexn/tejlexnli 'to run away' (cf. Ashkenazic Hebrew tejlejx 'you will go'); it is found only rarely in south Germany. Among the most interesting Jewish customs in the WY area is the hdlekras (cf. Landau 1899), the ceremony of naming the newborn child (Map 9). This ancient custom, virtually unknown among German Jews in the north, is practiced only in west-central and southern Germany, Alsace and Switzerland. When the mother goes to the synagogue for the first time after giving birth, a celebration is held at home on the Sabbath afternoon. Friends and relatives and all the children of the community are invited. The children raise the baby in its basket and say: hdlekras, vl sol s bobele hasen 'hilekras what should the child be called'; they then call out the designated name three times. Afterwards the teacher may say a short prayer and the children are sent home with bags of candy. There are many variations on this ceremony among the Jews who practice it. The formula used when lifting the child, and the prayers and poems at the ceremony differ considerably from place to place. The most striking variation, however, is the insistence by some informants that the ceremony is performed only when a girl is born, and by others, who insist just as 13 Evidence for the coexistence of davnen and orn, with slightly specialized meanings, in 15th-century North Italian Yiddish was recently discovered by Kosover (1964). 14 I have rarely heard the word in any form other than the past participle.
28
STEVEN LOWENSTEIN
strongly, that it is performed only for boys; still others tell us that it was held for all children. The reasons for these differences and for their geographic distribution have not been explained and even the origin of the
ceremony is still in dispute. What is very striking, however, is the sharpness of the boundary between areas where the custom was well known and those where it was unknown. 15 1S While the distribution of the hölekras ceremony as shown in Beranek (1965: M a p 23) agrees fairly well with our own data, he inexplicably describes the custom simply as "Namengebung bei Mädchen," without reference to the prominent variation that we have described.
ATLAS INVESTIGATIONS AMONG JEWS OF GERMANY
29
4 . 2 . W E S T VS. E A S T
Aside from a north-south cleavage, there are also WY differences between western Germany on the one hand, and central and eastern Germany on the other. In the west, for instance, etymological » is not ren-
dered in the customary fashion as o but rather as e; thus moes 'money' and orl 'gentile' are pronounced mees or mejs (Map 10) and erl (Map 11); cf. U. Weinreich 1965b: 42f.). This sound shift (of 6 to e) seems to occur only in association with B ; in other contexts o remains unchanged.
30
STEVEN LOWENSTEIN
Perhaps the e variant among the westernmost Ashkenazim indicates the early Yiddish value of »—in Classical Hebrew, the voiced pharyngeal fricative—which has long served exclusively as the Yiddish orthographic symbol for the vowel e (dialectally ej\ cf. U. Weinreich ibid.: 43).
Another peculiarity of the western area is the shift of b to v, so that we hear saves 'Sabbath' and reve 'Rabbi' instead of the more usual sabes and rebe.16 19 We have not yet determined whether our informants bear out the picture drawn by Beranek in MAJ (Folge 2, p. 11, Map 2) for the entire WY area and by Guggenheim (1964) for WY in the Alsatian-Swiss area alone.
ATLAS INVESTIGATIONS AMONG JEWS OF GERMANY
31
The extreme western part of south Germany is the home of many old and sometimes peculiar forms: 'Hanukkah' is called xanje,17 and 'cemetery' is gutort or gedort, not guter ort as in other areas. Survivals of Old Romance forms, previously more widespread, are also peculiar to this area; fyilsl 'maid' occurs especially to the west of the Rhine.18
after front vowels as [j]: xanike > xanige > xanije.
18 S. Pfeifer (c. 1897: 75, fn. 5) cites Bilsel from a notice dated 1765 in Reckendorf, northern Bavaria, far to the east of this area. Further inquiries have shown that informants in their seventies from Reckendorf, and from a town approximately 20 km. away still recognize the word. This information coincides very well with the picture conveyed by Beranek (1965: Map 25).
32
STEVEN LOWENSTEIN 4 . 3 . MISCELLANEOUS VARIATIONS
The influence of dialectal German phonology on the sounds of coterritorial Yiddish must always have been strong. It has been particularly marked in modern times. Thus, the characteristic weakening of consonants
in central German dialects has affected Yiddish words of non-German origin. Voiced stops and fricatives have merged with their voiceless counterparts. For example, tales 'prayer shawl' and dales 'poverty' are both pronounced alike as dales; the initials of zdxer 'male' and simxe 'joy' are
ATLAS INVESTIGATIONS AMONG JEWS OF GERMANY
33
rendered as z;k and g,f and v,p and b have also merged. In medial position g may be rendered x (or j). Thus brouges 'angry' (Map \2)>brouxes (or even broxes=broxes 'blessings') in central Germany.19
Another characteristic of Jewish speech—the dropping of word-final n, as in läfe 'to run', ore 'to pray', etc.—is also common in many German dialects. 19 The area in western Germany where Beranek (1965: Map 34) records so small that it can hardly be reconciled with our data. 2*
for -g- is
34
STEVEN LOWENSTEIN
In all of these cases, however, the actual degree of territorial correspondence between similar features in Yiddish and German remains to be investigated. Similar investigations in EY (cf. Herzog 1965a: 74ff.) have demonstrated the independence of Yiddish in this respect on Slavic language territory. 5. SPECULATIONS AND CONCLUSION
Within the WY area there are also a number of phonological variations that occur in EY as well. Some of them are in fact, criterial to the subclassification of the EY dialects (cf. Herzog 1969: this volume, pp. 58f.). So far we have not classified these differences on a regional WY basis, but certainly some of them will eventually fall into regional configurations here too. One is the alternation of u and i: some informants said that sike 'sukkah' was an older form used by their parents and grandparents but dropped in favor of suke.20 A similar variation is the regional alternation of o and ù : rùse 'anti-Semite', xùxem 'wise man', kiiecn 'rich man', sùlem 'peace' (as against ròse, xóxem, etc., though variation in each word appears to be independent).21 Furthermore, some informants offered xajder 'schoolroom', behajme 'cattle (sg.)', etc. in place of the usual xejder, behejme. The relation of these WY variations to similar variation in EY, as well as to differences in German dialects has not yet been explored. It may be significant to observe, however, that in our records at least they all occur in words of Hebrew-Aramaic origin, i.e. in words, which, unlike those of the German component, have no external model immediately available. Perhaps, then, they are vestiges of old Yiddish dialectal variation in WY, eliminated in the German component of Yiddish on German language territory but consistently preserved in the dialects on non-German territory in the East. As we indicated early in our discussion (§2), the phonological feature that has been considered criterial to the delimitation of WY itself—a in words like ham 'home' and kàfen 'to buy'—is heard only infrequently ; in the speech of many informants it is completely lacking. Although the work of collecting and evaluating facts about the speech of German-speaking Jews is still in its first stages, the results produced so far have been very promising. Hopefully, additional study will further clarify our knowledge of WY as well as of the customs of the Western Ash20
Our informants seem to confirm the difference in the distribution of u/i variants shown in Beranek (1965: Maps 55 and 78): while sike is common in south Germany but absent elsewhere, nedinje occurs everywhere but in southeast Germany. 21 Concerning o > u in Transcarpathian Yiddish, see U. Weinreich (1964b).
ATLAS INVESTIGATIONS AMONG JEWS OF GERMANY
35
kenazim. This has long been a neglected field, but it is of great importance. Germany was the birthplace of the Yiddish language and some of the data uncovered may aid in explaining puzzling characteristics of the language as a whole. In its own right, too, W Y is of great interest and presents a surprising testimony of survival of cultural traits subject to the strongest pressures to disappear.
ALSACE: A N OUTPOST WESTERN
OF
YIDDISH
RICHARD ZUCKERMAN 1. INTRODUCTION
Yiddish is still spoken by a segment, although an ever-decreasing one, of the Jewish population of Alsace, France. In the summers of 1963 and 19641 carried out an on-the-spot investigation of this most esoteric vestige of Western Yiddish (WY) on behalf of the Language and Culture Atlas of Ashkenazic Jewry (LCAAJ). Having established headquarters in Strasbourg, the capital of the province, I was fortunate enough to find several informants in the city itself. I also made side trips to Haguenau, Colmar, and Saverne in order to establish contact with additional informants. The questionnaire used was a modified version of the standard LCAAJ questionnaire. Items directed specifically at Slavic-origin material were eliminated, while a number of special WY questions were added. With the help of this questionnaire, tape-recorded interviews were made with informants from the following locations: Lower Rhine (LR)—Gerstheim (Ge.), Westhoffen (We.), Oberschaeffolsheim (O.), Duppigheim (D.), and Strasbourg (S.); Upper Rhine (UR)—Wintzenheim (Wi.), Grussenheim (Gr.), and Mulhouse (M.). 1 As will appear from the analysis which follows, the basic geographic division of the province into the northern LR region and the southern UR region, as reflected in the German dialects of Alsace, can also be observed in the Yiddish of the area. The various distinguishing features of LR Yiddish as opposed to UR Yiddish will be pointed out in due course. In order to compare Alsatian Yiddish (AY) and at least one variant of Alsatian German (AG), the same questionnaire was administered to a nonJewish speaker of AGS (M. Raymond Matzen, professor of German at the University of Strasbourg). The questions were posed to Yiddish speakers in AY and to M. Matzen in AGS. Where the interviewer's knowledge of AY or A G proved insufficient, French was used to avoid the use of standard 1 The Yiddish names of these localities are gc:rsda, *vesthofa, 'sefisa, dibfg, bor'g, vmdsana, gryss, mllhousa.
36
'stnus-
ALSACE: AN OUTPOST OF WESTERN YIDDISH
37
German (StG), which might have confused the respondents and colored their responses. By avoiding German, more authentic data were obtained. The field worker studying AY is beset by many problems. He must find speakers whose Yiddish has retained the dialectal features of their birthplace, and is relatively free of the influence of Alsatian dialects as well as of
J
^ 0 Strasbourg/ ° O Ol Oberschaeffolsheim 1
Westhoffen
Duppigheim0
/
J
V Gerstheim 0 I
FRANCE
o, /
Grussenheim®
GERMANY
W
Wintzenheim0
« Mulhouse0
1
SWITZERLAND Map 1 StG. But the opinion prevalent in Alsace is that Yiddish is on the same level as the local AG dialect or even lower, i.e. that it is not a "real language" and is hence unworthy of serious study. The attitude on the part of many is basically one of indifference. The stigma attached to the language makes some people reluctant to serve as informants. According to those who agreed to serve as informants, the situation of the language has changed within the memory of people still alive. While
38
RICHARD ZUCKERMAN
Alsace was under German rule (1871—1918), Yiddish held its ground. It then served as a unifying force; it symbolized the will to remain a distinct ethnic group and reflected Jewish dissatisfaction with life under the Germans. But when Alsace once again joined France in 1918, the Jews were among the first to master French and to make it their daily language. French represented a more liberal Weltanschauung and was a prestige language, making its acquisition and use attractive to Alsatian Jews. The threat of anti-Semitism and the desire, particularly during the 1940's, to be as inconspicuous as possible, further stimulated the Gallicization of the Jews. For many older Jewish businessmen, a knowledge of AG has been a virtual necessity. However, in the vast majority of homes parents use AG only occasionally; French alone is spoken to the children in a conscious effort to prevent them from developing an AG accent—always an object of ridicule. As a result, young urban Jews often know far less AG than their non-Jewish friends. As for Yiddish, only a few distinctively AY words have been retained. At the same time, a highly Hebraicized Yiddish continues, as in the past, to serve as the trade jargon of Jewish cattle dealers. To the Alsatian Jew, "real" Yiddish words are those of non-German origin, i.e. those stemming from Hebrew. The latter, however, have undergone considerable change in conformity with the evolving Yiddish phonological system. Thus to the average person their origin is often somewhat obscure, whereas the remaining components of AY can readily be compared with their AG or StG equivalents. In the course of its history AY has been subject to the constant pressure of AG dialects on the one hand, and to the more remote influence of StG on the other. The language has thus to a certain extent become "diluted" with German. For example, when an informant offers as his translation of la plume avec laquelle il écrit 'the pen with which he writes', the form difedar mid dc{:)ra ar srajbt,2 we observe an instance of StG influence; the idiomatic Yiddish equivalent is di fedar vy ar (da)mit srajbt. Among Alsatianisms noted in the speech of one or more of our informants were the following: bil 'swelling, bump' for Yid. bajl; snydsar 'mustaches' for snoudsar ; dy:ga 'staves (of a barrel)' for douga ; debi 'carpet' for debig. It is worth noting that Yiddish, in turn, has exerted an influence on AG and many popular Yiddish words of Hebrew-Aramaic (HA) origin have become current in 2 Except where phonemic contrasts are explicitly intended, our transcription reflects a great deal of nondistinctive, phonetic detail. The symbols used therein are generally selfexplanatory. Those that are not are defined when first encountered. Note that A G Y is lower than AY y. We have marked stress only where it is other than penultimate.
ALSACE: AN OUTPOST OF WESTERN YIDDISH
39
the speech of Christians, e.g. mo:ras han 'to be afraid'. In past years it was not unusual to find non-Jews able to speak AY with a proficiency ranging from elementary knowledge to complete fluency. It is important to point out that AY cannot be considered a "deformation" of AG for several reasons. First, AG has numerous dialects and subdialects which vary drastically on many points, mainly with respect to vowels. AY on the other hand, is relatively uniform (with only minor differences between the northern and southern dialects) and does not agree in detail with any one subdialect of AG. For example, A G has for the most part preserved Middle High German u and t as monophthongs, whereas AY, like StG and Eastern Yiddish, has diphthongized them. Secondly, AY shares many systemic features with other Yiddish dialects, especially with WY as spoken in other countries. We must therefore conclude that AY is an offshoot of a formerly more cohesive WY speech territory. This paper is based solely on the AY interviews made on behalf of the LCAAJ. It restricts itself to: (i) a brief description of the development of the AY consonant system; (ii) a detailed description of the stressed vowel system; and (¡ii) a description of selected features of the grammar. No attempt has been made to include comparisons with previous descriptions. 3 2 . DEVELOPMENT OF A Y CONSONANT SYSTEM
The following is a brief description of the development of some of the salient features of the AY consonant system. We take the phonemes of proto-Yiddish (PY) as our point of departure. 2.1. *d, *t. PY *d and *t have merged completely in AY. We might represent the phoneme as /D/. In initial, intervocalic and final unstressed position /D/ has the lenis voiceless allophone d: dabs 'poverty' (dales)4 and 'prayer shawl' (tales), dreda 'to tread, step' (tretn), rajda 'to ride' (rajtn), ejfsd 'slave' (eved). In stressed final position, we find fortis t instead: 3 For a bibliography to 1959, see U. and B. Weinreich (1959:nos.l43-146). Cf. now also Zivy (1966), with additional references. Regarding the morphology of AG, cf. Beyer (1964). 4 A Y forms are cited in italics; they are followed, as a rule, by the Northeastern Yiddish (NEY) cognate, parenthesized in roman type. When the Southeastern or Central Yiddish (SEY, CY) form is more appropriate for comparison, we cite it instead; note that Southern Yiddish (SY) encompasses both SEY and CY. Sometimes the Swiss Yiddish (SWY) or the Transcarpathian Yiddish (TCpY) form is given as well.
