1.1 Aspect

The category of aspect exists in both English and Russian, yet its encoding, functions, and meanings differ in the two languages. In English, syntactic aspect functions in combination with tense, presenting events as complete wholes (simple aspect; e.g., ‘I study at Yale’) or as actions in progress (progressive aspect; e.g., “I am studying for an exam right now”), and placing more or less completed (perfect aspect; e.g., “I have finished studying”) and incomplete events (perfect progressive aspect; e.g., “I was studying very hard”) at a prior point in time. Importantly, syntactic aspect is not an intrinsic characteristic of English verbs and is not marked in the infinitive.

In Russian, on the other hand, aspect is a syntactic category, independent of tense, and characterizes all forms of the verb, including infinitives, imperatives, and participles. This means that Russian speakers are required to mark verb aspect, regardless of whether the marking contributes to the meaning of the sentence (Zalizniak and Shmelev, Reference Zalizniak and Shmelev2000; Bondarko, Reference Bondarko2003). All Russian verbs, with the exception of быть (to be), belong to one of two aspectual categories: imperfective which refers to the process or state (e.g., идти [to be walking]) and perfective which refers to achievement or accomplishment (e.g., пойти [to start walking somewhere]). Many verbs constitute aspectual pairs, with one perfective and one imperfective member (e.g., идти–пойти). Perfective verbs are commonly used in two tenses (past and simple future) and imperfective verbs are used in all three tenses (past, present, and future with auxiliary verbs).

These aspectual differences do not correspond directly to tense–aspect distinctions encoded in English. For instance, Russian imperfective aspect encodes only some of the meanings of the English progressive, more specifically that of unfolding action either in the past or in the present. Russian also relies on imperfective verbs to describe repeated actions in the past (Мы ходили в школу [We walked/went to school]), whereas English in the same context relies on simple past tense markers or on additional lexical means (e.g. We used to walk/go to school).

As a result of these differences, each English motion verb may have at least two corresponding Russian verbs, a correspondence that creates major difficulties for English-speaking learners of Russian (Driagina, Reference Driagina2007). Polinsky's (n.d., Reference Polinsky, Brinton, Kagan and Bauckus2008a, Reference Polinsky, Corbett and Noonanb) studies of heritage speakers of Russian in the US identify aspectual distinctions as one area where simplification takes place in the Russian verb system.

1.2 Directionality

Directionality is a unique characteristic of Russian motion verbs indicating whether the action proceeds in a single or in multiple directions (Muravyova, Reference Muravyova1986; Mahota, Reference Mahota1996; Zalizniak and Shmelev, Reference Zalizniak and Shmelev2000). Unidirectional or determinate verbs refer to motion proceeding in a forward direction, from point A to point B (e.g., идти по улице [to be walking down the street]). Multidirectional or indeterminateverbs describe motion with more than one destination, aimless or random motion, and habitual or repeated motion (e.g., ходить по комнате [to walk back and forth around the room]; ходить в школу [to go to school every day]). As a consequence, a single English verb may have four corresponding Russian verbs, with a unidirectional and multidirectional verb in each aspect (e.g., идти-ходить [to walk somewhere – to walk back and forth, imperf], пойти–походить [to go/walk somewhere – to have walked back and forth, perf]). These differences are challenging for English-speaking learners of Russian who often confuse members of aspectual and directional pairs (Driagina, Reference Driagina2007). They also open up possibilities for simplification of the Russian system in contact with English documented in the speech of heritage Russian speakers (Polinsky, n.d., Reference Pavlenko, Wei and Moyer2008a; Bermel and Kagan, Reference Bermel, Kagan, Kagan and Rifkin2000).

1.3 Path

The two languages also differ in how they mark path of motion. In English, path is mostly marked through prepositions (e.g., climb in), in a few cases it is also encoded in the verb itself (e.g., enter). In Russian, path may be marked through both prepositions and prefixes. Frequently, the two perform the same function (e.g., влезть в [(in) climb into]). At times, however, prefixes and prepositions serve different functions, with the former marking aspect and the speaker's point of view and the latter path. This multifunctionality results in a very complex system of meanings whereby the meaning of one English utterance may be rendered in different ways, depending on the speaker's point of view:

The same Russian lemma with different prefixes and the same preposition may also require completely different verbs and prepositions in translation:

While there is some correspondence between prototypical or core meanings of the key Russian prefixes and prepositions and the English pre- and postpositions (e.g., в – in), this correspondence is limited, creating challenges for English-language learners of Russian and Russian-language learners of English, as well as opportunities for attrition and simplification.

