Yiddish–Lithuanian bilingualism: Incomplete acquisition,change through contacts,or both?

Sociolinguistics of Yiddish in Lithuania

The minority-language status of Yiddish in Lithuania dates back several centuries. Starting from the fourteenth century, the first Grand Duchy of Lithuania and then so-called ethnographic Lithuania was home to a substantial Jewish population. It was a place of traditional Jewish learning and later a locus for trends such as Zionism, Bundism (Jewish workers’ political movement), and Yiddishism. To the best of my knowledge, no systematic account exists on proficiency in Lithuanian among Jews prior to Lithuanian independence in 1918. Yiddish was most often the first language, and knowledge of Lithuanian varied depending on, for example, personal needs, occupation, and relations with Lithuanian neighbors. Jews were the largest minority group in Lithuania, constituting seven percent of the population before WWII (120,000 persons). From 1918, a language with a former low standing received the status of the sole state language and, for Jews, proficiency in Standard Lithuanian (not just communicative ability in any variety of Lithuanian) became highly desirable. As described in several studies (Atamukas, 1998, p. 149; Mendelsohn, 1983, pp. 234–235), proficiency in Lithuanian dramatically increased, and there were signs of internalization (adoption of if not outright shift to) of Lithuanian (i.e., creating a Jewish voice in Lithuanian via press, artistic expression, etc; see Verschik, 2010). In the 1920s, very few Jews attended Lithuanian-medium schools but, by the middle of the 1930s, the number of students in Lithuanian-language schools had increased at the expense of Hebrew and Yiddish-medium schools (Atamukas, 1998, p. 149). In 1940, Lithuania was occupied and annexed by the Soviet Union, and in 1941 it was occupied by Nazi Germany. The majority of the Jewish population was killed during the Holocaust. Under the second Soviet occupation (1944–1991), Jewish/Yiddish education was not allowed for the tiny remaining group of Jews. For various reasons that are outside the scope of the current article, Yiddish was usually not transmitted to subsequent generations, and a shift to Lithuanian or Russian occurred during the post-war years.

After the collapse of the Soviet Union, some Jews left Lithuania for Israel or the United States, but a small group expressed no wish to leave. Currently, about 4000 Jews live in Lithuania, most of them in the two biggest urban centers, Vilnius and Kaunas. Unfortunately, official statistics do not provide figures for proficiency in Yiddish; however, a handful of Yiddish-speakers—mostly elderly—are still around. I have been familiar with some of them for more than a decade. Their varieties of Lithuanian Yiddish, although not considered here, are relevant as a point of reference. Yiddish courses provided by Vilnius University and the Summer Program in the Yiddish Language, Literature and Culture (where I have been teaching since its establishment in 1998) have attracted both Jewish and non-Jewish students not only from Lithuania but from around the world. Some have been extremely successful and acquired impressive skills in Yiddish. However, such learners and users of Yiddish are not the subject of the present study.

The informants

As is clear from the previous section, there are very few remaining speakers of Yiddish in Lithuania today, and acquisition of Yiddish as L1 or, more precisely, as one of the L1s in bilingual L1 acquisition is very infrequent. Unfortunately, nothing has been written about Lithuanian–Yiddish bilingualism and, thus, any data that can be obtained are valuable.

My informants in this study, both males, were born and raised in Lithuania and currently live in Vilnius. At the time of writing, I have known them for five or six years. The younger informant (I1) was also my informant in an earlier study (Verschik, 2014a). Usually we communicate both in Yiddish and in Lithuanian. With the older informant (I2) I have been using Lithuanian, but once I learned that he has some skills in Yiddish, I decided to record his speech.