40
RICHARD ZUCKERMAN
sa:t {holts) 'piece of wood' (sajt hole), ('brihysa:t (sejd).
'eye-glass case'
Aspirated th can be analyzed as morphophonemic / D / + / h / , a conjunction which occurs only across a morphemic boundary and usually as a result of the loss of an unstressed vowel: thent AY /v/ after back vowels except before syllabic I: 6 ma:Par 'thin' (moger), masufia 'crazy' (mesuge), xofla 'Christian holy day' (xoge), broufias 'angry' (brejges). Intervocalic *-g- > LR/ (a) after front vowels and diphthongs, and (b) before syllabic /: me. ja plural of ma:pa 'stomach' (mogns); hja 'to lie' (lign); a:ja 5
The sonants / and r may constitute part of the intervocalic environment: dsiv( 'onion' (tsibele), vi harva 'how much, many ?' (Heb. harbe). Note that while the change *b— > —v— has occurred throughout the German component of AY, it has occurred only incompletely in words of Hebrew-Aramaic (HA) origin. 6 It would thus seem to merge with *-b- (cf. §2.2). Note that /v/ is a strongly fricative bilabial [J3] in the speech of the older generation. Younger speakers generally have labiodental [v]. The same is true of Alsatian German.
ALSACE: AN OUTPOST OF WESTERN YIDDISH
41
(< *ajjs) 'own' (ejgn); na:jl 'nail' (nogl); khujl 'pudding' (kugl). (iii) In the UR *-g->g in the stated environments except that we have recorded j (a) after r and / in both the UR and the L R : veljara 'to roll (dough)' (velgern), mor(nja 'morning, tomorrow' (morgn); (b) we have recorded two words of HA origin with j in both areas: sejats 'Gentile boy' (sejgec), da:jas 'worries' (dajges). (iv) The merger of *g and *k in all other environments yields AY /g/: The lenis, voiceless allophone g is derived from the following: initial *g:gyt 'good' (gut); intervocalic *k:bo:ga 'expert' (boki); nonfinal *k after tjihiyga 'to limp' (hinken); *k before sonants:gla:n 'small' (klejn), gnoul 'clew' (knojl), grarjk 'sick' (krank). The fortis, voiceless allophone k is derived from both stressed final *g and final *k; in unstressed final position -g. a:k 'eye' (ejg), sak 'sack' (zak). (v) / g / + / h / across morpheme boundaries yields kh: kholfa past participle of helfa 'to help' khalda past participle of ha Ida 'to hold, keep' 2.4. *x. For some speakers, x~h in certain words of HA origin. Thus, [ haniga 'Hanukkah' (xanike). In some other words of HA origin x~g~$\ satnig '15th letter of Hebrew alphabet' (samax), sadgan or sadsn 'matchmaker' (sadxn). 7 The lenis allophone y occurs in the environments that elicit the lenis allophones of /D, B, g/ (see above): le.-yam 'bread' (lexem), mm(3)y3 'afternoon prayers' (minxe), bejsay 'Passover' (pejsax). In the LR, /x/ and uvular /R/ may merge : 8 bejsay 'Passover' (pejsax) = bejsaR 'angry man' (bejzer); bra: yas 'blessings' (broxes) rhymes with dsa.Ras 'troubles' (tsores); 7
For younger speakers there is no phonetic distinction between f and s. This is the case in AG as well. 8 In the UR we have apical /r/ instead. Both /r/ and /R/ may be syllabic.
42
RICHARD ZUCKERMAN
sayas ( < saxRas) 'morning prayers' rhymes with bay as ( < baRxas) 'Sabbath loaf'. 2.5. *s. ?Y*s > AYs before t in the same morpheme: dy va:s (>AY/ in words of non-Germanic origin: o!i:fa 'olives', fa'sit 'visit' (vizit), do:fad 'David' (dovid), xefra 'society, group' (xevre), eysa/'Esau' (ejsev). 3 . STRESSED VOWEL PHONEMES OF A Y
In this section we will consider each of the vowel phonemes of AY in terms of its phonetic realization and in terms of the proto-Yiddish vowels from which it is derived. 9 3.1. /i/. The short unrounded lower high vowel ranges between [i] and [e*]; 10 it never occurs before r; /i/ is derived from */ l i 3 . Examples: bin, bis, is 'am, (you) are, is' (bin, bi(n)st, iz); khint 'child' (kind); Im 'bee' (NHG Imme); sidar 'since' (NHG seither); bfila 'prayerbook' (tfile 'prayer'), difla 'church' (tifle). 3.2. /e/. The short unrounded mid vowel is usually [e] but has an allophone [e] which normally occurs before nasals and before r. It has the following origins: (a) : ebfl 'apples' (epl), esa 'to eat' (esn), blef 'sheet metal' (blex),ebar 'someone', veh 'to want' (vein), mesar 'knife' (meser), 11 hent 'hands' (hent), emas 'true' (emes), gem 'gladly' (gem); (b) *e3: esl 'donkey' (ejzl), heva 'to lift' (hejbn); (c) *e 5 : levar 'liver' (leber), gneda 'to knead' (knetn), 1svevlheldsla (O.) 'match' (svebele), nevl 'fog' (nepl); (d) *ai (sporadic): sedajsada 'shade' (sotn/sotn), felajfala 'to fall' (fain), flesfflas 'bottle' (flas). A variation in the distribution of allophones occurs in Wi.: while [e] 9
As posited by M. Weinreich (1960). In the system of protovowels, subscript numbers differentiate otherwise similar letter symbols. The significance of these numbers is as follows: 1 designates a short monophthong in the protosystem; 2 —a long monophthong; 3—a short monophthong in open syllable, generally subjected to early lengthening; 4—the nucleus of a diphthong; 5—in conjunction with e only (i.e. *e5), a close protovowel with special distribution (cf. Herzog 1965a: 161). 10 AG cognates typically have [e*]. 11
AGS nuesr, AGM masr.
ALSACE: AN OUTPOST OF WESTERN YIDDISH
43
occurs before a preconsonantal nasal (hent 'hands'), [e] occurs before a prevocalic nasal Qvenigr 'less', glenr 'smaller'). 3.3. /a/. The short low vowel [a] is generally derived from *ay: alt 'old' (alt), 12gast 'pauper' (gast 'guest'); gfar 'village' (Heb. kfar). In the UR subdialect of Wi. /a/ before nasals has the allophone [a] sometimes approching [a]. LR /a/ also corresponds to *a3: gavl 'fork' (gopl), navl 'navel' (nopl 'nipple', 'navel'); however, in UR, as represented by M. and Wi., these cases have /a:/. 3.4. /o/. The short rounded mid vowel is [o] before r and and [o] elsewhere. 13 /o/ has the following sources: (a) *o1:ofa 'open' (ofn), ros 'head' (ros 'chief'), sot 'secret' (sod), dort 'there' (dortn), fart 'away' (fort 'nevertheless'), ig tor, er vort 'I will', 'he will; 14 doxdar 'daughter' (toxter), lox 'hole' (lox). (b) *o3: hosa 'pants' (hejzn), of a 'stove, oven' (ejvn), ^elaboga^elabofia, 'elbow' (elnbejgn); ojfo 'above' (ejbn, SY ojvn),fogl (G.) 'bird', but also fougl{U.)lfoujl(Q.) (fejgl). (c) In a few cases /o/ reflects unstressed *o2 (or *o3), which otherwise, i.e. under stress, is rendered as /ou/ (cf. 3.11): xs!oma" msjlax 'King Solomon' (slejme hamelex); it may also sporadically reflect *a2 (or *a3), which is otherwise rendered as /o:/ (cf. 3.10): Hosns"khoudis 'Hebrew'. Finally, *«i before r is realized as [o]; this is the usual development in CY (and, in some cases, in other Yiddish dialects) as well.15 E.g.: vorsm 'worm', vordsl 'root', gorgl 'throat', chrst 'thirst' (vorem, vorcl, gorgl, dorst); «¿v'p 'through', khorts 'short', doram'tower', masuga umadoraf completely insane' (durx, kurc, turem, . . . meturef, but CY dor(a)x, kor(a)c, toram, mstoraf); vorst 'sausage' (vurst), sorts 'apron' (sirc/serc), for 'for' (far). 18 Allophonic variation again characterizes Wi. /o/ is somewhat less rounded than in LR; moreover, it is realized as [o] before tautosyllabic /, s, and s: holts 'wood', los 'let', maxalas ros 'bareheaded'. 3.5. AY /u/. The short rounded lower high vowel varies in pronunciation between a fully rounded [u] and a somewhat lower, less rounded [u]. It is derived from *uy, and does not occur before r (see above 3.4). 12 13
14 15 10
AG cognates regularly have [4], Before r, AG cognates have [o] and, in some cases [u]; before x, they have [o],
Cf. AG dort, furt', AGS vur, vurt.
AG cognates regularly have [u]. Cf. AGS fe'r, older fur.
44
RICHARD ZUCKERMAN
Examples: un 'and' (un), nus, pi. nusa 'nut' (nus, pi. nis), smuds a 'to kiss', bugl 'back (anat.)' (TCpY pykl 'hump'), stus 'nonsense, nonsensical story' (stus), masusa 'mezuzah' (mezuze), suk 'mark' (coin)', sus 'horse' (sus), misuman 'cash' (mezumen), duma 'church' (tume), ma'sulama 'to pay'. In German-component words before nasals, Gr. shows free variation between /u/ and /o/. Examples: sumer/somar 'summer' (zumer), khuma\ khoma to come' (kumen), ganuma/ganoma 'taken' (ginumen), numa/noma 'only', utjgljotjgl 'uncle', sunjson 'sun' ( z u n ) ; f u n and fori 'of, from' (fun) vary in subdialects other than Gr. as well. 3.6. AY /y/. The high front rounded vowel occupies a somewhat special position in the AY system. 17 By and large, short [y] and long [y:] are in complementary distribution. Although the number of short [y]'s is much greater in LR than in UR, the distribution is almost totally predictable in each area, so that [y] might be analyzed as a contextual variant of the /y:/ phoneme. This would yield a symmetrical pattern in the vowel system with /y:/ functioning as the long counterpart of /u/. However, in UR, where /y:/ is generally long before /l/, we have nevertheless recorded [syta] 'synagogue' (sul), contrasting, e.g., with [smy: la] 'Samuel' (smu(j)el). Furthermore, even in LR (O.), where, as a rule, vowels have been shortened before /s/ and /f/, we find fys 'foot' (fus), ryfa 'to call' (rufn) as against ry:s 'soot', ry.sig, 'ajgary:st 'sooty' (CY a:ngari:st) and hy:f 'hoof' (NHG Hufe). Accordingly, we have to postulate a contrast between long and short /y/ and /y:/, although the words which display it are different in LR and UR, and the functional load of the opposition is extremely small in both dialects. Both /y/ and /y:/ have their origin in proto-Yid. *u2,3. A similar split in the CY development of *a2. 3 is described by Herzog (1965a: 183). We cannot provide a full description of the conditions of near complementarity of [y] and [y: ] which, when disturbed, gave way to the opposition /y/, /y:/. The following conditions, however, seem to prevail: (i) Throughout AY [y] occurs prevocalically: 18 ryig 'peaceful' (ruik), bfye snajda 'to harvest grain' (snajdn tvue), safyas 'Shavuoth holiday' (svues). In LR (e.g., O.) [y] also occurs as follows: 17 In AG cognates one finds AGS [Y:], AGM [y a] and isolated instances of [y:]; e.g. dy: 'thou'. 18 Under these conditions it may be followed by the hiatus-breaker j. The same phenomenon occurs in NEY as well.
45
ALSACE! AN OUTPOST OF WESTERN YIDDISH
(a) before m, n, l:blym 'flower' (blum), syn 'son' (zun), styl 'chair' (stul); (b) before originally voiceless consonants other than /x/ in checked syllables: fys 'foot' (fus), manyt, 'minute', pi. manyda (minut, minutn), vyst 'puss'; 1 9 (c) before a consonant followed by a: hysda 'to cough' (hustn), xysam 'deaf' (xusim 'bewildered'), mydar 'mother' (muter), 1hasysa 'gesundheit' (CY asi: sa). (ii) Throughout AY [y:] appears to occur as follows: (a) word-finally: i y : 'shoe' (sux), dy: 'thou' (du); (b) before originally voiced consonants: bry:der 'brother' (bruder); (c) before x and r : bady:x 'rich' (betuax 'convinced'), dy:x 'cloth' (tux), dy:xana 'to perform the priestly benediction' (duxenen), y:r 'clock' (TCpY y:r), gfy:ra 'strength' (gvure), gfy:ra 'burial' (kvure). In U R , [y:] also occurs before m, n, I: bly:m 'flower', sy:n 'son', sty:I 'chair'. 3.7. A Y /i/. The unrounded high vowel has the allophones [i] and [i:]. The former is distinct from the more open and lax [i] (cf. 3.1.) 20 /i/ is derived primarily from */2 but also from The conditions under which the allophones of /i/ occur appear to be as follows: (i) Throughout AY, [i] occurs before originally voiceless consonants in checked syllables, and to some extent before originally voiceless consonants that were followed by a or a syllabic liquid. Examples: dif 'deep' (tif), hit (hit), pi. of hyt 'hat', kifar 'barrel maker', sprisa 'splinter', 21 sxida 'slaughter' (sxite, CY sxi:te), bfise 'prison' (tfise), bifla 'book (dimin.)' (bixl). In a group of words in which the allophone [i] is derived from *iu it occurs before ( and S: sifar 'sure' (zixer), i f , mif, di(, sip 'I, me, you (obj.), oneself (3rd person)' (ix, mix, dix, zix/zex/zax), nifdara 'sober' (nixter, SY riixter),jisafla 'village' (Wi.) (jisev 'settlement'), vist 'ugly, foul, nasty' (vist 'desolate'). In other environments, [i:] occurs in: spadsi:ra 'to stroll' (spacirn), nasi: res
' r i c h e s ' ( a s i r e s ) , fardi:na
' t o e a r n ' ( f a r d i n e n ) ; madi. na
'region'
(medine 'country, state'); agrousigri. ra 'very cold weather' (krire 'extreme 19
20 21
Cf. A G S
vY:st.
