Lost in Between: the Case of Russian Heritage Speakers, Part Two
Ludmila Isurin, Ohio State University, and Tanya Ivanova-Sullivan, University of New Mexico
3.2. Lexical Findings
Our lexical findings are limited to only three phenomena with a relatively high number of occurrences in the HS narratives. They do not show regularity or competence problems across the board; rather, they may manifest certain strategies to cope with lexical retrieval problems in unrehearsed narration.
3.2.1. Lexical Extension
Lexical extensions are those instances in which a single word is generalized to cover various semantically related contexts.16 There were seven instances of lexical extension in the speech samples of four of the HS participants; they are all provided below.
|'The frog took the dog's paw.'|
|'The fish took the fishhook.'|
|'It is hard for him to take the fish with him.'|
|'…and he sees that the turtle took his fishing and he leaves.'|
|'And then the little boy picked up everything that he had with him when he came fishing.'17|
|When||he||swam||away from there||seemed||like||the other|
|'When he swam away from there, it seemed that the other beast had taken his stick.'|
|'And it took the stick again.'|
It is not surprising that this particular Russian verb, vzjat' is being generalized to cover various contexts; the same verb in English ('to take') performs similar functions. A similar generalizing strategy has been used in the L2 narratives but there a different verb, polučit' ('to receive'), was used. This word choice is clearly less successful, given the passive, patient-oriented semantics of polučit' and the different conceptualization of the actions represented by lexical labels of 'take' and 'receive', both in Russian and in English.
3.2.2. Conceptual Shift
Another observation concerns the way heritage speakers expressed emotions, a phenomenon previously noted by Pavlenko (2002) in a study of American Russian bilinguals. According to Pavlenko, in English emotions are usually conceptualized as passive states, whereas in Russian they are conceptualized as inner activities. This conceptual difference manifested itself in the language of Pavlenko's bilinguals and that of our heritage speakers in utterances where a predicate consisted of the verb to be (which is omitted in the present tense in Russian) plus an adjective (cf. the examples below). We applied Pavlenko's notion of "conceptual shift" to the analysis of our HS data and detected five such instances.
|'After the struggle/fight with the turtle, the dog is very frightened.'|
|*испугана [ispugana] ('to be frightened') instead of испугалась [ispugalas'] ('to become frightened')|
|'The boy was surprised and he rejoiced because the turtle lived, and that's all.'|
|*стал радостным [stal radostnym] ('became happy') instead of обрадовался [obradovalsja] ('rejoiced')|
|'…and the little boy with his dog is then standing there with the dog and the dog is very sleepy.'|
|'He is angry at the dog because it bit the turtle.'|
|*злой [zloj] ('angry') instead of разозлился [razozlilsja] , (became angry')|
Polinsky (2000a) argues that expressing certain events descriptively, with the verb to be and an adjective or noun, is a sign of an impoverished lexicon. While we agree that this might be true for her examples (e.g., 'byt' alcoholic,' 'get drunk' for napit'sja), in cases where emotions are expressed we should consider both the nature of the speaker's lexicon and the psychological mechanism of conceptual shift as observed by Pavlenko, since they are most often interrelated. In order to determine which factors played a role in producing the above examples, we must therefore look at a speaker's entire narrative and its lexical variety, including the overall scope of his/her lexicon. We should also take into consideration such psycholinguistic factors as the speaker's family background and degree of exposure to Russian.18
The last two examples in this section may indeed represent both conceptual shift and impoverished lexicon. They were produced by the speaker who was born in the U.S. and who had had the least exposure to Russian of all the HS participants. His lexicon is very small, as evidenced by the fact that he repeatedly uses a pool of "favorite" words throughout his narrative. Furthermore, he often uses English words (e.g., 'turtle,' 'frog'), invents words (polož for položil ('put'), uses incorrect, albeit semantically related, words (rybalka), ('fishing' for 'fishing rod'), and replaces specific words with more general ones (cf. more, ('sea') for reka, ('river') and vzjat' ('take') for podnjat' ('pick up'), or vytaščit' ('pull out'), etc. In addition, his grammatical proficiency is extremely low; his speech exhibits obvious imitation of English syntax (e.g., … i idët, i smotrit na gde sobaka byla ('and he is coming, and he is looking at where the dog was'), oblique cases are almost non-existent, etc.19
3.2.2. Word "Invention"