The I1 is in his early twenties. His parents were born in Vilnius, but his grandparents come from Plungė (Plungjan in Yiddish), a town in Žemaitija in the northwest of Lithuania. Yiddish is spoken by all members of the family; both Yiddish and Lithuanian are spoken at home. I1 has an older brother, who has his own family and speaks Yiddish to his daughter. For I1, Yiddish constitutes an important component of Jewish identity. Neither he nor his relatives are religious but one may say that the tradition of secular Yiddishism, once prominent in Lithuania, lives on in the family. In addition to Yiddish and Lithuanian, I1 is conversant in Russian (according to him, learned rather from the street than under a formal instruction) and in English. He graduated from the only Jewish school in Lithuania, where Lithuanian is the language of instruction and Hebrew is taught as a subject. I1 is not fluent in Hebrew but is able to read it. He has no literacy in Yiddish. I asked him whether he has proficiency in the Žemaitijan variety of Lithuanian. I1 denied any proficiency but mentioned that his parents are fluent in it and use it when visiting the area. He is able to understand the variety and enjoys to hear it spoken.

The I2 is in his early fifties. He was born and educated in Kaunas and currently lives in Vilnius. His elderly father is a Lithuanian Jew, still living in Kaunas; the mother was born in Belarus and moved to Lithuania after WWII. Both parents used slightly different varieties of northeastern Yiddish. According to I2, they also used Russian with each other and the children. His father is fully fluent in Lithuanian and his mother learned the language. I2 has an older brother, who emigrated from Lithuania to Israel in the 1990s. Both brothers attended a Lithuanian-medium school and speak Lithuanian to each other. I have no information about whether the siblings started using Lithuanian at school age or even earlier. Kaunas is a city with an overwhelming Lithuanian majority. After WWII Lithuanian rather than Russian was the preferred language among the remaining Jews in Kaunas; therefore, I assume the language was certainly not unknown to I2 and his brother before their entry to school. I2 reports some proficiency in English, a language he learned at school and later at medical university. As we shall see, his English, albeit not very fluent and not exactly idiomatic, has had some impact on his Yiddish speech. Currently, I2 speaks Yiddish to his father and Lithuanian with his family. One of his children, a young adult, has studied Yiddish at the Summer Program, but she does not speak the language at home.

For the current study, I recorded I1 for half an hour in summer 2011, and I am also using my earlier recording of I1, made in summer 2010 (one hour). I2 was recorded for approximately one hour in 2011. As I am familiar with the informants and interact with them more or less regularly during the Summer Program, I have also had a chance to observe their interaction in daily life.

It has to be stressed that, whatever their level of proficiency in Yiddish may be, my informants identify themselves as speakers of Yiddish and as multilinguals. As I1 put it in a rather straightforward manner, in a idiše mišpoxe me darf redn af idiš “in a Jewish family one has to speak Yiddish.” This circumstance is important because it makes the informants utterly different from those in Levine’s (2000) study of mostly second-generation speakers of Yiddish residents in the United States, where speakers were reluctant to speak Yiddish because of a general stigmatization of immigrant languages and internalized negative attitudes towards Yiddish in particular.

The speech of the informants: overall characteristics

Before turning to the case studies, I will provide a brief description of the informants’ speech. The focus will be on the following aspects: (a) regional traits (in case of I1, Zameter Yiddish, a variety once spoken in Žemaitija; in case of I2, general northeastern Yiddish (NEY) features (for an overview of Yiddish varieties in Lithuania see Mark, 1951; Jacobs, 2005, p. 60); (b) “non-target” features in Yiddish (some of them with a parallel in Lithuanian); and (c) comparison with the features reported by Levine (2000) in incomplete acquisition of Yiddish in the United States.

The I1 speaks Yiddish fluently: his utterances are well-formed, and there are few hesitations. His speech exhibits Zameter Yiddish features in phonology (some diphthongs, distribution of s and š) and general NEY features in morphology, such as the lack of the neuter gender (Jacobs, 1990) and case syncretism (the merger of dative and accusative, Wolf, 1969). I will return to case and gender in discussion of Levine’s (2000) list of non-target features. Occasional switches to Lithuanian cannot, in my view, be considered as evidence of incomplete acquisition: first, code-switching is a hallmark of Jewish speech in Lithuania (elderly speakers switch as well); and second, code-switching as such may be a norm in a multilingual community.

On the other hand, I1 has Lithuanian-like intonation and Lithuanian-like progressive palatalization in consonant groups before front vowels: š’v’es’t’er “sister” instead of švester, geb’l’ibn “remained” instead of geblibn. Thus, the features in phonology and prosody have a clear Lithuanian prototype.