Cf. AGSfe:*], AGM [i e]. Cf. AGS ¡pre";sa.
46
RICHARD ZUCKERMAN
cold'), a xami:md 'hot weather' (xamime), mi:j3s 'ugly' (mijes), msdsi.jd 'bargain' (mecije), bri.dsr 'brothers' (brider). (ii) In the LR subdialects, [i] also occurs before /, m, n in checked syllables, and before la. Thus, the following are some of the words recorded with [i] in LR but with [i:] in UR: mi(:)l 'mill' (CY mi:l), sti{:)l 'chairs', gri(:)n 'green' (grin), bi(:)n 'attic', si(:)n 'sons' (zin), gddsi(:)nam 'rich men', pi. of kho:dsn, spi(:)la 'to play' (CYspi:ln), maxi(:)h 'forgiveness' (mexile). Note, however, that fil 'many' and myn/min 'must' (muzn) occur with short [i] and [y] both in UR and LR. 22
3.8. AY /e:/. The long unrounded mid vowel, like its short counterpart (cf. 3.2), is usually [e:] but has an allophone [e:] before r.23 /e:/ has the following sources: (a) *e3 (occasional): sne: 'snow' (snej), dse:n, pi. of dsa.n 'tooth' (cejn(er)). (b) *e2,3 before r, x: ss:xl 'common sense' (sejxl, SEY sexl), nsfe.-rs 'sin' (avejra, CY navajra), gss.ras 'argument, dispute' (gzejre 'evil decree'), he:ra 'to hear' (hern), de:rh, dimin. of da:r, 'gate' (tejerl), e:rsf 'Sabbath boundary' (ejrev). (c) *e5: me:I 'flour' (mel), ge:l 'yellow' (gel), le:va 'to live' (lebn), le:sa 'to read' (NHG lesen). sve:r 'difficult' (sver), e.-ra/'eve' (erev). (d) *a 2 ,3 brojt before brlt2S 'board' > brejt. This statement provides both an accurate description of the change in 6 The isoglosses between i/e/as and between ejloj i2 have been transferred directly from Figs. 3 and 4 respectively and have not been plotted for the specific examples used in Fig. 5.
YIDDISH IN THE UKRAINE
67
progress in the central Ukraine and a reasonable reconstruction of earlier developments in northeastern Poland—developments which led to the complete change of 12S to ej without residue, and to the replacement of eji2 by oj with some degree of hypercorrection of ej22. In the southeast, hyper-
correction is rampant in both cases : when oj replaces eji2 it leaves a trail strewn with oj22 as well: tojg 'dough', slojf 'bow', lojter 'ladder', cojlem 'cross', kojver 'grave', etc.; similarly, ej replaces i 25 and yields cy'-forms from other i vowels: e.g. nejgn 'melody', rejfn 'to call'.
68
MARVIN I. HERZOG
4 . 2 . OTHER IMPLICATIONS
It is tempting to suggest that eji2 once extended even beyond the limits that we have projected and that (in its merger with ej22) it served as the original barrier to the spread of the CY innovation aj22. Yet, this would imply the further extension of I25 and its subsequent retreat from a part of northeastern Poland that still distinguishes /i/ from /I/—an assumption that appears to be untenable. On the other hand, having projected the former limits of i 38 in northeastern Polandas far as our data permit, we can assume these to have been the minimal former extension of eji2 as well. In their southeastward migration both l25 and eji2 probably travelled together. Alternatively, l2S can be assumed to have preceded eji2; i.e. *bret25 'board'>brit before *broyti2 'bread' > brejt—the reverse of the process by which first eji2 then l25 is replaced by its neighboring variant. 5. HISTORICAL CORRELATES
Fig. 6 provides extralinguistic support for the relationships between northeastern Poland and the Ukraine that we have inferred from the linguistic data. It indicates the Jewish communities in both areas that appear in the documentary record before 1648.® Different symbols designate the first documentary mention of (i) communities on record before 1566, (ii) others that appear in the record between 1569 and 1592, and (iii) still others first mentioned in the period between 1602 and 1648. The break between each period is motivated by a marked increase in the number of new communities that appear. In one case, it also corresponds to a significant political event—the Union of Lublin—that unified Poland and the Grand Duchy of Lithuania in 1569 and opened the frontiers of the Ukraine to settlement by large numbers of Jews and Poles alike. It is significant that the dates of settlement in the Mazovian-Podlasian border area in northeastern Poland correspond to the dates of earliest settlement in the Ukraine. The chronological relationship between these areas suggests a relationship among their settlers that is borne out by the linguistic evidence. The 16th century limits of settlement in Fig. 6 correspond precisely to the earlier limits of I25 and eji2 projected on the basis of their vestiges in the southern Ukraine and on the record of their former occurrence in northeastern Poland. Since we have shown, moreover, that i25 could have 9 Data for the Ukraine is derived from S. Ettinger 1956. For northeastern Poland, see Herzog (1965a: Figs. 6:2-6:5 and p. 269, n. 5).
YIDDISH IN THE UKRAINE
69
originally emerged as a significant regional feature only in northeastern Poland, we can infer the northeastern Polish origin of many of the earliest Jewish settlers in the Ukraine. This is borne out in turn by the historical record itself (cf. Shatzky 1938:27f.). 5 . 1 . RESERVATIONS
The assumption that the present-day configuration of isoglosses in SEY somehow reflects the direction and pattern of Jewish settlement before 1648 can give rise to serious objections. It is known that the Chmielnicki
70
MARVIN I. HERZOG
massacres of 1648-49 led to the virtual destruction of the Ukrainian communities shown in Fig. 6. Few of their inhabitants survived in situ. Of the small numbers who saved themselves, most fled to nearby Polish towns and to the Crimea. Many fled only to lose their lives when the Cossacks pushed forward into Poland itself. Despite these catastrophes, the arguments adduced for the correlation of the described linguistic developments with historical events before 1648 may be salvaged if we assume further that the resettlement of the Ukraine, which proceeded at a moderate pace in the two decades that followed the massacres, was led by survivors returning to their places of origin, and that in this way the spatial relationship among the preexisting language patterns was reestablished. Thereafter, mass migrations from both the NEY and CY areas raised the Jewish population of the Ukraine to nearly 100,000 by the third quarter of the 18th century. 6 . INFLUENCES FROM THE NORTHEAST AND SOUTHWEST
Can we now determine the geographical extent of influences from the northeast and the southwest on the formation of SEY? As an indication of NEY influence, we can safely choose both the distribution of eji2 and the loss of phonemic vowel length. Loss of length, because of its aggressive nature, may well have arrived at its present location (as far to the southwest as the river Zbrucz) by means of normal diffusion along established lines of communication. On the other hand, eji2 is recessive and may more readily have reached its earlier geographical limits on the lips of NEY speakers, whether from northeastern Poland or Belorussia, who subsequently succumbed to the agressive features of SY phonology. As for the extent of influence from the southwest, we can infer it from the evidence described below. 6 . 1 . PROTO-VOWEL 3 4
SY alone filled the gap in the system of vocalic length oppositions that resulted from the raising of proto-Yiddish *a 12 (NEY o, SY u\cf. Fig. I). 10 Fig. 7 shows that the new a in SY, a result of the monophthongization of proto-EY *aj3i, is realized as CY /a/ and SEY /a/. The feature that serves to delimit CY within the SY area—aj 22 (cf. Fig. 1)—emerged subsequently. The distribution of SEY a 3 4 suggests the limits of direct influence from the CY area on the development in the Ukraine. We cannot say with 10
In NEY, the gap was never filled—testimony to the early loss of the length feature in that area.
YIDDISH IN THE UKRAINE
71
certainty whether the underlying a reached its limits as a result of migration or through a process of diffusion. Migration may be a more reasonable explanation for the bundling of isoglosses at this point (aj/a3i, ej/oji2, l/ej2s) but even if we choose diffusion we must assume the preexistence, in situ, of a vowel system receptive to the emergence of a new length distinction. This in itself implies ties to the west in the southern Ukraine that were absent in the north. 6 . 1 . 1 . MULTIPLE ORIGINS OF S E Y a3i. Comparison reveals a significant discrepancy between the distribution of a3l in Fig. 7 (malexl 'mouth
72
MARVIN I. HERZOG
[dimin.]') and its distribution in Fig. 8 {tax 'river'). Yet, it would be a mistake to assume that the cooccurence of tax and majlexl in the northern Ukraine indicates either the advance or the retreat of a3i. Upon closer examination of parallel developments, we would find that the wider distribution of a3i in Fig. 8 is the result of two separate developments that have meshed geographically. Thus, Fig. 9 delimits the area, largely in the northern Ukraine, where all diphthongs have undergone monophthongization before x and tax < tajx in the north has joined with tax < tax in the south to yield a continuous monophthongal area that goes well beyond the limits of the development described in §6.1.
YIDDISH IN THE UKRAINE
73
7. PROTO-VOWEL 5 4
We conclude with a discussion of the most complex of all Yiddish vowel developments. Figs. 10-14 illustrate the many regional variants of typical members of the word-class with StY oj5i: hojt 'skin', kojm 'hardly', zojer 'sour', etc. A brief description of each map follows: 7 . 1 . REGIONAL VARIANTS 7 . 1 . 1 . TYPICAL DEVELOPMENT. Fig. 10 ('skin') shows that hout occupies most of CY ; enclosed in this area, hot occurs in northeastern Poland. In
74
MARVIN I. HERZOG
the southern Ukraine—Bessarabia, Podolia, and eastern Moldavia—and at a single location in northeastern Poland (Ostrol^ka), we find hut. Moving northward, we encounter a fair sprinkling of hujt (along with some
hut and hout) in an area otherwise characterized by hojf, haut and an instance of heut occur in northern Latvia (Courland). 11 7.1.2. VOWEL 54 BEFORE 11
X.
Fig. 11 ('soup'). Here we emphasize the
At several northeastern locations, the word for 'skin' was not elicited. In these instances, we have incorporated variants of krojt 'cabbage' into Fig. 10. Similarly, where the word for 'soup' was not elicited, Fig. 11 includes variants of bojx 'belly'.
YIDDISH IN THE UKRAINE
75
extension of the southeastern u variant in jux (cf. bux 'stomach') beyond the u area in Fig. 10. In other words, jux in the northern Ukraine cooccurs with hojt/hujt 'skin'. The significance of a similar discrepancy was discussed in §6.1.1. (cf. Fig. 9).
7 . 1 . 3 . VOWEL 5 4 BEFORE /. Fig. 1 2 ('mouth') focuses on northeastern Poland where we find the variant mol. Previous examples would have led us to expect moul, mol, or mojl instead. For the discussion in §7.2.1. below, it is also relevant to note the two instances of mou that are adjacent to the mol area.
76
MARVIN I. HERZOG
Figure 12. Vowel 5 4 : Before / mot 'mouth' • O ic •
in hojt in hout in hot in hut
area area area area
'/ mou
Northeastern Poland 7.1.4. VOWEL 5 4 BEFORE m. Fig. 13 ('hardly') illustrates the predominance of monophthongal variants (primarily before -m but also before -/)—kam in the northeast, including much of the northern Ukraine; komfköm in the southern Ukraine and in northeastern Poland. (Cf. sam/som 'foam', ramen/romen 'to clean', etc.) 7 . 1 . 5 . VOWEL 5 4 PREVOCALIC. Fig. 1 4 ('sour') is of the southeast alone: zover in all of the Ukraine is matched by zaver in southern Belorussia. (Under the same conditions and in precisely the same area, o/w12 in noent 'near' yields novnt/nuvnt; eji2 in ejer 'ear' yields ever primarily in Belorussia but in part of the northern Ukraine as well.)
7 . 2 . HISTORICAL RECONSTRUCTION
Previous studies of vowel 54 have dealt largely with the monophthongal variant u. Faced with its discontinuous distribution, U. Weinreich reasoned (1958b: 260) that "we have here a case which requires polygenetic rather than monogenetic interpretation." With more adequate data on hand, we would like to argue, instead, for the monogenetic origin of all the variants of 54. 7 . 2 . 1 . EMERGENCE OF VARIANTS. Fig. 15 is an attempt to diagram the differentiation of the variants of vowel 54 from a common origin—protoEY*ÖW. Each box represents the entire E Y area while the order of the boxes specifies a chronological sequence.
YIDDISH IN THE UKRAINE
77
Box 1 represents the earliest stage when *o12 in *hoz 'hare' and *au5i in *hauz 'house' were coextensive in EY. 12 Box 2 suggests that in the differentiation of proto-SY, the parent dialect of CY and SEY, first *o12 > u, then *auSi was raised to ou. Both changes spread eastward in that order while in the east itself, *o12 and au5i remained unchanged. ia For a reconstruction of proto-EY see Herzog (1965a: 164). Note that final z in *hdz and *houz is a morphophonemic notation. Word-final voicing, though distinctive in NEY, SEY, and the standard language, is neutralized in CY.