In this section we analyze the word inventions of the above-mentioned heritage speaker who was born in the U.S. (Participant No. 6). In his narrative, we recorded the use of odd ("invented") forms such as son ('dream') for sonnyj ('sleepy') and polož for položil ('he put'). Recall that this is the same speaker who had difficulty retrieving lexical items, including the Russians labels for 'turtle,' 'frog', and 'fishing rod.' He also produced sentences that were clear syntactic calques from English, e.g. the preposition stranding in the fifth example under 'lexical extension' above. His word-invention strategy presents a possible mechanism for coping with grammatical, semantic and word-formation problems. In other words, by inventing new rules and/or perhaps using rudimentary rules from his mental grammar, this speaker creatively manipulates the following existing morphosyntactic and semantic rules of Russian:
1) Use of short forms of adjectives to express temporary qualities. A standard Russian example would be the short form zanjat ('busy'), (compare the long form zanjatyj ('busy')) in a sentence such as: On segodnja zanjat ('He is busy today'). The word that our heritage speaker produced, son, means 'dream' in Russian and the adjective sonnyj ('sleepy') does not have a short form at all.
2) Use of verbal suffixes (mostly va-) to form imperfectives, e.g. the aspectual pair nalit' (PERF)-nalivat' (IMPERF) ('to pour'). Psychologically, for a speaker who has not mastered all of the various ways of forming aspectual pairs, and is mostly aware of aspectual pairs such as nalit'-nalivat', the most natural way to form a perfective would be to use a shortened form of the imperfective. The speaker could wrongly assume that the lexicalized perfective form položit' is, in fact, imperfective and produce a non-word polož.20
4. In-Group Analysis
In this section we compare the data on the narrative performance of heritage speakers with the information gathered through the background questionnaire. We have plotted the linguistic data against the psycholinguistic data in order to detect relationships between grammatical inconsistencies and such psycholinguistic factors as age of arrival in the U.S. and the amount of exposure to Russian. The preliminary analysis suggested that age of arrival in the U.S. and exposure to L1 were the two primary factors contributing to variability in the participants' performance on the narrative task. However, a closer look at the exposure data shows that that this factor is more complex than simply the daily use of the language at any given time. On the one hand, all but two participants reported that they currently spoke Russian only ten percent of their time daily (Participants No. 1 and 7 reported 30% and 40%, respectively.) This ten percent may have comprised no more than the time our participants spent in Russian language classes, given that none of them lived permanently with their Russian speaking families any more. On the other hand, the questionnaire also revealed that only two participants (Nos. 5 and 6) speak to their parents predominantly in English, while the remaining 5 speak Russian. In addition, one participant (No. 4) spends every other summer in Russia. Thus, because of the inherent difficulties in defining and quantifying "exposure," age of arrival in the U.S. was our primary focus of analysis.
The heritage learners' age of arrival in the U.S. varied between 0 and 10, with only two participants (Nos. 2 and 3) having had 2 and 3 years of formal Russian instruction in Russia, respectively. At the time of the study all participants were enrolled in 2nd/3rd year Russian classes, which in theory corresponds to intermediate language proficiency. Since proficiency level was not measured in this study and given that daily exposure to Russian was reported as mostly comparable across the group, age of arrival in the U.S. was used as the major variable, but in each case we also considered other factors, such as language spoken to parents, daily use of Russian, and contacts with Russia.
The following two categories of language deviations/errors were chosen for the in-group analysis: aspectual/tense errors and incorrect use of cases. Not only were these deviations the most common among the HS, but they may also be of further interest in the cross-group analysis. In addition, word order patterns were separately analyzed to determine the extent to which age of arrival in the U.S. may have had a detrimental effect on the acquisition of VS word order in Russian. The data will be presented in the following way. First, a correlation will be sought among the three structural variables effect on the acquisition of VS word order in Russian. The data will be presented in the following way. First, a correlation will be sought among the three structural variables (e.g., tense/aspect errors, case errors, and VS use), then each structural variable will be compared with the extralinguistic variable of age.