In morphosyntax, I1 often departs from the normal Yiddish word order, especially in negative sentences: ix nit vejs zejer “I don’t know for sure,” instead of the usual Yiddish word order (in all regional varieties) ix vejs nit zejer. Verb-second (V2) position in the main clause is not always consistent: dortn mir redn af litviš “we speak Lithuanian there” instead of dortn redn mir af litviš (in all regional varieties). On the one hand, it is possible to consider the features in question not as a result of transfer (or, in a different terminology, copying, see Johanson, 2002a) from the dominant language but as a development toward less marked word order (free word order being less marked than V2). As Thomason (2001, p. 279) suggests, universally unmarked features are easier to acquire and are predicted to occur more commonly. On the other hand, however, changes in word-order and non-core morphosyntax, together with an impact in phonology, are characteristic of one of the intermediate stages of contact-induced change in language maintenance (Thomason & Kaufman, 1991, pp. 74–75). Therefore, the question remains open whether the changes in word order are a result of Lithuanian impact because, first, in Lithuanian negative sentences the negative particle precedes the negated component: aš nežinau (literally, I not know) “I don’t know” and, second, the word order is relatively free and Lithuanian does not have a V2 rule.

The most striking feature in I1’s speech is the lack of past tense auxiliary: er gelernt “he learned,” cf. er hot gelernt; zej aroušgekumen “they came out,” cf. zej zajnen arojsgekumen. This will be dealt with in the “Past tense auxiliary deletion” subsection.

The I2’s speech is less fluent: it is marked with hesitations and self-repair. At times, he has Lithuanian intonation as well. Regional NEY features in phonetics are not consistent since standard features occur as well: consider the realization of š in the NEY pronunciation of mentsn “people” (Standard mentšn) but the Standard pronunciation of švester “sister” (instead of expected NEY svester). The realization of diphthongs varies as well: grejser “big” (Standard grojser) but Standard ojsgebojt “built” (NEY ujsgebujt). Gender and case appear to be more complicated issues; clearly, no neuter gender appears, but an overall impression is that gender assignment is less consistent than in the case of I1. With very few exceptions, in general Yiddish nouns lack overt case marking in Standard Yiddish and in regional varieties. Case is marked on definite articles and agreeing adjectives (der tate “the father,” nominative case singular, mit dem tatn “with the father” dative case singular), and for this reason I would for now refrain from postulating the erosion of case systems, despite the fact that it is often discussed in the HL literature (see Montrul, 2010, pp. 6–9 for a summary on HL morphosyntax and syntax).

Exactly as I1, I2 departs from the usual Yiddish word order: ix nit zog “I am not saying,” cf. ix zog nit; zejer grejse gelt darfst “(you) need really big money,” cf. darfst zejer grejse (Standard Yiddish: grojse) gelt.

Also I2’s speech demonstrates a complete lack of past tense auxiliary: ven geven krizis, bank gezogt “when there was the crisis, the bank said,” cf. ven s’iz geven der krizis, hot der bank gezogt; Rusland nit gehat gelt un di transportne kompanje zix farmaxt “Russia lacked the money and the traffic company closed,” cf. hot nit gehat “had not” and hot zix farmaxt “closed.”

Interestingly, there is a whole list of non-target features that are present in I2’s but not in I1’s Yiddish. The following are examples:

Occasional non-target past participles in irregular verbs: gehabn instead of gehat “had,” avekgefort instead of avekgeforn “left.”

Now let us look at the list of features provided by Levine (2000, pp. 38, 39–40, 43, 47, 50–52, 73). The comparison between his and my informants is presented in Table 1.

OPEN IN VIEWER

Table 1.

Informants’ speech in Levine (2000) and the present study.