78
MARVIN I. HERZOG
Box 3 describes a development in the east where the articulation of the u-offglide of au/ouSi is assimilated to the articulation of a following labial m: kaum\koum 'hardly' > kSmjkdm.
Figure 14. Vowel 5 4 ' Prevocalic 'sour'
• zover \ zaver o zuver
YIDDISH IN THE UKRAINE
79
FIGURE 15. THE DEVELOPMENT OF VOWEL 54
Box 4 refers to the southeast only. Before vowels, the off-glides of vowels 12 and 54 are spirantized: zauer6i 'sour' >zaver; noentl2 [noant] 'near'> novnt. Box 5 illustrates a major change in the northeast that sets it apart from the southeast: vowel length is lost and long and short vowels merge; the vowel of hdz12 'hare' merges with the vowel of hdttl '(he) has'; zokn12 'old man' and zdkniX 'socks' become indistinguishable. Box 6 applies to northeastern Poland alone: ou s4 is reinterpreted as o (houz 'house' > hoz).
80
MARVIN I. HERZOG
Boxes 7 and 8 require a reminder that the earliest changes (cf. Box 2) proceeded in a fixed order: first *o12>ii, then *au5i>ou. Box 7 now suggests a reversal when these innovations reached the east: first the raising of *a«54 to ou (or o) penetrated the eastern area with the following results: (i) In the northeast where length had already been eliminated either completely or in part, *au5i > ou except in northern Latvia where auS4 remains (or was reinstituted under local German influence). (ii) In the southern Ukraine, where long and short vowels were still distinct, *au5i was replaced by ou or o (perhaps imported directly from northeastern Poland, cf. Box 6) and merged with o12; thus hozBi 'house'= hdz12 'hare'. Subsequently (cf. Box 8), the change of *d12 to m reaches the east. In the northeast, where the loss of length has already led to the merger of *o12 and *oil (cf. Box 5), only ou5i is affected (houz 'house' >huyz; joux 'soup' >juux). In the southeast, however, where vowels 12 and 54 have merged (cf. Box 7) the total merger is affected by its encounter with the raising of *o12: hoz 'house', 'hare' > huz. Box 9 describes a development that appears to be restricted to the southeast (though further investigation may establish its relationship to the entire SY area). Before length was lost in the southeast, *&1X in h&nt 'hand', mdn 'husband', malexl 'angel (dimin.), etc.,> 5, the variant d5i also became 5 (cf. k&m 'hardly', zdver 'sour' in Boxes 3, 4). Box 10 indicates that the absence of vowel length in the northeast led to the reinterpretation of huyzSi 'house' usually as hujz but also as huz. (In the latter instance there were mergers with u 51 , S2 ; thus bux6i 'stomach'= bux52 'book'.) The loss of length in the southeast effected no new mergers. Box 11 suggests the effects of contact between hujz5i and houzsi after the monophthongization of uj before x in the northern Ukraine (cf. Fig. 9). An intermediate variant hojz emerged—perhaps in the area of BrestLitovsk (cf. U. Weinreich 1958b: 261). Box 12 shows that oj5i (supported, incidentally, by the standard language) 13 diffuses in all directions at the expense of other variants. Another variant—o54 in mol 'mouth'—in northeastern Poland requires some explanation, although its place in the chronology seems unimportant. The key to its development appears to be the velar quality of / in the area. In two of the sample communities (cf. Fig. 12), nonpalatal / has been 13 Standard orthoepy has, generally speaking, not been a very momentous force in Yiddish. It does appear, however, that anything but oj for vowel 54 has been disparaged.
YIDDISH IN THE UKRAINE
81
completely vocalized. Every sequence of vowel plus / is realized as a diphthong with a u-offglide: gel 'yellow' > geu and mol 'times' > mou. Similarly, every sequence of a «-diphthong plus / is reduced to a diphthong: mouI 'mouth' > mou. This must have been widespread in northeastern Poland where the subsequent restoration of / reduced «-diphthongs to vowel plus /. In part of the area, 'times' and 'mouth' are therefore rendered alike, mol. Hypercorrections like jolx 'soup' (cf. Fig. 11), bolx 'stomach', zolber 'clean' and polk 'drum' are not infrequent. 7 . 2 . 2 . FURTHER PROBLEMS. The foregoing account is by no means complete. It makes no mention of the absence of mol 'mouth' in western Poland where / has also been vocalized. It fails to account for the island of «54 in northeastern Poland (cf. Figs. 10, 11)—probably a late development (see Herzog 1965a: 217). It offers no explanation for the absence of an abi variant (before m) of StY flojm 'plum' (cf. kam, Fig. 13), nor does it attempt to explain the prevalence of af5i 'upon' in the southeast where we have reason to expect of (cf. Fig. 15, Box 19). It seemed preferable to set these unanswered questions aside for future consideration.
THE GEOGRAPHIC MAKEUP OF BELORUSSIAN
YIDDISH1
URIEL WEINREICH
1. After decades of preoccupation with a valid classification of the dialects, Yiddish dialectology has recently turned to a different major goal. At the head of our current agenda is the characterization of dialectal regions and the explanation of their present-day makeup in terms of migrating populations and radiating innovations. It is by no means hard to understand the long-time preoccupation of scholars with classifying dialects by criteria drawn from historical phonology. This interest grew not only out of the specific kind of training received by a generation of Yiddish linguists, but also out of the nature of the material. The strikingly regular divergences among Yiddish regional sound systems (particularly the vowels) have been, and remain to this day, irresistible grist for the structuralist mill. It is the author's view, however, that a classification of dialects based on their phonologies is nothing but a classification of their phonologies; it is a typological exercise as applicable to genetically unrelated dialects as to the varieties of a single language. The right of phonology to serve as the privileged basis of dialect classification is still to be established. This right cannot depend, in any case, on the descriptive tractability of the phonological material. It will only be established if phonological differences are shown by detailed sociolinguistic research to have overriding value as indices of subgroups in dialectally heterogeneous (i.e. real) speech communities. Whether phonological indices do indeed function in this way is a completely open empirical question, and the slowly accumulating literature seems to contain few indications that the issue will necessarily be resolved in favor of phonology. While this question remains pending, research has turned to the analysis 1 In an abridged form this paper was presented to the Conference on Yiddish Dialectology in New York, June 13,1965. A slightly revised Russian version was prepared for a volume of papers in honor of Academician Konrad Krapiva, to be published by the BSSR Academy of Sciences in Minsk.
82
GEOGRAPHIC MAKEUP OF BELORUSSIAN YIDDISH
83
of geographic dialect regions as results of intersecting social and cultural forces. In the search for significant geographic configurations, no special priority is assigned to the "top" of the hierarchically organized structure postulated for language. On the contrary, recent dialectological work has been turning up patternings, for Yiddish no less than for other cases, in domains of language (irregular morphophonemics, vocabulary) that are not particularly highly structured in terms of intradialectal oppositions or rules.2 The present treatment of Belorussian Yiddish (BrY) may be thought of as an addition to a growing series of relatively successful analyses of such geographic patternings in Yiddish: the region south of the Carpathian Mountains (U. Weinreich 1964b), the area of northern Poland (Herzog 1965a), and certain aspects of Yiddish in the Ukraine (Herzog 1968; 1969: this volume 58-81). In the long run, such investigations will not only throw light on the concrete history of the Yiddish language, but will help us to understand the unequal "diffusability" of various types of innovation (cf. Herzog 1965a: 5) on an empirical basis, rather than in the a priori manner that has been prevalent since the phonological successes achieved by comparative linguistics. Although no region of the Yiddish speech territory deserves to be omitted, the Belorussian area makes a particularly attractive object of study. Its antiquity contributes to its importance, for the history of Yiddish on the territory of the present-day Belorussian Soviet Socialist Republic goes back to the very beginning of Ashkenazic settlement of the Grand Duchy of Lithuania. In Brest and Grodno, Jewish communities were founded no later than in the 14th century,3 and ever since Yiddish speakers continued to increase in number and to expand eastward and northward. Towards the end of the 18th century the Jews already constituted significant proportions of the population of cities and towns in almost every corner of Belorussia, and in many a town Yiddish was the language of the majority. In addition, we must reckon with a certain dispersion of Jews in the villages. Even if the Jews had not played an important part in the growth of trade and crafts and in the modern social movements, the intimacy of their coexistence with a Slavic population would endow this area with special interest as a case study in "bilingual dialectology"—the comparative analysis of geographic fragmentation of coterritorial speech communities.4 For the study of Yiddish dialect regions, uniform materials are now being 2
See Wolf (1969: this volume 102-215). The same period also saw the rise of Jewish communities in Troki and Luck, which lie outside the borders of contemporary Belorussia. See Shiper (1914). 4 See e.g. U. Weinreich (1962a). 3
84
URIEL WEINREICH
collected under the auspices of the Language and Cultural Atlas of Ashkenazic Jewry (hereinafter LCAAJ).5 Compared to the rest of the Yiddish language territory, its easternmost segment possesses the additional research advantage of having been covered by the Yiddish Atlas of the USSR (Vilenkin 1931),6 which included the eastern (pre-1939) portions of Belorussia and the Ukraine. Of these two regions—at least from the point of view of bilingual dialectology—Belorussia takes a certain precedence over the Ukraine, inasmuch as the cartographic coverage of Belorussian dialects is further advanced: we are fortunate today in being able to consult the published Dialectological Atlas of the Belorussian Language (hereinafter DABM). 2. In the period of Yiddish dialectology in which criteria of historical phonology were given priority, "Belorussian Yiddish" 7 hardly emerged as an independent topic of discussion. Phonological criteria rather yielded a broad "Northeastern" dialect, in which a sample phrase such as 'large garden' corresponds to grejser gortn, in opposition to the "Southern" Yiddish of Poland and the Ukraine, where the same phrase appears as grojser gurtn. (We are concerned here with the proto-Yiddish vowels8 corresponding to MHG o and a.) In fact, there are further specific features of Northeastern Yiddish (NEY), to be found not only in its vocalism, but also in its consonantism (the merger of the opposition of hushing and hissing sounds, an opposition which has been only partially restored—cf. U. Weinreich 1952; 1964a:348, 353ff.). The grammar of this dialect is characterized by a radical reconstruction of the gender system and of the morphology of nouns in general (Wolf 1969). In vocabulary, too, there are NEY peculiarities, such as holt hobn\\lib hobn 'to love', dekl\\stercl, pdkriske 'pot cover,' etc. (The first named form is from NEY.) Thus we see that the well-known vocalic features distinguishing NEY can be supplemented by other, even nonphonological ones. 6 For further information on the Atlas, see U. Wcinreich (1962a, 1962b); Herzog (1965a:6-10). 6 As noted by 2irmunskij (1939), this work had the distinction of being the first dialect atlas to be published in the Soviet Union, if one leaves out of account the far more sketchy monograph by Buzuk (1928). Because of the fine-grained network used by Vilenkin and his opportunity for verifying all findings in situ, his results will by no means be superseded by the LCAAJ, which uses emigrant informants and a thinner grid. On the other hand, the LCAAJ has the advantages of direct questioning of informants (rather than by correspondence) and of covering a great deal more than the phonology. 7 Belorussia is distinguished by a traditional Yiddish name of its own—rajsn. The old Yiddish term lite includes ethnographic Lithuania as well as Belorussia. a On the reconstruction of the proto-Yiddish vowel system, see M. Weinreich (1960a).
GEOGRAPHIC MAKEUP OF BELORUSSIAN YIDDISH
85
For the internal subdivision of NEY the vocalism is far less serviceable. But if we are willing to start in other domains of the language, we find that, for all its unity vis-à-vis the outside, the Jewish Northeast is nevertheless cut up by numerous inner boundaries. On Map 1, where NEY is set off
from its southern neighbors by isoglosses I (gortn\\gurtn) and II (grejser\\ grojser), we also find isoglosses which subdivide the NEY massif into a western, "Lithuanian," and an eastern, "Belorussian" half. The latter at the same time set off NEY from Southern Yiddish (SY), although in the northern Ukraine they fan out here and there. For the isoglosses enumerated below, the first form mentioned belongs to BrY 1—s'iz mir tut-vej\ sHut mir vej 'my . . . hurts'; 2—voncn\\vonces 'mustaches'; 3—behejmes\\ ki(en) 'cows'; 4—mojl 'mouth' in the masculine ¡feminine, neuter genders; 5—teler 'plate' in the feminine||masculine gender; 6—fencter 'window' in the feminine || masculine or neuter genders. For the sake of completeness it should be added that LCAAJ materials
86
URIEL WEINREICH
also reveal certain other subdivisions of the territory under discussion. For example, one is struck by a configuration of isoglosses which separates innovations in the center from relics along the northern boundary (cf. the replacement of «-diphthongs by new /-diphthongs in words of the type moul\\mojl 'mouth'; the replacement of the form gor 'Hanukkah spinning top' by the designation drejdl; and others). In other cases, there appears a vestigial area in the extreme northwest (cf. y- and «-diphthongs in the words of the type greyser, greuser 'large'). Sometimes, on the other hand, it is Courland and Samogitia that are marked by innovations (bud 'store, shop' instead of the general NEY krom; bord 'beard' in the masculine, instead of the general Yiddish feminine gender).9 Moreover, BrY as we have defined it (see Map 1) is in turn cut by various isoglosses. However, we are not concerned here with establishing an authoritative classification of dialects and subdialects. As suggested in §1, we wish rather to sort out observable innovations according to their geographic distributions, and then to sketch the culture-historical forces which brought the perimeters of a large number of innovations to a halt at such and such lines or in such and such areas—wherever we now find the pattern of isoglosses. For the present we will bypass the delimitation of BrY from its neighboring dialects (Southeastern Yiddish [SEY] and Central Yiddish [CY]), and will concentrate on isoglosses which traverse the NEY massif itself. 3. It must be remarked that the temporal frame in which the diversification of BrY took place is a relatively narrow one—considerably narrower than the antiquity of Jewish settlement on Belorussian territory would in itself suggest. This becomes clear from an examination of Map 2. The small dots here signify Jewish communities founded before 1650; the larger dots represent communities which appeared between 1671 and 1775. Obviously, Jewish communities already existed before 1650 in the major commercial centers of the Grand Duchy: in Vitebsk, Polock, Vilna, etc. However, dense Jewish settlements were to be found only in four provinces (voevodstva), corresponding approximately to "Black Russia" (Cernaja Rus'): in the provinces of Brest and Novogrudok and in the southern portions of those of Troki and Vilna. It is up to the historian to explain why the wave of Jewish migration from the west avoided the larger part of Belorussia up to 1650 while at the same time it flowed so intensively into the Ukraine. Whatever the explanation, it appears that in response to the Cossack massacres of Ukrainian Jewry in 1648-49 the Jewish migratory stream turned northeastward. And so, in the subsequent century and a quarter 9
For details, see Wolf (1969).