In Figure 1, case errors, tense/aspect errors and VS use are plotted for each individual participant.
Figure 1. Heritage learners: case errors/aspect errors/VS use
Figure 2 illustrates erroneous performance on tense/aspect with respect to age.
Figure 2. Heritage learners: age versus tense/aspect errors
The above data suggest a possible relationship between age of arrival in the U.S. and aspectual/tense errors. Among the heritage speakers, the age of the first exposure to English ranges from 0 (Participant No. 6 was born in the U.S.) to 10 years old. The only participant without erroneous performance was the one who regularly spends time with family in Russia (Participant No. 4). The rest of the participants did have problems correctly using tense/aspect in Russian, although there is not a linear relationship between erroneous performance and age of first exposure to English. Neither the language used for communication with parents nor age of arrival clarifies the picture. Participant No. 6, who was born in the U.S. and who speaks English to his parents, makes fewer errors (7.1%) than Participant No. 1 (8.3%), who came to the U.S. at the age of 4, speaks Russian to her parents, and has a higher daily exposure to Russian (40% versus 10%). Given that only one participant had no problem with the correct use of tense/aspect in Russian and that that participant frequently lives with a Russian-speaking family in Russia, we hypothesize that intensive regular exposure to the language in the home country is more important for the acquisition of Russian tense/aspect than the age at arrival in the U.S.
Figure 3 shows that case errors are negatively correlated with age of arrival, with Participant No. 3, who arrived at the age of 10, having no such instances and Participants No. 1, 6, and 7 having 3.7%, 4.7%, and 5.7% error rates, respectively. However, it should be noted that Participant No. 5, who came to the U.S. at the age of 6 and speaks English to his parents, showed no case errors. This again suggests that age of arrival cannot be considered separately from other extralinguistic factors, such as the amount of daily exposure to L1 and the language spoken in the family.
Figure 3. Heritage learners: age versus case errors
The next chart (Figure 4) portrays reverse word order plotted against the age of participants' first exposure to English. It is notable that age is positively correlated with the acquisition of reverse word order in Russian. Early exposure to narrative patterns in Russian, in which VS order predominates in story-telling, may explain the fact that the participants who were either born in the U.S. or came to the country under the age of six used the VS order less frequently or not at all (Participants No. 1 and 6). To summarize, we suggest that the acquisition of cases and VS word order may depend on age of arrival in the U.S., while the acquisition of tense/aspect may depend more on regular and extensive exposure to the target language than on the age factor.
Figure 4. Heritage learners: age of arrival versus VS use
Further analysis of the questionnaire revealed that all participants are highly motivated to maintain their Russian and that their parents and family heritage encourage their interest in the language. All but one reported writing as their most problematic area. Five participants admitted having a problem with remembering Russian words, correctly conjugating verbs and using the correct case endings. This shows a high metalinguistic awareness among the participants in the heritage learner group. All 7 speakers also reported English as the language they think and dream in and the language they find easier to use when they are tired. Only one participant reported Russian as the language she switches to when she is emotional or feels like cursing. The rest chose English. This additional information suggests that the age of first exposure to the second language and the amount of daily exposure to the native language, despite their importance, should be considered along with an array of other factors that shape the ultimate performance of heritage speakers.
5. Cross-group Analysis
Further analysis of the data was focused on comparing the performance of the heritage speakers with that of the two other participating groups, i.e., English speakers who have studied Russian as an L2 and monolingual native speakers of Russian (the latter will be referred to as the control). Our goal is to present a more complete picture of the heritage language by comparing it to that of the other two groups. Thus, the heritage speakers in this study are seen as belonging to a continuum of speakers and not as isolated individuals. Such a comparison has been made before but with a primary focus on comparing heritage speakers and native (or monolingual) speakers. We argue that our incorporation of L2 learners offers a more detailed picture of the speaker continuum.