Levine (2000)	Younger informant	Older informant
Considerable effort to sustain their speech in Yiddish (p. 38)	–	+/–
Lack of subject–verb agreement (p. 38)	–	To some extent, third person singular = third person plural may be considered as such; no instances other than that
Preference of Germanic-origin items over words of Hebrew/Aramaic origin (pp. 39–40)	Cannot be considered a diagnostic feature because of wide regional, stylistic, and idiolectal variation in Yiddish
Non-grammatical plurals (pp. 50–52)	–	–
So-called code-blended forms (English stem and Yiddish morphology) (p. 43)	Not applicable, see explanation regarding code-switching in “The speech of the informants: overall characteristics” subsection
Non-target past participles (p. 43)	–	Very few
Non-target assignment of gender and case (either generalized de or ungrammatical use of articles) (p. 47)	Problematic because of northeastern Yiddish regional developments
In past tense, the auxiliary hobn preferred over zajn; at times zajn is used grammatically (p. 73)	–	–

As is demonstrated in Table 1, only some features appear in the speech of my informants. Lack of subject–verb agreement appears in I2’s speech in third person only where the Lithuanian prototype exists. Notably, Levine (2000) reports a confusion of auxiliaries and the spread of hobn “have” and does not mention the lack of any auxiliary. In sum, there is little similarity between the features.

Past tense auxiliary deletion

Yiddish has two past tenses: past and past perfect. The latter belongs mostly to written registers and is seldom used in everyday speech. The focus here will be on the former. Past tense in Yiddish has the structure “auxiliary + past participle,” which is more or less isomorphic to German Perfekt, for instance, Standard Yiddish zi hot gešpilt mid dem kind “she played with the child.” It covers both the meaning of simple past and present perfect. Like in German, there are two auxiliaries, zajn “to be” and hobn “to have.” Some 20 intransitive verbs require the auxiliary zajn, and all other verbs require hobn: er iz gegangen “he went/has gone,” zi hot farkojft di dire “she (has) sold the flat.” All verbs that form past tense with zajn are irregular, whereas hobn takes both regular and irregular verbs. The rules of the auxiliary choice are very much like in German.

In the NEY area, the generalization of hobn to all contexts has been attested (Mark, 1951, p. 457; see also Jacobs, 2005, p. 70). However, NEY is a group of varieties once spoken over a vast territory (the Baltic countries, Belarus, Northern Ukraine, and some areas in Poland) rather than one variety and, according to Mark (1951) and other scholars, at least three varieties of Yiddish in ethnographic Lithuania existed: Suvalker Yiddish; Stam-Litvish Yiddish (that is, Yiddish of Lithuania proper); and Zameter Yiddish (the variety spoken by I1). Mark did not link the overgeneralization of hobn to any particular location. I am fairly familiar with most Northern varieties of NEY and have never encountered this feature; during my contacts with the elderly speakers of Yiddish in Lithuania (Zameter Yiddish and Stam-Litvish Yiddish) I have not attested this either.

In itself, overgeneralization of hobn is not extraordinary. First, think of languages such as English where only one auxiliary to have has been retained in present perfect. Thus, this could happen in varieties of NEY without any external reason. Second, the literature on the acquisition of German as L1, a language closely related to Yiddish, reports overuse of haben “to have” in young children (Mills, 1986; Wittek & Tomasello, 2002). The reasons are twofold: (a) the complexity of the rule for learners in the first years of language acquisition (e.g., it is necessary to choose the auxiliary and to form the past participle, while keeping in mind that some participles are irregular (Wittek & Tomasello, 2002, p. 568)); and (b) the higher frequency of haben in adult German speech. Opinions vary on when children completely master the complex rule; some scholars believe that it occurs relatively late (for references, see Mills, 1986). If the same patterns are true for the acquisition of Yiddish, then one can assume that the informants in Levine’s (2000) study had stopped actively using (or even hearing) Yiddish before the rule was mastered. Mills (1986) and Wittek and Tomasello (2002) do not mention the auxiliary omission; neither does Levine (2000). Levine (2000, pp. 80–83) extensively quotes the famous study by Leopold ([1939–1949] 1970) on his daughter’s German–English bilingualism because the auxiliary confusion is attested there as well. Significantly, German in the United States at that period of time was stigmatized and, according to Levine, this might be one reason why the child failed to acquire the rule in full.