87
GEOGRAPHIC MAKEUP OF BELORUSSIAN YIDDISH Map 2
YEARS OF FOUNDATION OF JEWISH COMMUNITIES •
up to 1650
o
1667-1675
O
1676-1700
•
1701-1725
•
1726-1750
•
1751-1775
'VDvinsk _ CT-v—.
-a.
•>
•
S
'Vitebsk
•
'S. i f \ • I \ *J * I Novognidok - - \ • ;
.K
• Mogilev
(
•Bobrujskv
v _ __
# .Mstislavl' U • s
i
\ \ \
'Silicic
I
\
à .-•I / \
"Minsk
.—1 .\
V
o o
•
/
^
•i—r
•V
v—' V
./v'-t —
«
—--——
boundaries of the Polish Staie boundaries of provinces
I \
\
JL
(i.e. up to the partition of Poland), the provinces of Vilna, Minsk, Vitebsk, Polockand Mstislavl', i.e. "White Russia" proper, came to be dotted with Jewish settlements. This chronological scheme forces us to the conclusion that, until the eastward movement of the Jews halted altogether, i.e. until the middle of the 18th century, 10 the isoglosses that run through present-day BrY could not yet have existed. In this connection, Map 3, 11 which shows the Yiddish designations for 'duck', is particularly instructive. Of the two forms, entl and kacke, the former is the more ancient (cf. Western Yiddish [WY] ant [U. Weinreich 1962a: Map 2], M H G ente). The form entl could not have 10
As Map 2 shows, the rate at which new communities were founded fell after 1750, so that we may be permitted to assume the middle of the 18th century as the end of the main settlement stream. 11 The numbers on Map 3 identify the birthplaces of our informants. For greater perspicuity, these numbers are not repeated on Maps 4ff. The locations can, however, easily be checked by juxtaposition with Map 3.
88
URIEL WEINREICH
" W
Map 3
'DUCK' }
A Daugavpils f
• •
Sa Vilnius A-
in Yiddish
•
no answer
s
®\
)
"A
N
Polack.
fi
"A j
A^
'._/"w Maladzeina
"V
A " , MINSK T•hi,-»r A«
j Hrodna —1 ID
L
Mahilew *
Navahrudak -
in Belorussian
vutka kilka
if
21
J
I
0 kacke
A
/
AJ
*
• Biafystok
5
Sluck.
v \
both designations
S k" 1
V I Brest
V..-'
f
/ V Babrujsk
i S,
Hornel'
•'0
v
Pinsk^l
L
Mazyr •
s
u
©
• Cernihiv
arisen independently in eastern Belorussia; consequently, it must have been imported; in fact, it most likely appeared with the arrival of the emigrants from the Brest-Novogrudok-Troki area between 1650 and 1750. But if in western Belorussia at that time the form kacke had already been prevalent, the emigrants would have exported it with them, and there would be no entl in the east today. For this reason we conclude that up to the middle of the 18th century, entl was the dominant form in the west as well. One could, of course, suppose that even before 1650 entl was in the vocabulary of the original Jewish settlers in the major eastern urban centers, and that it subsequently diffused from there among the newly established communities of eastern Belorussia, displacing the imported kacke. It is hardly probable, however, that the old communities, few in number and poor in population, were powerful enough to impose their imprint on the later and stronger stream of migrants. Still less probable is a third hypothesis, according to which the original synonymy kacke=entl was replaced throughout the length and breadth of Belorussia by a later territorial
r-
GEOGRAPHIC MAKEUP OF BELORUSSIAN YIDDISH
89
specialization of designations, kacke\\entl. The argument, at first blush quite seductive, that Yiddish kacke could be based on the Belorussian (BR) dialectal kacka,12 is negated by the fact that the eastern BR vutka in turn could hardly have been the basis for Yiddish entl, despite the remote etymological kinship of these two words: we cannot attribute such comparativistic farsightedness to the average members of a speech community. Thus we claim that up to the middle of the 18 th century, BrY was either uniform or chaotically mixed. There could not yet have been any territorial subdivisions. We can adopt that time as the terminus a quo for the crystallization of the modern internal geographic makeup of BrY. We will now attempt to analyze the subdivision of BrY under three headings : influences from the south (§4); influences from the southwest (§5); and differentiation in situ (§6). 4. When we examine Map 4, we notice a step-like series of Yiddish "horizontal" isoglosses. (We will return to the Belorussian isoglosses of this map in §8.) These lines have been transferred to our map from Vilenkin (1931) and therefore end at the Soviet-Polish border in its pre-World War II location. They are as follows: 13 1—z\z before stressed i (e.g. zilber 'silver'); 2—unstressed em\\om in world-final position (e.g. ojlem 'public'); 3—consistent appearance of o|| alternations with oj, uv in the word grojer 'gray'; 4—vestiges of «¡consistent oj in words of the type bojx 'stomach'; 5—unstressed ex\\ax in word-final position (e.g. lajlax 'bed sheet'); 6—z\\z before e, i in post-stress syllables (e.g. buzem 'bosom'); 7—o\\oj in the word blojer 'blue'; 8—retention||loss of av in the word geddjern 'to last' (i.e. gedavern); 9—vestiges of ay ]| vestiges of ov in the word zojer 'sour'. In each of these cases the southern form in BrY agrees with the form common in Ukrainian Yiddish (UkrY) or, at least, in its northernmost varieties. In the majority of cases (1,4, 5, 6, 8) the southern forms also coincide with those prevalent in StY. It is hard to believe that forms contradicting the literary norm could, in the period under discussion, have diffused in a southerly direction from some unknown center in the north. It is far more likely that the forms which agree with the norm should have diffused from the south—from the Kiev area and from Volhynia. In the adjacent Yiddish dialect of the Ukraine we encounter similar phenomena of radiation from the south to the north. In the territory where 12
DABM (297) shows the widespread distribution of kacka in western Belorussia. The sources of isoglosses 1, 2, 3, 4, 5, 6, 7, 8, 9, are Maps 68, 59, 33, 38, 60, 67, 34, 41, 40, respectively, in Vilenkin (1931). The first named form in each case refers to the north. The lexical examples are identified in their standard Yiddish (StY) forms. 13
90
URIEL WEINREICH
Vilenkin (1931) shows the prevalence of the pronunciation grojser 'large' (cf. our Map 1), we find a strip of ex 'also' ( < ejx; StY ojx). The location of this strip suggests that here an earlier dialectal ej was displaced by southern oj, except before x. There is another suggestive parallel. In a fairly wide strip in the northern Ukraine we find -ix-, -ir- as against southern -ex-, -erand as against northern -ux-, -ur-; sample words are pex\\pix\\pux 'down feathers', kerc\\kirc\\kurc 'short' (see M. Weinreich 1965). In contexts other than before r, x, we have u in the same area as ur, ux, and i in the same areas as ir, er, ix, and ex. There is little doubt that the forms in u are the most conservative. Hence we appear to be dealing here with an area in which an original u was in relatively recent times engulfed—"precipitously," as it were—by a southern-derived i taken as the general equivalent of u, and overlooking the special conditioning role of velar continuants which produced the "regular" transition ux, ur>(ix, ir>) ex, er. We propose to regard as instances of the same process both groups of radiations, namely: (1) innovations proceeding from the south within
GEOGRAPHIC MAKEUP OF BELORUSSIAN YIDDISH
91
Ukrainian Yiddish, but stopping short of BrY; and (2) innovations which flowed north in UkrY and penetrated into BrY, being reflected there in the step-like series of isoglosses shown on Map 4. Lexical examples would be welcome at this point. Perhaps what Map 5
92
URIEL WEINREICH
shows us is an instance of the same emanation from south to north in its effect on a particular vocabulary item. No less than nine times do we find the borrowed form stol'e 'ceiling' in Yiddish to the north of the area where it is at present current in Belorussian dialects. Still more striking are the locations for which Yiddish informants reported the form stel'e. These latter forms seem to show particularly long-distance northward penetrations in Yiddish in comparison with the Slavic speech area in which they originated. 5. The second geographic process which we can observe on the basis of the dialectological material is the influence from the southwest, proceeding from the centers Brest-Novogrudok-Grodno. As we have already indicated (§3), this region was the cradle of BrY. Apparently it continued to furnish linguistic innovations for all of Jewish Belorussia even after it ceased to send out emigrants. Incontrovertible examples of such southwestern influence, based on
GEOGRAPHIC MAKEUP OF BELORUSSIAN YIDDISH
93
Vilenkin (1931), are represented on Map 6.14 Isogloss 1 is the northeastern front of the restoration of the c: c opposition (as against a single affricate); isogloss 2 is the corresponding front of the s:s restoration. The Yiddish confusion of hissing and hushing sounds, which is here shown disappearing before our eyes, apparently arose, in accordance with the discussion in §3, in the southwestern corner of Belorussia in a sufficiently early period to be able to be diffused throughout the entire Grand Duchy of Lithuania. Later on, it would seem, the same area proceeded to restore the opposition between the hushing and hissing consonants. In the most recent period this restoration gained the support of the literary norm and came to expand not on an areal, but on a social basis;15 but Vilenkin's maps still depict the maximal reach of the spatial expansion. The form gix 'rapid' in place of the older gex (cf. MHG gäch, modern German jäh) is delimited in the northeast by isogloss 3; it, too, constitutes an innovation with a similar geographic distribution. (Concerning the Belorussian isogloss bundle a-b, see §8.) The dynamics and the relative recency of the dialectal process in BrY manifest themselves even more vividly on Map 7.16 As Zirmunskij concluded years ago (1939:143), the old w-diphthongs (proto-Yiddish u 4 ) in words of the type mojl 'mouth' were, in some parts of Belorussia, displaced first by uj, and later by oj. Map 7 shows that the shift uj -> oj (isogloss II) in the north did not catch up with the shift (o)u -» uj (isogloss I); in the south, on the other hand, the uj oj shift overtook the earlier (o)u -> uj shift so completely that no «/-interstice between the old (o)u zone and the very recent oj zone is visible. Map 7 also displays some vestigial "pockets" in which these shifts had not yet, at the end of the 1920's, encompassed all the words to which they were in principle applicable. For nonphonological parallels to this geographic process, we may reexamine Map 3. The lexical innovation shown there—the form kacke in place of the older entl—has approximately the same center and the same radius as the phonological innovations depicted on Maps 6 and 7. It is to be expected that the full materials of the LCAAJ will contain many additional grammatical and lexical examples of the same radiation. In comparison with the radiation from the south discussed in §4, the influence from the southwest is both more recent and more distinctive. Map 40 in Vilenkin (1931) allows us to unravel the complex picture and to arrange the areas displayed from the point of view of relative chronology. 14 The sources of isoglosses 1, 2 and 3 are Maps 69, 71, and 25, respectively, in Vilenkin (1931). 15 See U. Weinreich (1952:374) and the literature cited there. 16 The map is based on Vilenkin (1931: Maps 35-41).
94
URIEL WEINREICH
Without a doubt, for example, the areal split between u and oj in the word zojer 'sour' is superimposed on an earlier series zover\\zaver\\zoer, zuer. In a case like this the question about the temporal order of the linguistic phenomena (cf. §3) becomes even more intriguing. Within the limits of 200 years we apparently have to establish two subperiods: an earlier one, in which the influence from the south was dominant, and a later one, in which the leading role passed to the southwest. With regard to Map 3, another important question arises. As the shading indicates, the western dialects of the Belorussian language also contain the word kacka. Although the boundaries of the corresponding word areas do not coincide in every detail, the congruence of the Yiddish and Belorussian isoglosses is in general truly remarkable. (One must not forget that the Belorussian materials are more recent by 20 or 25 years than the Yiddish ones, and reflect a rural rather than urban milieu.) Assuming that both the Belorussian and the Yiddish forms ultimately go back to the Polish
GEOGRAPHIC MAKEUP OF BELORUSSIAN YIDDISH
95
kaczka, we still face this question: did Yiddish borrow this word directly from Polish,17 or did it penetrate into Yiddish via Belorussian? Map 3 by itself will not resolve this puzzle; we approach a solution of it, however, by means of Map 8, which displays the dialectal designations of
in Yiddish H
padloge
O Mk Pol A
podlege in Belorussian most padldha ~;
pamAst (sporadically)
'floor'. At first blush this map is no less ambiguous than Map 3, since here, too, the distribution of the parallel innovations in Belorussian and in Yiddish is almost identical. It is remarkable, however, that the Yiddish forms have been recorded exclusively with g, while the Belorussian forms in all likelihood are pronounced with [y].18 As a rule Belorussian poststress intervocalic [y] is rendered in Yiddish lexical borrowings as h, j, or zero (cf. smdha 'dried spittle' > smahe, smaje; kryha 'ice floe' > krie, and others). 17
In the Yiddish of Poland, 'duck' is also kaike\ cf. U. Weinreich (1962a: Map 2). Avanesov (1964:104) mentions the fricative pronunciation of the phoneme among the features common to all Belorussian dialects (except for some few "foreign" words). 16
96
URIEL WEINREICH
If the designation of 'floor' in western BrY had come from the local Belorussian dialect, it would have been rendered as *padlohe, *padloje. Since this is not the case, we must conclude that the Yiddish word originated in Polish. On the other hand, one cannot fail to notice the pretonic a; it is as if the Yiddish speakers in Belorussia had submitted to a:d neutralization (akan'e) every distinctive Slavicism, regardless of its actual origin, and in spite of the existing opposition within Yiddish, of pretonic a: o (e.g. xazejrim 'pigs': xolile 'God forbid'). The probable Polish origin of Yid. kacke does not imply a total denial of Belorussian influence; such influence was definitely exerted, and contributed to the formation of the geographic makeup of Yiddish. (Examples will be cited in §6.) The influence, however, was never wholesale: every borrowing, as was already seen in connection with Map 5, has its own history. 19 The analysis of Map 8 proves that even with regard to Polish influence, to which both Yiddish and Belorussian were subjected, it is necessary to discover the separate channels of communication with the west that prevailed in each speech community. We will come to a similar conclusion in connection with the problem of Russian influence (§8). We have discussed the phonetic shape of the Yiddish designation of 'floor' current in western Belorussia—padloge. But Map 8 poses yet another riddle: the origin of brik, the corresponding Yiddish synonym in eastern Belorussia. It is tempting to regard this word as a local caique on most, that is as an expansion of the meaning 'bridge over a river' on the model of the Belorussian word.20 However, brik in its meaning 'floor' exists also in the Yiddish of the western Ukraine and in Rumania (cf. U. Weinreich 1960:51, map). One would therefore have to reconstruct two stages for Yiddish in East Slavic territory: the first, in which brik arose as a caique on the model of Ukr. mist, BR most; and the second, in which the Polishpodloga expanded in Yiddish as well as in the East Slavic languages. If the proposed chronological scheme for Yiddish is correct, the Yiddish material may turn out to be useful in the dating of parallel processes in the East Slavic languages, too. Let us observe in passing that the question of contacts between NEY and the Polish language is far from simple. Actually our maps display two types of Polish impact on NEY: (1) "urban" Polish words penetrating Yiddish by leaps and bounds, without subjugating any continuous terri19
Similarly unpredictable is the specific geographic distribution of Yiddish forms of Hebrew-Aramaic stock; see now U. Weinreich (1965b). 20 It is only by a calquing mechanism that we can explain the appearance, in eastern Belorussia, of Yid. barg 'attic' (.
o o « 3 ?
