This doctoral dissertation focuses on Russian-English bilingualism and code-switching in New York City, applying the main functional models in code-switching research to data gathered in a field study the author conducted between 1998 and 2000. The data are quantified, and an attempt is made to correlate the linguistic competence of the speakers with their code-switching behavior.

The book consists of seven chapters. Chap. 1 deals with the history and recent developments in research on code-switching, the most visible form of bilingualism. “Code-switching” (CS) as the alternation of languages in conversation was first coined by Haugen 1956, but Weinreich 1953 had earlier written about a “switch” in language depending on the situation. CS research differentiated into a structural approach that analyzes the grammatical features and a functional approach. The latter is taken by Blom & Gumperz 1972, who describe the contextualization function of CS. Auer's (1984, 1995) sequential and Myers-Scotton's (1999) Marked/Unmarked Model are also used. Since the late 1980s, CS has been established as an independent field of study with its own publications and the founding of the European Science Foundation on Code-switching and Language Contact.

In chap. 2, the field study that provided the empirical basis is introduced. After first contacts and questionnaires, 50 half-structured and 50 unstructured interviews were conducted. Since code-switching is an in-group form of communication, a network-based study was essential, followed by informal “kitchen sessions.” The author observed three networks: a group of Russian-English adolescents who were rehearsing a Pushkin production in a theater; a group of 10 friends (including the author), all but one of them students, who met often on weekends and during school holidays; and a network consisting of 8 young adults. Because of problems of recording, only the second network was the object of further analysis. The transcriptions were made in a mixture of Russian and English alphabets following conversation analytic conventions of representation with everyday orthography. The majority of the conversation recorded was in Russian. Almost all of the speakers spoke English with a Russian accent. When English words were inserted into Russian speech, verbs carried Russian prefixes and/or suffixes so they fitted into the Russian verb system (e.g.,

for ‘push’,

for ‘cut’).

The third chapter describes the sociolinguistic background of the Russian-English bilingual community, involving a long and complex history of immigration to the U.S. In 1994, approximately 13,000 Russian speakers lived in New York, primarily in Brighton Beach and the bordering districts. Of these, 92% came with refugee visas, the majority being Jewish. The strong and rapid upward mobility of the last wave of Russians is a striking fact, with good knowledge of English being one of its main conditions.

The fourth chapter gives an overview of theoretical concepts of bilingualism research. The author concludes that the observed speakers have not been in the U.S. long enough to develop diglossia, but there was a certain possibility of English emerging as the High language and Russian as the Low language The Russian-speaking community in NYC is described as a bilingual community in a society dominated by a monolingual mainstream. The community in itself is very heterogeneous, with many members from non-Russian-speaking parts of the former Soviet Union (e.g. Ukraine, Belarus) and thus bilingual even before entering the U.S.

The fifth chapter examines the most influential theoretical constructs in the functional branch of CS research and applies them to the data of the text corpus. For instance, Auer 1999 describes four types of CS: discourse-related alternational CS, discourse-related insertional CS, participant-related alternational CS, and participant-related insertional CS. Borrowings of single lexemes occur with CS in the speech of monolinguals, but CS is a constraint on bilingual speech. The author decides to consider all single lexemes transferred from one language to the other as “insertions” and codes them as CS in the text. Participant-related insertional CS concerns the participant's competence; discourse-related insertional CS is used because the speaker intends to make the utterance coherent with the context. In the corpus, more participant-related insertions than discourse-related ones are counted. According to the “Triggering model,” insertions of single lexemes are “triggers” for the switching of longer stretches of discourse.

The sixth chapter analyzes structural and discourse features of Russian-English CS. The individual choice of language of the 10 Russian-English bilingual speakers is examined using conversation analysis. The quantification of the data shows tendencies of the CS behavior of the speakers in relation to their linguistic competence. Under the mostly English insertions into the Russian Matrix Language, the author differentiates between Cultural Borrowings that represent objects or concepts new to the Matrix Language culture (e.g.,

‘Manhattan’) and Core Borrowings that have viable equivalents in the Matrix Language (e.g.,

‘project’). The data show cultural borrowings and a few insertions on the way to become core borrowings. Thus, the Russian lexicon takes on more and more lexemes from the dominant English language, starting with the semantic fields to which the speakers have the greatest exposure (e.g. university). The speakers displayed differences in linguistic behavior according to their linguistic competence and language preference: The speakers whose dominant language was Russian but who were almost equally competent in both Russian and English tended to use more participant-related alternation in the English-Russian direction. Speakers who were equally competent in both languages with no specific preference for Russian tended to use more discourse-related CS in the Russian-English direction. Speakers with dominance in Russian or preference for Russian tended to use more insertional CS than alternational CS; that is, the more balanced the bilinguals, the more they preferred alternational CS and avoided insertions. When the two parts of the group were mixed, the more balanced speakers tended to accommodate to the speakers who were Russian-dominant or preferred Russian; they used more insertional CS, generally alternated less, and used Russian almost exclusively. Drawing on Auer's language-alternation phenomena (1999), the group was described as being at a very early stage of code-switching. Chap. 7 gives a detailed summary and draws conclusions.

Although this work fills a big gap with its detailed empirical approach to CS and appropriate use of a network study, I do have some critical remarks. Regarding the theory, in contrast to Myers-Scotton's markedness/relational choice model (1999; Myers-Scotton & Bolonyai 2001), which is said to rely on a large amount of sociolinguistic background, Auer's approach to conversation analysis is seen as starting with the text and thus to be based on more solid ground. Thus, Auer's sequential model is given preference over Myers-Scotton's relational choice model. But then it is unclear why Gregor predominantly uses Myers-Scotton's concept of a “Matrix Language” that refers to the notion of “competence” in generative linguistics and argues psycholinguistically. The theoretical mixture of a conversation analytic approach based on empirical and observable data concerning the “performance” of the speaker, and a psycholinguistic approach based on background knowledge about the speaker and his competence, is difficult to apply to the empirical base.

A less important point of my critique concerns the transcription. After a discussion of the use of the Russian or English alphabet, the author decides to use both mixed together to make the switching point more visible. In CA the transcription is essential, but one of the main rules is – and should remain – that interpretation should be avoided at the stage of transcription. The decision to categorize a word as belonging to one or the other language should have been made later. Since even Russians often use the roman orthography RUSLIT for e-mailing, transcribing the Russian language with the roman alphabet here would have worked pretty well, and the decision for English would have been less problematic. Also questionable is the transcription of spoken forms that do not closely correspond to the written form: Whereas in the English parts these spoken forms are reflected in the transcript, the author decided in the Russian parts to exclude forms that were not prominent.

The most important result of this linguistic research concerns the connection between the different kinds of CS and the linguistic competence of the different speakers. It is very interesting to see the effect of mixing the balanced with the unbalanced speakers. The speakers most competent in both languages used more participant-related CS in the Russian-English direction to accommodate to the interlocutor, and the others used a more discourse-related CS in the Russian-English direction. Speakers with a preference for Russian tended to use more insertional CS, whereas the more balanced bilinguals preferred alternational CS and even avoided using insertions. This knowledge helps to explain the function of CS and thus advances its study.