Some recent studies on bilingual acquisition where one of the languages is German (for instance Salustri, 2002–2003) do report auxiliary omission in German. Salustri found that in monolingual German children the ratio of bare participles ranges between 55 and 99.5 percent; the same tendency is attested also in bilingual children’s German.

For the language learners, extensive use of bare participles may be due to the fact that bare participles are frequent in colloquial speech in formulae such as: erledigt! “(I’ve) done (it)!”, verstanden? “(have you) understood?” Although their use is restricted, they may be frequent in the input. ¹

Together with auxiliary confusion, Levine (2000, p. 43) mentions non-target participles. As summarized in Table 1, I1 produced no non-target participles. Complicated cases where an irregular participle is formed from a prefixed verb were handled perfectly: oušgelernt “learnt by heart,” aroušgegangen “went/gone out,” and so forth. This holds both for the earlier recording I did for my previous study (Verschik, 2014a) and for the current article.

The I2 has occasional non-target participles such as avekgefort “departed, left” instead of avekgeforn, but in another context he produces the target version of the same participle; there is thus variation in the language of the same speaker. On the whole, I2 uses target participles from irregular verbs: gekrogn “got,” gevorn “became,” and farnumen zix “dealt with.” Evidence from I1’s interview implies that the two things, the handling of the auxiliary and the participle formation, are not necessarily connected.

In a one-hour interview with each informant, the past tense auxiliary zajn “to be” appeared only once in I1’s speech and twice in I2’s speech in the phrase s’iz geven “it’s been.” I believe the phrase was reproduced as an unanalyzed unit: the subject and the verb here are contracted (the full form is es iz “it is”) and the collocation is frequent in Yiddish.

Tables 2 and 3 summarize participle use by I1 and I2, respectively. In Table 2, data from both interviews with I1 (in 2010 and 2011) are included (all participles are target-like).

OPEN IN VIEWER

Table 2.

Participles in the older informant’s speech.

Types	Gloss	2010	2011
1	geven “been”	33	27
2	(ouš)gelernt (zix) “studied, learned, acquired”	12	7
3	(arouš)gegangen “gone (out)”	3
4	(arouš)gekumen “came (out)”	3	2
5	gezogt “said”	3
6	gekent “known”	2	4
7	gelebt “lived, resided”	2	4
8	geendikt “finished”	1	3
9	bakent zix “got acquainted”	1	1
10	gearbet ‘worked’	1	1
11	geblibn “stayed”	1	2
12	gešpirt “anticipated, felt”	1
13	gemejnt “meant”	1
14	(iber)gegebn “given (further), transmitted”	1
15	gešemt zix “ashamed”	1
16	gebitn “changed, altered”	1
17	gedarft “had to”	1	1
18	geborn “born”	1	1
19	gefeln “pleased”	1
20	(iber)gemaxt “(re)done”		6
21	gexavert zix “befriended”		1
22	ibergefirt “switched” (in this context)		1
23	geštelt “placed”		2
24	oušgegrobn “dug-out”		1
25	gevorn “become”		1
26	gehat “had”		2
27	gezen “seen”		4
28	geredt “spoken”		4
29	gehert “heard”		3
30	ongehejbn “started”		1
31	geštelt “put, placed”		2
	Tokens total 151	70	81

OPEN IN VIEWER

Table 3.

Participles in the older informant’s speech.

Types	Gloss
1	geven “been”	27
2	gearbet “worked”	9
3	geendikt “finished, graduated”	9
4	gekumen “came”	7
5	gehat “had”	6
6	gelernt “learned”	5
7	gefunen “found”	4
8	gelebt “lived, resided”	4
9	gedarft “had to”	4
10	(ojs)gebojt “built”	4*
11	geblibn “remained”	3
12	avekgeforn “left, departed”	3**
13	gezogt “said”	3
14	farštanen “understood”	3***
15	gekrogn “obtained, got”	3
16	gemaxt “made”	3
17	gevelt “wanted”	3
18	(durx)gegangen “went (through)”	3
19	gezuxt “looked for”	2
20	gebitn “changed, altered”	2
21	gelejent “read”	2
22	farmaxt “closed”	2
23	arendirt “rented”	1
24	gevorn “became”	1
25	farnumen “taken up, occupied”	1
26	gekent “knew; could”	1
27	umgekumen “perished”	1
28	ibergelejent “re-read”	1
29	baštelt “ordered”	1
30	geredt “spoken”	1
31	geštarbt “died”	1****
32	gešinen “seemed”	1
33	getraxt “thought”	1
34	gezen “seen”	1
35	gehert “heard”	1
36	genumen “taken”	1
37	faršlosn “closed”	1
	Tokens	126