The same categories examined for in-group analysis were also chosen for cross-group analysis, i.e. aspect, case, and VS word order (Table 2). We limit our discussion to summarizing the number of errors/type of word order. No detailed analysis will be given on the performance of monolinguals and L2 learners. The rate at which participants made errors with verbal aspect or cases, as well as those instances where they used the VS word order, are represented in Figure 5.
Table 2. Cross-group Comparison
VS Word Order
Figure 5. Cross-group comparison
Needless to say, there were no tense/aspect or case errors in the control group. As is clear from the graph (Fig. 5), the L2 learners made 3 times as many aspectual errors than the heritage speakers (12.7% and 4.1%, respectively), and almost 3 times as many case errors (6% versus 2.4%, respectively). It should be noted that the aspectual errors made by the L2 learners were different in kind from those made by the heritage learners. It is clear from the L2 learners' narratives that they had not mastered the use of aspectual pairs or that they lacked knowledge of a specific form (mostly perfective). As for word order, heritage speakers used VS twice as often as L2 learners but 3 times less than the control group (2.11%, 1%, and 6%, respectively). The lower percentage of VS sentences produced by HS as opposed to monolingual Russians could be interpreted either as a language transfer from English or as a solution to the higher cognitive load of a VS sentence (Isurin, 2005).
When VS usage among the L2 learners is matched with the length of the L2 acquisition period, it becomes apparent that the use of Russian reverse word order increases with exposure to the language. The four L2 participants who did not use VS at all had spent only two years on L2 acquisition.21 As was previously mentioned, the acquisition of VS word order seems to depend on age of arrival in the U.S. for heritage speakers, which can be translated into the number of years spent on Russian language acquisition. Thus, we suggest that length of time spent learning the language can be crucial for the acquisition of VS order for both groups. But given the small scope of the present study, we refrain from any far-reaching conclusions.
As the title of our paper suggests, Russian heritage speakers may indeed be "lost in between" in the continuum of language speakers. They outperform English-speaking learners of Russian in such linguistic areas as the correct use of aspect/tense and cases, but they fall well behind the native speakers in these same areas. Moreover, they seem to be more comfortable than L2 learners in using the predominantly VS word order of Russian narrative discourse. At the same time, their mastery of that category is far from that of the native speaker. These findings are not accidental but rather are determined by the linguistic uniqueness of the population whose first language was either incompletely acquired or underwent a certain changes due to L2 influence.
The findings reported in the present study clearly show a continuum of speakers, ranging from "perfect" (the control monolingual group) to "imperfect" (the L2 learners), with the heritage speakers somewhere between the two. This observation is valuable for describing HS language more adequately and in greater detail by separating into morphosyntactic categories, while keeping in mind the bigger picture of grammatical and lexical competence.
Clearly, HS language is different from that of monolinguals and the language of L2 learners. Our paper attempted to show specific characteristics of this language as well as a possible correlation between language deviations and psycholinguistic factors such as amount of exposure to L1, age of arrival in the L2 country, etc. Given the heterogeneous nature of the Russian HS group in the U.S., and the differences in their backgrounds and language attitudes, this approach seems warranted. The results of the present study indicate that larger scale research is needed. We suggest that the pool of participants be extended, additional rigorous tasks be included, and different skills be tested, e.g., writing, speaking (interviews, film narration), translation, etc.
Our findings have implications for teachers of Russian as a foreign language, general linguists, sociolinguists, and psycholinguists who wish to attain better insight into first language learning and first language re-acquisition by Russian immigrants in the U.S.
The authors express their gratitude to two anonymous reviewers for their valuable comments on an earlier draft of this paper. They would also like to thank Andrea Sims for editing the original version of the paper.
Andrews, D. R. (1999). Sociocultural perspectives on language change in diaspora. Amsterdam: John Benjamins.
Andrews, D. R. (2000). Heritage learners in the Russian classroom: Where linguistics can help. ADFL Bulletin, 31(3), 39-44.
Bermel, N., & Kagan, O. (2000). The maintenance of written Russian in heritage speakers. In O. Kagan & B. Rifkin (Eds.), The learning and teaching of Slavic languages and cultures, pp. 406-436. Bloomington, IN: Slavica Publishers.