A
fem, there seems little reason to consider them new neuters. It would seem rather that in these cases, as in the case of grub 'pit' (4.13, Map 4:21), the attested MHG genders cannot be projected to an earlier stage of Yiddish. As we pointed out in 1.3, the description of gender variation in Yiddish is limited. Consequently each case must be dealt with separately. The essential shallowness of our present analysis, though unavoidable, deters us from attempting to fix upon a particular gender as historical without reasonable geographic support. Even then we emphasize the very tentative nature of the gender assignment in the absence of historical investigation.50 If our interpretation is correct, neither rejnkejt nor zamd can serve as examples of the process adduced by Herzog, namely, that new neuters were semantically motivated, i.e. that they emerged in order to bring certain mass nouns within the scope of the interdialectal equation: Northeastern di (mass) = Southern dos.51 Two further maps deserve consideration in this connection. Map 4:47 (stroj 'straw') would be a candidate for a new neuter if its gender history were not so obscure. The near universality of the feminine variant in Map 4:47 might suggest that the neuter is an innovation, but the distribution displayed there can also be interpreted as the result of the further southward penetration of the feminine form from what was once a Pattern II distribution. If so, we have reason to believe that the neuter variant is historical and the feminine an innovation, in accordance with the argument 50
Old and Middle Yiddish texts are still largely unexplored. Moreover derivatives in -kejt are true feminine in NEY, not mass. Even had the gender of such derivatives changed from di to dos in SY, it could not, a fortiori, be the result of the process suggested by Herzog (1965a: 118). 51
YIDDISH CASE AND GENDER VARIATION
7*
193
194
MEYER WOLF
at the beginning of the section. (The gender attested for MHG is also neuter, which perhaps supports this possibility.52) Had the information for Map 4:48 (ves 'laundry') been more complete,
52 The relationship of the gender of MHG Streu (fem.) 'litter' to Stroj 'straw' is as obscure as that of MHG grap to grub in 4.13.
YIDDISH CASE AND GENDER VARIATION
195
we might have seized on this map as a clear example of a new neuter. The large compact regions of the feminine in the Northeast, the Southeast, and Southwest suggest that responses from locations where the form was not elicited would reinforce the impression that the neuter instances of ves are typical of the border region between the mass and nonmass areas; (cf.
196
MEYER WOLF
Map 2:8). Border speech is frequently subject to hypercorrection. Here the opposition SY neut/NEY fem leads border speakers to change to neuter a word which appears to be universally feminine. This process is similar to the phonological hypercorrection that leads speakers of SY to create a form lutvak 'Lithuanian Jew' from the universal litvak on the basis of an "overlearned" correspondence, SY i/NEY u. This one instance, however, hardly permits us to determine whether the phenomenon is restricted to mass nouns, i.e. the change is semantically motivated, or whether it extends to all di nouns, i.e. the change is purely formal. 53 In view of the difficulties which continue to arise in considering the gender history of nouns, it seems advisable again to call into question the practice of relying on attested MHG genders as evidence for the historical genders of Yiddish. The danger of depending too strongly on a MHG dictionary for phonological and etymological purposes has been emphasized by M. Weinreich (1954:73); it appears that equal care must be taken with respect to gender, particularly since the study of gender variation both in standard German and in the German dialects has barely begun. 2irmunskij's discussion (1962:443ff.) shows how preliminary is the knowledge of gender systems in the German dialects and Kaiser's study of gender variation in the first half of the eighteenth century (1930:181ff.) is unmatched by studies of later periods, and reveals how little is known about variation in standard German. Kaiser's work indicates that clarification of the history of Yiddish gender is in part dependent on the knowledge of gender variation in dialectal German. The fact that he lists neuter variants for nouns such as Butter, Honig, Tau and Sand (none of them neuter in standard German) suggests that the very origin of the mass gender (though not its form) may be related to a similar phenomenon in German, a tendency to transfer mass nouns to a single gender. At the very least, an understanding of gender variation in dialectal German would permit us to distinguish specific Yiddish gender developments from those pre-Yiddish ones that were inherited by Yiddish but failed to affect standard German, either in its modern or its medieval form. 54 4 . 3 2 . HISTORICAL NEUTERS
For the following words, there is nothing to suggest that their historical 53
See 4.32.4 for further consideration of this question. A number of words in Kaiser's list are accompanied by comments that in "Oberdeutsch" the gender is "schwankend", "ungewiss" or shows more than one variant, e.g. Eck, Fahn, Gesang. One wonders if the dialects from which the German component of Yiddish is drawn differed extensively among themselves in regard to gender, or even if each showed a considerable amount of gender fluctuation. 64
rv
«Riga 3
BALTIC
197
YIDDISH CASE AND GENDER VARIATION
Liepaja
SEA
. Daugavpils
Smolensk < prtJPj*.
0 Biafystok
\
Mass
N]èut~ lass \
frue Fem~ Mass
Map 4 : 49 vaser water':
Summary of Gender Variants
• True Fem
xNeut
o BLACK SEA
gender is other than neuter. A brief discussion of each may reveal some of the processes involved in the retreat of neuter forms. 4.32.1. vaser 'water'. Map 4:49 is a summary map showing regional variation in the gender of the historical neuter vaser. The fundamental division on the map is that of a Pattern II neuter (neuter in the Southwest,
198
MEYER WOLF
nonneuter elsewhere), but with a rather more solid incursion of nonneuter forms into the Southwest. Of special interest here is the regional evolution of a true feminine form in the Southeast. Thus, in the area labeled Mass we find the normal mass gender paradigm di vaser 'the water '//« kalte vaser 'in cold water'; in the area labeled Neut we find the normal neuter paradigm dos vaserI in kalt vaser; in the area labeled Neut ~ Mass we find fluctuation between these two paradigms. However in the area marked True Fem ~ Mass, we also find variation of the type in kalte vaser ~ in kalter vaser. The adjectival inflection -r is of course the marker of the true feminine dative in the Northeast and the feminine dative elsewhere. This is the only case which has so far been recorded of a mass noun being transferred into the true feminine in the three-gender area. In general the mass noun is taken over with all its inflectional peculiarities, as shown, for instance, by the form of the variation in the area marked Neut ~ Mass. It seems curious that instances of true feminines from penetrating mass gender forms should be so rare; one could expect them as the primary way of assimilating new di forms to the three-gender system. 4.32.2. kind 'child'. As Herzog (1965a: 135) has pointed out, the noun kind is of great importance to the history of the neuter, since it is one of the few words showing traces of a former neuter article in the Northeast, in the now idiomatic di skind 'the child'. Map 4:50 is a summary of the nonneuter variants of kind in three nominative contexts. Although, on the whole, masculine and feminine variants seem randomly distributed, a predominance of ¿//-forms is discernible in the East-Northeast. This predominance whose significance we will touch on below, does not reappear on Map 4:51. On the contrary, Map 4:51 indicates that in accusative contexts, the feminine variant of kind is extremely rare. On this map we also see three instances of the interesting di skind.55 Herzog (1965a: 135) suggests that the observed nonneuter forms can be traced to an original s kind a common allegro variant of dos kind: where the s loses its function completely, we have di skind. . . and di kind. Masculine der kind seems to be the result of the following series of changes: dos kind (nom., acc,)/dem (dat.) > dis kind/dem > di (s)kindjdem. The dative dem of this anomalous paradigm may then have provided the basis for the analogical reinterpretation of the nominative as de r. 55 Mark (1951:451) reports that di skind has a connotation of particular tenderness. Such a specialized usage suggests that the form is petrified. The distribution of this sense, as opposed to the neutral sense, might well provide us with finer geographic detail about the disappearance of the neuter. Moreover, it would be useful to know whether the neutral form that cooccurs alongside of di skind is di or der kind.
YIDDISH CASE AND GENDER VARIATION
199
The unedited records of the L C A A J do not as yet provide conclusive evidence on the dative of kind. If in the Northeast dem should turn out to be usual in dative contexts, with instances of the true feminine der rare, Herzog's account will be confirmed in all its detail and will explain the predominance of ¿em-forms in the accusative, i.e. dem is original, while
200
MEYER WOLF
the nominative der is a later development. Again, the predominance of di in the East-Northeast may indicate that the neuter survived longest there, and we are observing essentially the first stages of the process described by Herzog. The area was found to be conservative in matters of case as well (cf. Maps 3:11, 3:18).
YIDDISH CASE AND GENDER VARIATION
201
4 . 3 2 . 3 . flejs 'meat'. If the maps of kind suggest that in a transition area, the gender of a noun often varies when it is inflected for case, our next maps suggest that variation may also be a function of construction. In Map 4:52 flejs occurs in a definite accusative context {brotn d. . . flejs 'roast the meat') and exhibits essentially a Pattern I distribution. In Map 4:53, a Pattern II distribution emerges when flejs occurs with an attributive adjective in a nominative context (gekoxt. . . flejs 'cooked meat'). There is then, a considerable area where flejs appears to be a deviant neuter—where dos flejs and gekoxte flejs cooccur. We do not know whether this deviation is characteristic of all or even many neuter mass nouns in this area; if it is, we are faced with a structural change of great interest. Otherwise our maps represent the process of change only in a single item, as it passes from the neuter into the mass gender. 4 . 3 2 . 4 . ferd 'horse'. Map 4 : 5 4 suggests that the Northeastern usage with respect to anaphoric pronouns diffuses somewhat more rapidly than Northeastern genders. As we pointed out in 2.13, in the South, anaphoric pronouns depend on the gender of antecedent nouns rather than on the sex of the noun's referent; thus der-nouns take er 'he', day-nouns take es 'it'. In the Northeast, however, the anaphoric pronoun is dependent on the sex of its referent. The border area phenomenon dos ferd 'the horse' with anaphoric pronoun er seems to indicate that the Southern usage has been displaced by the Northeastern before the shift neut > masc affected ferd. The Southeastern instances of der ferd with anaphoric es very likely reflect fluctuation, both in gender and in pronoun usage. The two instances of es to the North, however, are extremely puzzling: they reflect a phenomenon that has so far been observed only sporadically in LCAAJ materials. Further editing of the materials may provide enough information to determine whether Northeastern nonexpletive es represents a survival of the neuter.
4 . 3 3 . SUMMARY
We have summarized the fate of the neuter on Map 4:55. Five nouns are represented by separate isoglosses, each of which deviates somewhat from the shaded area which represents the transition zone for all the other historical neuters that we have considered. It is evident that as one moves northward, the number of neuters that retain their historical gender decreases until the neuter as a category disappears and, finally, even vestiges of the category itself disappear.
202
MEYER WOLF
203
YIDDISH CASE AND GENDER VARIATION
Map4 : 54 ferd 'horse1 And Its Anaphoric Pronoun
i noun gender not elicited - pron. not elicited + both not elicited
BLACK SEA
5. ORIGINS OF THE NORTHEASTERN GENDER SYSTEM 5 . 1 . SUMMARY OF FINDINGS
The twofold aim of the present study has been to explore the facts of geographic variation in Yiddish case and gender and to make explicit the
204
MEYER WOLF
problems that inhere in an attempt to fashion a general description of that variation. Against the background of case and gender usage in StY and NEY, we provided a schematic picture, based on a small amount of data, of the partial loss of the dative/accusative distinction over most of the Southwest
YIDDISH CASE AND GENDER VARIATION
205
and Northeast. Subsequently, a comparison of pronominal case mergers with case mergers in the nominal modifiers revealed that the distinction was better preserved in the East-Northeast than in other Northeastern subregions, while in the North it has been lost completely. Only in the Southeast has the distinction been consistently preserved (cf. Map 3:19). We delimited the North, the East-Northeast, the Southeast, Northeast Poland and the West-Southwest as areas of gender regionalization (Map 4:51). The vagueness of our regional nomenclature, however, prevents the full identification of these gender areas with the areas of case merger even when they bear the same names. It is apparent that disturbances in both the historical gender system and in the historical case system occur in the same general regions. Paramount among our findings is the fact that what we call the morphological variants of EY case and gender have very specific distributions and that these distributions are associated with a small number of geographic subregions of the EY language territory. 5 . 2 . SOURCES OF THE N E W GENDER SYSTEM
Having acquired a measure of geographic knowledge about case and gender in Yiddish, we now wish to examine the manner in which that knowledge contributes to the solution of the central problem of Yiddish diachronic morphology: how the case and gender morphology of presentday NEY emerged from that of an earlier stage of Yiddish. For NEY gender this question has traditionally been formulated in terms of two more particular ones: (a) Why was the neuter gender lost? (b) What is the source of the nonneuter gender deviations? We shall deal first with an hypothesis that invokes the influence of coterritorial languages in answer to these questions.56 We shall then depart from our more strictly geographic approach applied earlier and attempt to add a historical explanatory dimension to our study by reformulating the questions in light of our discoveries and offering several hypotheses of our own. 5 . 2 1 . THE COTERRITORIAL INFLUENCE HYPOTHESIS 5 . 2 1 . 1 . THE HYPOTHESIS. Attempts to account for the radical transformation of the historical EY case and gender system have generally been 56
Various other hypotheses are ignored here. One (Roback 1945) has attributed differences in the gender system to regional differences in character among East European Jews. Another (Shapiro 1939) has attributed the Northeastern two-gender system to an unknown two-gender substratum language purported to have been spoken by oriental Jews who preceded the Yiddish-speaking Ashkenazic Jews in the Slavic lands.