Note: *non-target ojsgebojen 1 time; **non-target avekgefort 2 times; ***non-target farštejt 1 time; ****the single occurrence is non-target (cf. target geštorbn).

As shown in the beginning of this subsection, evidence on monolingual and bilingual acquisition of German Perfekt and of incomplete acquisition of German/Yiddish differ: some find confusion of the auxiliaries, as well as ungrammatical participles (Levine, 2000; Wittek & Tomasello, 2002); some report auxiliary deletion in monolingual and bilingual acquisition of German (Salustri, 2002–2003). Thus, there is no uniform pattern, and the question remains open as to why in some cases we encounter confusion of auxiliaries (or overgeneralization of the most frequently used auxiliary) and the deletion of auxiliary in others.

Unfortunately, I have no information about the kinds of varieties the parents of the informants speak and when exactly auxiliary deletion occurred. It is reasonable to suggest that I1’s parents, a post-war generation, had a kind of limited input in the sense of the quality and the quantity (fewer registers, i.e., no literacy, and fewer opportunities to use the language). I2’s parents belong to the older generation that grew up when Yiddish was widely used and normally transmitted to their children. Based on my experience with the speakers from the same area and the same generation, I would assume that I2’s father acquired a full version of Yiddish, so to say. However, we have no knowledge about whether and how his variety of Yiddish changed and exactly what kind of input I2 received. Note that studies on limited input have been conducted mostly in immigrant communities, in which case the point of arrival to the new country can be clearly established. Yiddish in Lithuania is an indigenous minority language and, therefore, it is impossible to establish a particular point in time when the impact of another language began, at least without a detailed biographical record.

Importantly, in Lithuanian, the dominant language of the society and the language acquired simultaneously with Yiddish, present perfect auxiliary is optional: jis (yra) išejęs “he has left/gone.” As there are other features in the informants’ Yiddish that can also be attributed to the impact of Lithuanian (recall Table 1), I believe that one should not ignore a possible influence of the Lithuanian model. Clyne (2003, pp. 118–120) provides several examples from Germanic languages in contact with English in Australia and shows that the auxiliary is vulnerable in a contact situation.

Borrowing of DPWs

The DPWs, also known as discourse markers, form a broadly defined class of words that includes conjunctions, particles, question words, and adverbs (on this and various terms referring to this word class see Wertheim, 2003, pp. 153–154). As Wertheim (pp. 153–154) has it, it is notoriously difficult to find a suitable cross-linguistic definition of what comprises the class, and the terms to label the class vary to a great extent (see Wertheim, 2003, p. 155 for an approximate list that different scholars have suggested).

Various explanations have been proposed for the high borrowability of DPWs. Interestingly, I have not found any studies on DPWs within the framework of incomplete L1 acquisition and HLSs. It has been suggested that structural properties, for instance autonomy (see Matras (1998) on borrowability hierarchies), and semantic qualities of DPWs (in the sense that they only modify a proposition but do not change its meaning) are responsible for their borrowability (Maschler, 1994). For instance, some scholars discuss DPWs in the context of contact-induced language change. Indeed, introduction of Russian DPWs in Turkic languages, especially conjunctions, has led to substantial changes in the morphosyntax of the Turkic subordinate clause because Turkic languages have few conjunctions (see Muhamedowa (2009) for Kazakh). Others have noticed competence or identity-signaling and contrasting functions, often in the context of the loss of domains and registers and subsequent language shift (Clyne, 2003, pp. 225–233; see also Wertheim, 2003, p. 144 on what she calls “contracting languages”).