1.4 Manner of motion

Despite their common categorization as satellite-framed languages, manner of motion is another area where the two languages differ in important ways. English has several high frequency generic verbs of motion, such as to go, to get, and to come, which are unmarked for manner. In Russian, on the other hand, there is a very limited number of verbs that are not marked for manner of motion, such as прибыть (to arrive) or отправиться (to set off). Russian lacks translation equivalents of to go, to come, and to get and requires its speakers to differentiate systematically between walking on foot (идти/ходить) and driving/riding (ехать/ездить). This difference also opens up possibilities for simplification, and both Polinsky (n.d.) and Andrews (Reference Andrews2001, Reference Andrews2004) noted that under the influence of English, heritage speakers of Russian may lose the obligatory distinction between walking and driving/riding and resort to one set of verbs, commonly идти/ходить (to walk), to refer to these motion events.

Given these differences between the two languages, we can now proceed to ask whether the Russian verb system remains intact or undergoes changes in contact with L2 English in late Russian–English bilinguals.

2.1 Participants

The narratives analyzed in the present study were collected from native speakers of Russian (L1 Russian corpus) and English (L1 English corpus) and from Russian–English bilinguals (Bilingual corpus). Clearly, contributors to both the L1 Russian corpus and the Bilingual corpus are native speakers of Russian. To differentiate between them the first group is labeled “L1 Russian speakers” and the second “bilinguals”. Contrary to the prevalent tradition, I decided not to refer to L1 speakers of English and Russian as “monolinguals” – all had exposure to foreign languages in secondary or higher education; some also had exposure to a second language in their environment. To minimize potential influence of these languages, only speakers who reported minimal competence in foreign or second languages were selected for participation in the study.

To take into account linguistic variation that stems from geographic, socioeconomic, and socioeducational factors, participants were recruited among students in two types of universities and in three geographic locations in each country: (a) prestigious elite universities (University of St. Petersburg in Russia; Cornell University in the US) and (b) large state universities (Tomsk State University and Khabarovsk State Pedagogical University in Russia; Pennsylvania State University and Temple University in the US). The three corpora were collected at these universities in the course of three studies (see Table 1 for an overview).

Table 1. Distribution of participants across narrative tasks in the three corpora.

Bilingual Russian corpus

The Bilingual Russian corpus, consisting of 94 narratives, was collected from 70 late Russian–English bilinguals; in addition, four narratives were collected from four childhood bilinguals. All late bilinguals reported Russian as their native language and English as a second language. Some also reported basic knowledge of a second language (Kazakh, Ukrainian, or Latvian) and/or a foreign language (French or Spanish). The following participants took part in the studies:

• • Study 1. Twenty-nine narratives were elicited from 27 bilinguals (14 females, 13 males), aged between 18 and 31 years (Mean age = 21.7 years; SD = 3.7), undergraduate and graduate students at Cornell University. Their age of arrival (AOA) in the US varied between 10 and 26.5 years (Mean AOA = 16.1 years; SD = 4.4). The length of exposure (LOE) to English in the US varied between 1 and 17 years (Mean LOE = 5.6 years; SD = 3.2). All participants learned their English upon arrival, some attended middle or high school in the US and others came directly to college. They used Russian with their families, relatives, and Russian-speaking friends, and English with English-speaking friends, as well as for educational and everyday interaction purposes.

• • Study 2. Sixty narratives were elicited from 38 bilinguals (19 females, 19 females), aged between 17 and 34 years (Mean age = 25.1 years, SD = 4.8), undergraduate and graduate students at the Pennsylvania State University. Their AOA varied between 6 and 31 years (Mean AOA = 20.9 years; SD = 5.7). The LOE in the US varied between 6 months and 12 years (Mean LOE = 4.2, SD = 2.7). Most of the participants had studied English in secondary and higher education establishments in Russia and other post-Soviet countries (e.g., Ukraine, Kazakhstan) between 3 and 19 years (Mean = 8.9 years; SD = 4.2) and came to the US as students. All interacted in Russian and English on a daily basis and were recruited through the Russian club.