206
MEYER WOLF
hampered by the manner in which the problem was formulated. The questions, as posed above, suggest an impoverishment of the NEY gender system; not only has a gender category been lost, but even among the categories preserved, the historical classification of nouns has been disturbed. Mark (1944) offered a single answer to both questions: gender deviation in NEY is due to the influence o f the languages with which Yiddish has been in contact. More precisely, the loss of the neuter gender was due to the contact of NEY with coterritorial Lithuanian and Belorussian. Also implicated in the loss, though Mark considered them of lesser importance, were the influence of Hebrew-Aramaic (which also lacks a neuter gender) and a putative tendency toward gender simplification in Yiddish. In general, Mark held that the gender of nouns in NEY fluctuated greatly; nonetheless, he claimed that many nouns tended to stabilize and to show the same masculine or feminine gender as their translation equivalents in Lithuanian or Belorussian. Thus, kop 'head' is said to have shifted from masculine to feminine under the influence of both these languages, where the words for 'head' are feminine (Lith. galvá, Belor. halavd). On the face of it, the hypothesis seems to be a reasonable one. The influence of Slavic on Yiddish phonology, morphology and syntax has been demonstrated again and again (Landau 1927-29; Joffe 1927-28; Shklyar 1933; Jakobson 1953; U. Weinreich 1958a): contact has often left its mark in widespread calquing—in the case of verbs, for instance, German-origin morphemes serving as adverbial complements have been functionally adapted to match the Slavic usage of similar elements (cf. Wexler 1964). By way of contrast, the influence of Lithuanian, mostly in the form of lexical loans (U. Weinreich 1964a: 350), has been of relatively minor significance.
5 . 2 1 . 2 . PROBLEMS WITH THE HYPOTHESIS
However reasonable the coterritorial influence hypothesis may appear on the surface, it does not stand up to closer scrutiny. Its formulation depends on the homonymy of the terms used to describe the gender systems of two languages which, though historically related, have diverged far enough inflectionally so that their nominal systems are not synchronically comparable. The gender of nouns in StY is ultimately defined in terms of the three inflectional possibilities of agreeing modifiers, corresponding to the three-category case system of StY. The gender of nouns in the coterritorial Baltic and Slavic languages, on the other hand, is defined in terms of the inflectional possibilities of both the nouns themselves and the agreeing modifiers—with six to eight possibilities, depending on the number of case
YIDDISH CASE AND GENDER VARIATION
207
distinctions in the particular language. A gender, we are saying, can be considered neither a formal nor a substantive language universal. In principle —even though the principle is not always adhered to—a gender is a conventional marker that indicates the inflectional class of a noun (or of the elements which the noun governs) in a particular language. If both Yiddish and the Baltic-Slavic languages have such markers and the markers go back to a common historical source, this by itself does not warrant their synchronic cross-linguistic identification. We do not deny the possibility of a cross-linguistic identification of corresponding gender classes in Yiddish and the coterritorial languages. Rather, we suggest that no such relationship has yet been demonstrated and that until it is, great caution must be exercised in associating the Yiddish feminine with the Slavic or Baltic feminine for linguistic purposes. 5 . 2 1 . 3 . VALIDATION OF THE HYPOTHESIS. Any serious demonstration of the hypothesis that the gender classes of Baltic and Slavic exercised some influence on NEY gender reclassification would have to include structural, statistical and geographical evidence of the sort outlined below (we use Belorussian as a categorical representative of the coterritorial languages. Our source is Kolas 1953).
Structural Evidence The classes termed feminine and masculine in both Belorussian and Yiddish have at least one characteristic in common—most nouns denoting females belong to the feminine gender in both languages. Similarly, most nouns denoting males belong to the masculine. Do the gender classes share any other synchronic properties? Posing this question suggests an analogical formula, whose implications ought to be explored. Using only the male and female characteristic, suppose we wish to suggest that Yiddish left 'spoon' shifted from the masculine to the feminine (or stabilized its gender as feminine) under the influence of Belorussian lyzka 'spoon'. The analogical formula would be: zancyna (BF): lyzka (BF):: froj (YF) : left (X) where (BF) means Belorussian feminine, (YF) Yiddish feminine, zancyna and froj 'woman'. We must keep in mind, however, that such an analogical formula also admits of proportions such as isci (Belor. Irregular Verb): gejn (Yid. Irregular Verb) etc. Unless it can be shown that there are restrictions which forbid the latter but permit the former, the analogical formula is no more than a record of a Yiddish gender shift, not a description of the factor which conditioned its result.
208
MEYER WOLF
Statistical Evidence The following is a schematic sketch of one kind of statistical investigation that could be sharpened and applied. (1) Compile a list of, say more than 300 simplex Yiddish nouns and determine their historical genders as accurately as possible. Prepare a corresponding list of Belorussian translation equivalents of the Yiddish nouns. (2) On the basis of the proportions of each gender to the total number of nouns, calculate the number of Yid.-Belor. pairs whose gender could be expected to correspond by chance alone. (3) Determine the NEY gender of the nouns on the Yiddish list and count the number of NEY-Belor. pairs whose genders agree. (4) Compare the figures from the calculations of (2) and (3). If they are substantially the same, this would point to a shift in NEY which was random with respect to Belorussian, since the shift destroyed as many pairings as it created. On the other hand, if the comparison should reveal that the number of corresponding pairs increased and was greater than chance, it would be reasonable to assume the operation of Belorussian influence. Should the comparison yield some other combination of results, it would either indicate lack of influence or it would evade linguistic interpretation. We have neither built in the safeguards to guarantee the results of the proposed statistical investigation, nor have we determined the test of significance to be used in the comparison suggested in (4). Our aim has been to outline the kind of statistical evidence that would provide reasonable support for the coterritorial influence hypothesis. If the results of such a test were positive, they would still only serve to suggest that Belorussian had influenced NEY in some way. To show that the influence operated in the particular way suggested by the example kop 'head' (masc > fem because halava is feminine in Belorussian), it would be necessary to show (1) that in a large number of cases, a significant majority changed in conformity with the Belorussian model and (2) that the minority which did not could be accounted for in some other way. Furthermore, it would be necessary to show that, among the Yiddish nouns that already agreed with Belorussian, there were few instances of gender change, and that those that did occur were explainable on other grounds. Geographic Evidence Even if structural and statistical evidence were not forthcoming, the
209
YIDDISH CASE AND GENDER VARIATION
reasonableness of the coterritorial influence hypothesis might find support in geographical evidence. On Maps 4:27 (rod), 4:28 (kisri), 4:31 (bret), 4:35 (bojx) and 4:37 (nus), we saw that there was a clear difference in gender between the North and the rest of the Northeast. To demonstrate coterritorial gender influence one would have to show that the gender of these five nouns in the North differs from their gender in the rest of the Northeast in just the way that the corresponding nouns in the Baltic languages differ in gender from the nouns in Belorussian. For instance, if with respect to Map 4:35 (bojx 'belly'), we found that in Belorussian 'belly' is masculine, but in Lithuanian it is feminine, this would count as an instance in favor of cross-linguistic gender influence.57 Again, even if supporting instances were found, the failure of correspondence in other cases would remain to be accounted for. If the reasons for failure were not made explicit, the proposition that the coterritorial gender influenced the Yiddish gender would amount to a misleading restatement of the agreement of some Yiddish innovations with the coterritorial gender—not an explanation of it.
5 . 2 2 . THE STRUCTURAL REORGANIZATION HYPOTHESIS
As we have suggested, the historical gender problem in Yiddish had always been cast in such general terms that only general processes such as coterritorial influence could be invoked to account for it. U. Weinreich (1961) showed, however, that the NEY system, far from having been impoverished, was rather reorganized and augmented. And, as we have pointed out above, one of the products of this reorganization—the intermediate subgender of the feminine with its postprepositional, masculine67 As an indication of the support that may be expected from such geographical evidence, we have tabulated below (Fig. 5:1) the comparative genders of the five nouns mentioned in the text:
Fig. 5:1 Coterritorial Gender Correspondences Belorussian
Yiddish Northeast
'wheel' 'pillow' 'board' 'nut' 'belly'
kóla padüSka dóSka aréxi iivot br'üxa (vulgar)
N F F M M N
Lithuanian North
F F F M
rod kiSn bret nus
M M M F
M
bojx
F
M F F M M M
ràtas pagalvé lentà riéStas skilvis pilvas (vulgar)
210
MEYER WOLF
like article contraction—led investigators to exaggerate the extent of gender fluctuation in the Northeast. Perhaps the subregional paradigmatic differences in NEY, unknown or overlooked by earlier investigators, contributed to the impression of widespread fluctuation as well. By taking into consideration the facts uncovered by our investigation and by adopting the view that the NEY gender system underwent structural reorganization, we can now reformulate the central problem of Yiddish diachronic morphology, which seeks to explain how the case and gender morphology of an earlier stage of Yiddish developed into that of presentday NEY, as three specific questions: (1) What was the fate of historical neuters in the NEY gender system? (2) What were the sources of the morphology of the mass and intermediate subgenders ? (3) Why did certain feminine nouns shift to the masculine and certain masculines to the feminine ? We shall propose answers to these questions below. Before going on, however, we wish to clarify the aims of our hypothesis by remarking on its nature. It would be misleading to claim that it explains the N E Y phenomena in the sense of offering causes. The causes of language change are the same as the causes of cultural change in general—innovation and subsequent diffusion of innovation. One of the goals of a dialectological description is to reconstruct the history of the diffusion of an innovation from its current distribution and from historical information of other sorts (Moulton 1961; U. Weinreich 1964a; Herzog 1965a:235ff.). Since we cannot adequately identify the hypothetical areas in our account with real areas, we fall short of this aim. However, this goal is pursued by the dialectologist only in his role as culture historian; in his role as linguist, he pursues a second goal—-he attempts to reconstruct the series of innovations which led from an older known or reconstructed stage of a language to the system as we observe it today (Herzog 1965a: 161 ff.; U. Weinreich 1958b). It is toward this goal, with respect to NEY case and gender, that our hypothesis is aimed: for if anything emerges from the careful comparison of the historical Yiddish system with that of NEY, it is that the latter is the result of a complex series of innovations. 5 . 2 2 . 1 . THE NEUTER-FEMININE SYNCRETISM. W e p o s i t , a s t h e initial
innovation on which subsequent ones were built, the partial falling together, in a subarea of NEY, of the historical neuter gender with the feminine. That is, the structural similarity between the feminine adjectival inflections and those of the neuter (nom = acc in both the feminine and neuter para-
YIDDISH CASE AND GENDER VARIATION
211
digms; nom = acc adjectival inflection is -e for both genders) led to the replacement of the neuter article dos [dos ~ das ~ s] by the feminine article di [di ~ de]. 58 Thus: Early Yiddish *dos alte bet > NEY (subarea) di alte bet 'the old bed'. As we observed in the case of flejs (4.32.3), gender shifts do not necessarily proceed uniformly throughout the whole paradigm. Consequently, we may assume that even after dos (alte bet) had been replaced by nom/acc di, fluctuation in the dative (dem altn bet vs. der alter bet) continued for some time (cf. the case of kind 4.32.2, where however, the recorded variation is in the accusative). 59 At this point there were two kinds of feminines: historically feminine zax 'thing' functioned with a true feminine paradigm; formerly neuter bet fluctuated in the dative between mit der {alter) bet and mit dem (altn) bet. In allegro speech, mitn bet fluctuated with mit der bet in contrast to mit der zax which had no competing reduced form. 60 It is in this fluctuation that the reorganization of the NEY system may well have begun. 5 . 2 2 . 2 . THE MASS INFLECTION. T h e
key to the origin of the
mass
gender may also lie in an analogical development1. In Yiddish, as in other Germanic languages, there is a partial formal resemblance between mass nouns and plural nouns: the indefinite article has a zero allomorph in construction with both mass and plural nouns, but with count nouns it is always a(n). Thus, epl 'apples' and puter 'butter', but an epl 'an apple'; also, frise epl 'fresh apples', frise puter 'fresh butter', but a friser epl 'a fresh apple'; mit (frise) epl 'with (fresh) apples', mit (frise) puter 'with (fresh) butter', but mit a (frisn) epl 'with a (fresh) apple'. 61 This resemblance may have led to the following analogical reformation 58 Recall the development of di skind (4.32.2). Perhaps metanalysis played a mediating role in the replacement of dos by di. 59 The fate of the strong/weak neuter distinction and readjustment of the diminutive gender pattern from automatic neuter to the gender of the underlying noun are not involved in the process we are outlining here; consequently we ignore them. 60 In its rare occurrences cur{—cu der) 'to the' appears to be the only postprepositional contraction of the feminine article in EY. However, it has not been recorded in the Northeast. 61 a puter may be possible in the sense of a kind of butter. In that case it would be the singular of puter(s) or putern 'kinds of butter'. The motivation for assuming a zero allomorph for the indefinite article with mass and plural nouns lies in the parallelism of patterning with the negative indefinite article:
a bux 'a book' bixer 'books' puter 'butter'
kejn bux 'no book' kejn bixer 'no books' kejn puter 'no butter'
212
MEYER WOLF
of the semantically definable class of mass nouns into a new gender class : mit epl : mit di epl : : mit puter mit frise epl mit frise puter
: X62
We have assumed here that the syncretism of the mass and plural forms began in oblique contexts and spread from there to the nominative. We find support for this assumption in the fact that mass nouns do not enter into the plural agreement rule operating between subject nominals and verbs (cf. 2 . 2 2 ) . Had syncretism begun in the nominative, we would have expected the plural agreement rule to be extended to embrace mass nouns. 5 . 2 2 . 3 . POSTPREPOSITIONAL CONTRACTION. The survival of postprepositional contractions of the article with feminine concrete inanimates challenges us to explain their failure to survive with feminine mass nouns; 63 where it may be possible to detect the survival of the article contraction in a semantically specialized function. The historical gender of ajzn 'iron' appears to be neuter, therefore we would expect it to appear as di ajzn in NEY. Indeed it does, but only in its mass sense. In the count sense 'definite amount of/kind of' we find der ajzn, apparently a back formation from mitn ajzn, which is ambiguous in StY, because of its definite article, as to whether 'iron' in the mass or 'iron' in a limited amount is intended. If this is so, the occasional mass to count noun conversion by article alternation di/der noted by Mark (1951) may be more general than it seems, for even in the case where a single form di zalc appears for both mass and count senses, di may in the first case represent di of the mass subgender, while in the second case it may represent di of the intermediate. In any case, the semantic specialization of the postprepositional contracted article would account for its absence from the mass paradigm.