Isurin, L. (2005). Cross linguistic transfer in word order: Evidence from L1 forgetting and L2 acquisition. In J. Cohen & K. McAlista (Eds.), ISB4: Proceedings of the International Symposium on Bilingualism, pp. 1115-1130. Somerville, MA: Cascadilia Press.
Kagan, O., Akishina, T., & Robin, R. (2002). Russian for Russians. Bloomington, IN: Slavica Publishers.
Ke, C. (1998). Effects of language background on the learning of Chinese characters among foreign language students. Foreign Language Annals, 31, 91-100.
Kondo-Brown, K. (2002). K. Family and school factors influencing academic performance of bilingual Shin Nisei students in Hawaii. In E. Tamura, V. Chattergy, & R. Endo (Eds.), Asian and Pacific Islander American education: Social, cultural, and historical contexts, pp. 149-74. South El Monte, CA: Pacific Asia Press.
Olshtain, E., & Barzilay, M. (1991). Lexical retrieval difficulties in adult language attrition. In H. Seliger & R. Vago (Eds.), First language attrition, pp. 139-151. Dordrecht: Foris Publications.
Pavlenko, A. (2002). Conceptual change in bilingual memory: A neoWhorfian approach. In F. Fabbro (Ed.), Advances in the neurolinguistics of bilingualism, pp. 69-94. Udine: Forum.
Polinsky, M. (2000a). The composite linguistic profile of a Speakers of Russian in the U.S. In O. Kagan & B. Rifkin (Eds.), The learning and teaching of Slavic languages and cultures, pp. 437-465. Bloomington, IN: Slavica Publishers.
Polinsky, M. (2000b). Language acquisition: Full and incomplete scenarios. Paper presented at the SLING2K Workshop, Bloomington, IN. Available at http://www.seelrc.org/glossos/issues/8/polinsky.pdf
Polinsky, M. (2008). Heritage language narratives. In D. Brinton, O. Kagan, & S. Bauckus (Eds.), Heritage language education: A new field emerging. New York: Routledge. Available at http://ling.ucsd.edu/~polinsky/pubs/frog%20stories.pdf.
Zemskaja, E. A. (Ed.) (2001). Jazyk russkogo zarubež'ja [The language of the Russian diaspora]. Moscow/Vienna: Wiener Slawistischer Almanach. (Wiener slawistischer Almanach 53.)
Notes, Part 2
14. We adopt the term 'reference-tracking' from Polinsky's discussion of null copying across clauses in Polinsky (2000b, p. 23). (back)
15. Cf. Polinsky's conclusion about the avoidance of null copying in AR: "Overall, it seems that the elimination of null copying is due to the general increase of redundancy rules observed in American Russian: the speaker who lacks confidence that the message will be parsed and decoded properly, introduces more 'instructional' elements that are supposed to guide the hearer in the processing" (Polinsky, 2000b, p. 23). (back)
16. Cf. the same phenomenon observed by Bermel and Kagan (2000) in HS writing. (back)
17. This sentence is also a syntactic calque from English involving preposition stranding (or extraction), almost always impermissible in standard Russian. In this utterance the speaker obviously imposes the English mechanism of extraction onto Russian. However, he is also trying to mask it by not allowing the overt preposition c 'with' at the end of the sentence, but rather implying it as a silent copy. A total of six syntactic calques appeared in the HS data, distributed among five of the participants. (back)
18. We employ this procedure not only in this section but also in the other sections, which deal with aspect, case, word order, etc. We do not describe every single heritage speaker the way we do in this section due to space limitations, but we do present general findings at the end of the paper. (back)
19. For further analysis of the relationship between lexical proficiency and grammatical level, cf. Polinsky (2000a). (back)
20. The perfective form of this particular verb is lexicalized for this speaker, perhaps due to its greater frequency in regard to accomplished (telic) actions. (back)
21. Unlike the other L2 speakers in the study, these four had probably not yet attained a proficiency level of intermediate-high. (back)
Published: Tuesday, January 08, 2008