• • Study 3 (in progress). Five narratives were elicited from five bilinguals (four females, one male), aged between 18 and 37 years (Mean age = 25.8 years; SD = 7.5), undergraduate and graduate students at Temple University. Their AOA varied between 6.5 and 25 years (Mean AOA = 15.9 years; SD = 7.3). The LOE varied between 6 and 14 years (Mean LOE = 10.3 years, SD = 3.2). All participants learned their English upon arrival, some attended middle or high school in the US and others came directly to college. They continued to use Russian with their families and, in two cases, for work-related purposes.

In all three studies, the participants were fluent enough in English to be enrolled in regular undergraduate and graduate classes; none were enrolled in the Intensive English Language Program.

Given the large variation in the three bilingual groups in the age range, AOA and LOE, participants were further subdivided into four groups according to their AOA and LOE (see Tables 2 and 3). In terms of AOA, the groups were split as follows: (i) 6–13, participants who arrived as younger children and completed most of the secondary school in the US; (ii) 14–17, participants who arrived as teenagers and spent some time in a US high school; (iii) 18–22, participants who completed secondary school in Russian and attended college in the US; (iv) 23–31, participants who completed both secondary school and college in Russian and arrived in the US as graduate students.

Table 2. Distribution of Russian–English bilinguals across four groups in terms of AOA.

Table 3. Distribution of Russian–English bilinguals across four groups in terms of LOE.

The distribution of participants in Table 2 shows that the Cornell and Penn State corpora are distinct in terms of AOA (the Temple corpus is too small at this point to be included in this comparison): The majority of Cornell participants are early arrivals (ages 6–17, 70%), while the majority of Penn State participants are late arrivals (ages 18–31, 74%). The two groups are also distinct in LOE: Almost half of the Cornell corpus consists of participants who spent between 7 and 12 years in the US (44%), while in the Penn State corpus almost half of the participants spent between 6 months and 3 years in the US (47%). These differences were taken into consideration in the analysis.

Because I was interested in examining areas where L2 influence on L1 begins, I chose to focus on late bilinguals. To make my discussion comprehensive, however, I will also consider narratives elicited from four heritage Russian speakers who volunteered to participate in the study. Three of the participants were age 19 and one age 24. Their AOA varied between 6 months and 3 years (Mean AOA = 1.9; SD = 1.0) and LOE between 16 and 23 years (Mean LOE = 18.2; SD = 3.2). All four grew up speaking mostly Russian at home and mostly English outside the home. Their narratives were excluded from the bilingual corpus because as heritage speakers they may be incomplete acquirers.

2.3 Procedure

Each participant was interviewed individually. In Study 1, L1 speakers were asked to recall two films with different contents (Ithaca Story/The Letter or Kiev Story/Pis'mo). They were shown the videos in randomized order and after each screening were asked to recall the film speaking directly into the taperecorder. Due to scheduling constraints, some L1 English speakers recalled only one film, which explains a higher number of participants in this group. Bilingual participants recalled one film in Russian and one in English. In all but two cases language order was randomized. Two participants expressed preference for Russian and produced both narratives in Russian, hence 29 Russian narratives were produced by 27 participants.

In Study 2, L1 speakers were asked to recall three films (The Letter/Mr. Bean 1/Mr. Bean 2), their recalls were taperecorded at the end of each screening. Unfortunately, a tape with ten L1 Russian narratives of Mr. Bean 1 was damaged prior to transcription, hence only 19 Mr. Bean 1 narratives were available for analysis. Bilingual participants were asked to recall two films in Russian and one in English or vice versa, with language choice and order randomized. Once again, however, some participant preferences were taken into account and the numbers of narratives elicited by each stimulus are somewhat different (22 narratives were elicited by The Letter, 20 by Mr. Bean 1, and 18 by Mr. Bean 2).

In Study 3, participants were asked to narrate the story portrayed in the picture book Frog, where are you (Mayer, Reference Mayer1969), as they looked at the pictures.

All recalls were taperecorded, transcribed in the original language and analyzed for the uses of verbs of motion.

2.4 Data analysis

To ensure that all corpora produced sufficient numbers of motion verbs, the narratives were first analyzed in terms of narrative length and proportion and number of motion verb tokens across the corpora. Then, two types of analysis were performed on the motion lexicon. The first involved identification of deviations from standard Russian usage. To do that, one first has to define what counts as standard Russian, as a verb of motion, and as deviation. Russian linguists working on L1 attrition, in particular Zemskaia (Reference Zemskaia2001, Reference Zemskaia2004) and Glovinskaia (Reference Glovinskaia2004), argue that literary Russian and the speech of educated inhabitants of Russian cities, and more specifically Moscow and St. Petersburg, constitutes the standard, while the speech of native Russian speakers born in Ukraine, Belarus, or Kazakhstan, diverges from the standard. Both dialectal variation and instances of code-switching and lexical borrowing are seen as deviations in this approach, with code-switching and lexical borrowing also taken as evidence of attrition.