If indeed the factors just described led to the gradual elimination of the 62
We chose epl for its phonemically identical singular and plural forms in order to emphasize that the semantic parallelism which motivated the formal change was between the mass and the plural; the constituent morphemes of the plural were irrelevant. If this were not so, it is perfectly possible that from flejs 'meat' we might have gotten a form "flejsn on the model of lis, 'table', tisn 'tables'. The form *flejSn would not differ in meaning from the former flejs; its semantic content, however, would have been distributed over two morphemes: *flejS 'meatness' and *-n 'massness'. Since this did not happen, we have chosen epl rather than tiin in the interests of an uncluttered example. 63 Neuter animates are so rare that their loss of intermediate inflection can be accounted for in terms of the analogical shift of each individual word (e.g. vajb 'wife') as the true-feminine class became a semantic class. The semantic class of kind 'child'— a personal noun, unmarked for sex—presents difficulties whose clarification may shed light on its survival as a sort of neuter in the Northeast where, for example, vajb is universally feminine.
YIDDISH CASE AND GENDER VARIATION
213
dative fluctuation mitn X~ mit der X, in the paradigms of abstract and mass nouns, then such fluctuation came to be characteristic of a class of nouns which can be semantically characterized, namely the concrete inanimates. Since there was no way to distinguish these intermediates from the true feminine concrete nouns that take di in the nominative, we may assume the gradual absorption—by semantic analogy—of true feminine concrete inanimates into the intermediate subgender. In the case of bord 'beard' (U. Weinreich 1961), this process has apparently not yet begun; koldre 'blanket' (Map 2:7, 2.21.4) may be at an intermediate stage. 5.22.4. FEM > MASC AND MASC > FEM SHIFTS. These gender shifts may have resulted from the spread of the system through dialect contact. As suggested in Herzog (1965a: 114; cf. also 4.13 above), the gender of an intermediate noun, in a dialect lacking the intermediate subgender, may be taken either as feminine, on the basis of its nominative and other full paradigmatic forms, or as masculine, on the basis of its postprepositional contracted article. Thus, for the historically feminine noun brik 'bridge' in the Northeast (cf. Map 4:9), we assume the series of changes represented in Fig. 5:2: (1) true feminine>intermediate in area A; (2) in the spread of the intermediate from area A to area B, brik is regionally reinterpreted as masculine (i.e. mitn brik 'with the bridge' is taken to indicate der brik 'the bridge'); (3) the masculine form of area B is then diffused back into area A and throughout the Northeast, submerging all nonmasculines and yielding the distribution of Map 4:9. Figure 5:2 Stratigraphic diagram of posited gender shifts of brik 'bridge' from true feminine to masculine in NEY Masculine
(3) (2)
Intermediate
Masculine
True feminine
(1)
Intermediate
True feminine True feminine
A
B
The same process may also account for Maps 4:1; (grenec 'border') and 4:11 (cung 'tongue'). A similar series can be constructed for the shifts shown on Maps 4:17 (haldz 'neck, throat') and 4:18 {left 'spoon'). In these shifts of masculine to feminine, the postprepositional contracted
214
MEYER WOLF
masculine article was taken as a mark of the intermediate subgender (i.e. mitn left 'with the spoon' is taken to indicate di left). If it is indeed the case that Fig. 5:2 reconstructs for us the emergence and diffusion of gender forms through the Northeast, then maps such as 4:35 (bojx 'belly') and 4:36 (teler 'plate') may well capture for us an intermediate stage, i.e. they may depict the point—different in each case—at which a historical masculine has been reinterpreted and regionalized as a feminine in different subareas of the Northeast. If this is so, a composite of the two maps may well suggest the source from which the intermediate gender first spread. In Map 5:1 we have superimposed the two maps in question (4:35, 4:36); we have also included the Northeast Poland masc < intermediate area of Map 4:27 {rod 'wheel'), since as was indicated in 4.21.1, the intermediate has been diffusing through this area for some time. In the inter-
YIDDISH CASE AND GENDER VARIATION
215
stices of the three distributions, we observe a small area around Hrodna to the south and around Vicebsk to the north. Of course, the distributions we have reproduced on Map 5:1 outline the source area distortedly, if at all. Nonetheless, there is some reason to believe that the Hrodna area was the cradle of NEY (Herzog 1965a: 236). Further historical investigation may shed some light on the problem. We have proposed this series of regional contacts despite our lack of an adequate historical dialectological study of the Northeast on which to base it. It is hoped that such a study will offer independent historical support to the assumption that the center of innovation shifted from an area A to a second area B as our account requires, and will associate the labels with actual subregions. 5 . 2 2 . 5 . CONCLUSION. We consider our speculative attempt to account for regional gender variation in Yiddish an appropriate conclusion to a factual study of the variation itself. The comparison of the coterritorial influence hypothesis with our set of structural hypotheses is an appropriate conclusion in another sense as well. Underlying the former hypothesis is the assumption that the study of diachronic morphology involves only the historical identification of morphological details—the genders of nouns and the causes of gender change. Implicit in our approach to diachronic morphology, however, has been the leading principle articulated by Diver (1958:54), that
an insistence on viewing the history of morphological details in the context of other forms of the language and of the opposed concepts they represent should lead to a clearer view of the details and to a broader understanding of the manner in which languages change.
ON ACCENTUAL VARIANTS IN SLAVIC C O M P O N E N T OF
THE
YIDDISH
EUGENE GREEN 1. INTRODUCTION
In his seminal paper on Yiddish stress and word structure, U. Weinreich notes the desirability of a study of "the origins of the Yiddish stress patterns and of the integration of elements of heterogeneous origin in a unified stress system" (1954:16). As a step to achieving these purposes, the survey and analysis presented in this paper center on stress variants of tri- and polysyllabic words in the Slavic component of Yiddish. The principal question that the study poses—how to account for these stress variants— issues from the fact that they appear in Yiddish dialects with the stress on different syllables: 'floor', for example, is rendered in some areas as padloge, in others as podlege; 'horseshoe' appears alternately as pidkove and as podkeve. Instances abound. The synchronic stress system of Yiddish does not account for these differences, for, according to Weinreich, all the dialects are alike in possessing a stress whose place is, in principle, unpredictable; it can fall on almost any syllable of the word. 1 The perspective thus needs to be historical and dialectal; it asks: (1) What features of accentual systems in the older German and Hebrew-Aramaic components as well as in the Slavic languages themselves are significant to the development of these stress variants? (2) How did these features enter into the process of integrating these stress variants in the dialects of Yiddish ? Earlier studies have proposed explanations based on the influence of 1
See U. Weinreich (1954) for an analysis of phonemic stress in Yiddish. Further instances of words exemplifying unpredictable stress include roiinkes 'raisins', meSpoxe 'family', and telegram 'telegram'. Accentual variations in the Yiddish component of nonSlavic origin are rare: e.g. inevejnik and inivenik 'inside' in the German component; koj(Ji)anim and kejanim 'members of the priestly class' in the Hebrew-Aramaic component. Only tri- and polysyllabic words are considered in this paper, because the variation in stress is either on the initial or penultimate syllables. Yiddish has hardly any Slavicisms with final stress; ¿epuxa 'nonsense' is one, but it sounds very "foreign" in Yiddish. Bisyllabic words may have either syllable stressed as in gezunt 'health' and take 'indeed', but distinctions of initial or penultimate syllables are misleading in them. 216
ACCENTUAL VARIANTS IN YIDDISH
217
German or Slavic stress patterns, but with only partial success. Leibel (1965:71), for example, ascribes the accenting of initial syllables in Polish loanwords in Yiddish primarily to the German pattern of root-initial stress (see § 2.1). In his view, the initial stress on xoljeve 'boot leg', mednice '(copper) washing pail', and pdstronik 'string', clearly different from the penultimate stress in their Polish counterparts, is due to the influence of words in the German component of Yiddish like vinter 'winter' and naket 'naked'. This explanation has three shortcomings: (1) it is limited to Central Yiddish (CY) in Poland and ignores the dialects to the east; (2) it fails to account even in CY for words like pocontek 'first sale' and cikave 'curious' that retain the penultimate stress of the Polish model and could not be influenced, therefore, by the German pattern of stress; and (3) it overlooks the fact that the great majority of words in the German component are bisyllabic, so that one cannot fairly say whether their stress is initial or penultimate. Jakobson (1953:212) also offers the German root-initial stress as a model for Slavicisms, but even more interestingly points to Czech and Old Polish rules of initial syllable stress as most influential for the stressing of tri- and polysyllabic words in the Slavic component of Yiddish. This rule, moreover, is held to be effective even after changes in the Polish system of stress itself and after the contact of Yiddish with Belorussian and Ukrainian and their systems of free, unpredictable stress (see § 2.2). In support of Jakobson's view, one could cite again the loanwords from Polish listed above and add to them other loanwords from Belorussian and Ukrainian like recene 'buckwheat gruel', Br. hracany; suneces 'wild strawberries', Br. sunicja; kropeve 'nettle', Ukr. kropyva; kromisles 'yoke', Ukr. koromyslo; krenece 'well, spring', Ukr. krynycja. It would seem then that the Old Polish and Czech rules have long applied to borrowings from all the Slavic languages in contact with Yiddish. The weakness in Jakobson's account is that there are too many exceptions. It does not explain why the Old Polish and Czech rules had no effect upon the penultimate stress of such Slavicisms as cerate 'oil cloth', cerulnik 'barber', and katoves 'jest'; nor does it account for variations in stress in forms like kaluze\kaleze 'puddle' and zavjese||zavise 'hinge'. In both Leibel and Jakobson's studies the principal difficulty is that too much weight is put upon the power of a model, either German root-initial stress or Old Polish and Czech initial stress, to account for accent in the Slavic component of Yiddish. The result is an oversimplification that ignores significant details of stress variants and dialect differences. If German or Slavic models alone will not do as sources of explanation, it is appropriate to look closer to home, to Yiddish itself, to see how the 8
218
EUGENE GREEN
language in its development fused its components together and thus put its own creative stamp on tri- and polysyllabic Slavicisms. This fusion results from the interaction of several elements: stress patterns in the older components of Yiddish as well as the patterns of the Slavic languages in contact. Thus, we need to begin with a description of these patterns and the ways they interacted with one another. 2 . RELEVANT FEATURES OF THE ACCENTUAL SYSTEMS
From the very first, Yiddish was an innovating language, developing patterns of stress that differed from those in the stock languages. On German language territory, the chief accentual innovation is related to words from the pre-Ashkenazic reading tradition, words in the Hebrew-Aramaic component. And to the east, as early as in its encounter with Czech and Old Polish, Yiddish impressed its own character on the first borrowings. 2 . 1 . RULES OF ACCENT IN THE OLDER COMPONENTS
A survey of the vocabulary in the German and Hebrew-Aramaic components indicates clearly enough that the rule of root-initial stress in the German stock language did not hold full sway in Yiddish, even in the early centuries. To be sure, words in the German component show only slight (cf. fn. 1), if any, variation. As in German, the root is stressed in Yiddish bruder 'brother', fingerl 'ring', and kumen 'to come'; likewise Yiddish adheres to the grammatical rule that places the stress on a separable prefix preceding a root as in aropfaln 'to fall down' (fain 'to fall') and ibergejn (gejn 'to go'). Moreover, it might appear at first glance that the rule of root-initial stress extended to words in the Hebrew-Aramaic component. Leibel (1965:69) suggests, for example, that the accented initial syllable in many words acts as if it were a root: xaneke 'Hanukkah' bog(e)res 'adolescent girl', and xasene 'wedding'. The analogy between the place of stress in these words and in the German component seems evident. But as Leibel's analysis also implies, it is not a necessary analogy. Instead, one can look to the order of vowels that the antecedents of "Merged Hebrew" words had in the pre-Ashkenazic reading tradition for a more plausible account of accent in the Hebrew-Aramaic component.2 If the order of vowels in these words comprises sequences of first syllable long, second syllable short, third syllable long or short [ — ] or vowels of equal duration in the first and second syllables (both short ] or both 2
The pre-Ashkenazic reading tradition is that which had been "current in Italy, France, and S p a i n . . . " (Leibel: loc. cit.). "Merged Hebrew" consists of the HebrewAramaic elements that have entered Yiddish.
ACCENTUAL VARIANTS IN YIDDISH
219
long [ — ]), then the stress falls on the antepenultimate syllable: Leibel's examples include göliyes 'diasporas' ( < galuyot); khasene 'wedding' (