Matras (2009, p. 143) argues that reinforcing contrast and proficiency or signaling identity by the means of borrowed DPWs are not sufficient explanations of borrowing; neither are the structural properties (that is, the syntactic autonomy) of DPWs. He observes that contrast fades eventually once borrowed DPWs replace the native ones. I concur with Matras, especially in light of Auer’s (1999) model of fused lects, where in the ultimate stage code-switching becomes conventionalized and the contrast disappears. Instead, Matras (1998, p. 282) suggests that their borrowability is conditioned by their specific functions. DPWs come from a pragmatically dominant language, that is, the language that is dominant in “regulating mental processing activities” (Matras, 1998, p. 286). According to Matras (1998, pp. 286, 291), a bilingual speaker may experience a considerable cognitive load of managing two varieties. When such a speaker interacts with other bilinguals, the strain of keeping the two varieties apart may be relaxed. The very nature of utterance modifiers is such that they help to direct the hearer’s attention (Matras, 1998, speaks of “grammar of direction”).

To sum up, explanations for a high borrowability of DPWs come from different points of view: structural; pragmatic; sociolinguistic; and functional–cognitive.

The following list (by no means exhaustive) demonstrates that borrowing/copying of DPWs has been attested in many contact situations involving different languages: starting from Weinreich (1953, p. 30), who commented on a higher borrowability of certain word classes (the English conjunction but in United States’ Yiddish); Salmons (1990) on English DPWs in Texas German; Savić (1995) on English markers in immigrant Serbian; Maschler (1994) on Hebrew markers in the English of American immigrants to Israel; Leinonen (2009, pp. 320–321) on Russian DPWs in Komi; Keevallik (2006) on Swedish markers in immigrant Estonian; and Hlavac (2006) on English DPWs in Croatian spoken in Australia, and many others. Borrowing of Russian DPWs has occurred in many Turkic languages: Uzbek; Tatar; Karaim; and Kazakh (see Wertheim (2003) for an overview, and especially on Tatar). Borrowing of DPWs in itself does not signal a disappearance or shrinking of a variety: for instance, Latvian un “and” and jā “yes” are conventionalized borrowings from German. Some Slavic particles such as take “indeed, exactly,” and xotš/xotse “although” have become integral components of Yiddish; the most northern varieties of NEY have borrowed German ja “yes” that replaced NEY je. In certain case studies, borrowed DPWs have been claimed to exist alongside with the native ones (Hlavac, 2006), yet others have attested a wholesale importation of DPWs and convergence (Salmons, 1990). Apparently, whether native DPWs are replaced completely or partially, or whether DPWs can occur with their borrowed equivalents, cannot be predicted in straightforward terms; wholesale replacement, partial replacement, and co-occurrence per se are probably not indicators of maintenance versus shift.

Let us turn to the present case of Lithuanian Yiddish. In Verschik (2014b), I have looked at conjunctions in Yiddish–Lithuanian bilingual speech. Here, I will present a brief overview of that study. It has to be noted that many conjunctions in Yiddish are already borrowings of either Hebrew–Aramaic or Slavic origin. In the present case, some conjunctions are retained, some are replaced by their Lithuanian or English equivalents (see discussion below), and some vary. I am going to focus on the following points: what DPWs appear in the informant’s speech; whether they have replaced the corresponding Yiddish DPWs; what languages are involved; and what this can possibly mean as far as input and contact-induced change are concerned. Analysis of DPWs from a code-switching perspective remains outside the scope of this study.

Notably, no DPWs from another language appear in I1’s speech; yet, there is one item that requires an explanation. The above-mentioned literature focuses on borrowing in the sense of form-meaning importation (Matras (2009), and Matras and Sakel (2007), refer to this as “MAT-replication” (matter replication); Johanson (2002a), uses the term “global copying”). Still, Keevallik (2006) discusses different processes, such as extension of possible usage of Estonian DPWs in other contexts, modeled on similarly sounding DPWs in dominant Swedish. Similar examples are attested in Verschik (2008, p. 166) where, based on Estonian a(ga) “but,” Russian contrasting conjunction a “but” is used in contexts where another Russian conjunction но “but” would be used (i.e., when the contrast is unexpected). In the similar vein, instead of the Yiddish correlative conjunction saj…, saj… or i…, i… “both…and…” (the latter option is of Slavic origin), the usual un “and” is reduplicated, modeled on Lithuanian ir…ir… “both…and…” in (1a):