Such a view of the standard is problematic for several reasons. First of all, linguists who adopt this approach draw on their own native speaker intuitions and on literary examples to judge whether particular usages conform to the standard. Yet, as Polinsky (Reference Polinsky2004) points out, codified standard literary Russian differs in several aspects from the spoken varieties and should not be used as a baseline comparison. Second, similar processes, in particular lexical borrowing from English, are now taking place both in diaspora and metropolis Russian (cf. Ryazanova-Clarke and Wade, Reference Ryazanova-Clarke and Wade1999), thus, lexical borrowings are not necessarily evidence of attrition – rather, they may be evidence of conceptual enrichment and expansion of the lexical repertoire (see also Pavlenko, Reference Pavlenko, Schmid, Köpke, Keijzer and Weilemar2004).

In the present study, the issue of the standard is resolved through the corpus-based approach, where lexical choices made by bilingual speakers are compared to those made by native speakers of Russian in the same context. This approach allows us to judge bilinguals against a real, rather than an idealized, reference group. To take into account geographic variation, both groups included some participants who grew up as native Russian speakers outside of Russia, in Belarus, Kazakhstan, Latvia, and Ukraine (on Russian in post-Soviet countries, see Pavlenko, Reference Pavlenko2006, Reference Pavlenko2008c).

Only verbs with inherent motion-related semantics referring to an agent or patient changing position or location counted as verbs of motion in the present corpus (e.g., to drop, to carry, to jump, to roll, to swim); verbs, such as to open or to close, that may or may not involve such change did not count as verbs of motion.

To decide on what counts as deviation, the stimuli were divided into segments with particular motion events (e.g., a woman sitting down, a deer throwing a boy off the cliff, etc.). These events were matched with lexical choices in the L1 Russian, L1 English and bilingual corpora. Items counted as deviations when they satisfied two conditions: (a) a morphosyntactic or pragmatic error was committed, and (b) no L1 Russian speakers made the same lexical choices. These instances further counted as evidence of L2 influence on L1 if they patterned with lexical choices made by L1 English speakers.

This corpus-based approach allowed me to eliminate some items that might otherwise have been judged as erroneous. For instance, one bilingual participant narrating the Frog Story stated that the characters махают [are waving], violating the standard Russian conjugation pattern (машут). This violation was not counted as a deviation because similar violations were also found in the L1 Russian corpus. On the other hand, morphosyntactically correct verb tokens counted as deviations if they diverged from L1 Russian speakers’ choices in a particular context and thus violated semantic and pragmatic constraints of Russian. For instance, lexical choice in the utterance она ходила по парку [she was walking around the park] was judged as erroneous because the speaker used a multidirectional verb in the context where L1 Russian speakers used unidirectional verbs to refer to the character walking in a specific direction.

The second analysis involved examination of lexical diversity in the motion lexicon of bilingual participants. As discussed earlier, because Russian makes distinctions that are not encoded in English, L2 influence on L1 may be displayed as the loss of lexical diversity. To examine this possibility, I first compared the corpora in terms of verb types, using the type–token ratio (TTR; on limitations of this approach, see Dewaele and Pavlenko, Reference Dewaele and Pavlenko2003). To equalize Russian and English, I counted as types only basic stems, that is motion verbs without prefixes, prepositions, and postpositions. Russian aspectual verb pairs, as well as directionality verb pairs, counted as a single stem in this analysis (e.g., to walk vs. идти/ходить/пойти/уйти/зайти, etc.).

To examine the use of aspectual, directionality, manner, and path distinctions, I then compared L1 Russian and bilingual corpora in terms of specific subsets of verbs, which included prefixed derivatives of the main verb pairs, such as идти/ходить [to walk], плыть/плавать [to swim], or ехать/ездить [to drive/ride]. Thus, if in a particular corpus L1 Russian speakers used a variety of prefixed derivatives of a particular set of verbs, these choices were compared to those made by bilingual speakers, looking for evidence of simplification of the motion verb repertoire.