(1a) in der hejm mir redn un idiš,     un  litviš

     in the home we speak and Yiddish, and Lithuanian

    “at home we speak both Yiddish and Lithuanian”

    Compare (1a) to monolingual Lithuanian in (1b) and monolingual Yiddish in (1c):

A completely different picture appears in I2’s speech. Partly, Yiddish DPWs have been retained; some are at times replaced with their English equivalents, and others have been completely substituted with equivalents from Lithuanian and English.

The general picture is as follows. During the interview, 86 tokens of non-Yiddish DPWs were produced. These 86 DPWs comprise conjunctions and some particles (yes, not). Certain DPWs cannot be clearly classified as Lithuanian, Yiddish, or English because of their similar shape in the languages or because of an unclear realization by the informant: his English has a Lithuanian accent. For instance, the particle for “but” is sometimes closer to Lithuanian bet “but” and sometimes to English but. An overview of all DPWs is presented in Table 4.

OPEN IN VIEWER

Table 4.

Non-Yiddish discourse-pragmatic words (DPWs) in the older informant’s speech.

DPWs (language), (tokens)	Yiddish equivalent (tokens)	Remarks
but/bet (ENG/LIT), 29	ober, 0	Ambiguous because of the similar shape
and (ENG), 20	un, 39	In three cases un used as a part of compound numerals: in najn un naintsikn jor “in year (nineteen) ninety-nine”
if (ENG), 17	ojb, 0	The cognates have material similarity and this may have triggered the replacement
yes (ENG), 5	ja, je, jo, 17	NEY variants are ja and je
not (ENG), 4	nit, 79	Similar shape facilitates variation
as (ENG), 3	vi, 0
o(r)…, o(r) (ENG), 3	oder…, oder…, 0	English either…or used according to the Yiddish model (oder “or”)
kad “that” (LIT), 5	az, 0

Note: ENG, English; LIT, Lithuanian.

Table 4 shows that some conjunctions do not occur in Yiddish at all (ober “but,” ojb “if,” oder…oder “either…or,” and az “that”) and have been replaced; others co-occur in several languages. Material similarity facilitates use of non-Yiddish DPWs, for instance, not and Yiddish nit, and, and Yiddish un (Clyne 2003, pp. 163–165 on the concept of facilitation). Fluctuation between Lithuanian bet “but” and English but is of the same nature. This is not surprising, given that all three are Indo-European languages and Yiddish and English are both Germanic languages and share a lot of lexicon. Lithuanian and Yiddish, on the other hand, share a lot in morphosyntax and have both changed under the Slavic impact. Unlike in the case of Russian conjunctions in Turkic languages, insertion of Lithuanian/English conjunctions does not affect morphosyntax. Observe (2a) and (2b) for English conjunctions and (3a) and (3b) for Lithuanian ones:

(2a)

and mama  gezuxt         but nit  gekent      im    

and mother search: participle (PART) but not can: PART he: accusative case (ACC)

gefinen

find

“and mother searched but was unable to find him”

As we have seen, from the point of view of contact linguistics, borrowing of DPWs is rather common. What is striking in this particular case is the use of English, the language in which the informant is not exactly fluent. At this stage, I have no knowledge about whether any replacement of Yiddish DPWs (by those of Lithuanian or Russian origin) had taken place in the speech of I2’s parents. Whatever kind of input in Yiddish I2 had in his childhood, I find it rather unlikely that English DPWs appeared in his parents’ speech: proficiency in English was not common in Lithuania when I2 was a child. In the interview, he mentioned that he had learned English at school. Thus, English DPWs is something that occurred in his Yiddish later on. It may very well be that the full set of Yiddish DPWs was acquired and then underwent changes. Whether one calls this L1 attrition or contact-induced language change is probably a matter of worldview and choice of research paradigm.