Multilingualism in Lithuanian cities: aims and outcomes of a home language survey in Vilnius, Kaunas and Klaipėda

On the basis of the experience gained from the Multilingual Cities Project, carried out in 6 Western European multicultural cities (Extra & Yağmur 2004), a home language survey was carried out at almost all primary schools in Vilnius, Kaunas and Klaipėda. The total sample consists of almost 24,000 pupils, most of them in the age range of 8-10 years old. After an introduction to the aims of the project, the design of the questionnaire and the collection, processing and analysis of the resulting data, information will be provided on the size and composition of the sample and the distribution of reported home languages. The top-9 of reported languages contains the vast majority of all home languages referred to, i.e., Lithuanian, Russian, English, Polish, German, Belarusian, French, Ukrainian, and Latvian. For these 9 languages groups, crosslinguistic perspectives will be offered on language profiles and language vitality in terms of reported language proficiency, language choice, language dominance, and language preference. Pseudolongitudinal perspectives will be offered for each of these four dimensions in the age range of 8-10 years old.

The aims of the MCP were to gather, analyze, and compare multiple data on the status of immigrant minority languages at home and at school, taken from crossnational and crosslinguistic perspectives.In the participating cities, from the North to the South of Europe, Germanic or Romance languages have a dominant status in public life.Figure 1 gives an outline of the MCP.For the final crossnational report we refer to Extra & Yagmur (2004).The rationale for collecting, analyzing and comparing multiple home language data on multicultural school populations derives from at least four different perspectives: -From a demographic perspective, home language data play a crucial role in the definition and identification of multicultural school populations; -From a sociolinguistic perspective, home language data offer valuable insights into both the distribution and the vitality of home languages across different population groups, and thus raise public awareness of multilingualism; -From an educational perspective, home language data are indispensable tools for educational planning and policies; -From an economic perspective, home language data offer latent resources that can be built upon and developed in terms of economic opportunities.
Home language data put to the test any monolingual mindset in a multicultural society and can function as agents of change (Nicholas 1994) in a variety of public and private domains.From an educational perspective, it remains a paradoxical phenomenon that language policies and language planning in multicultural societies often occur in the absence of basic knowledge and empirical facts about multilingualism.
The Lithuanian home language survey was part of a larger project called "Language Use and Ethnic Identity in the Major Cities of Lithuania".The main goal of the larger project was to examine the relationship between language use, language attitudes and peoples' ethnic identity in the largest cities of Lithuania (Ramonienė 2010).The project was financed by the Lithuanian State Science and Studies Foundation.In order to reach the main goal, the three largest Lithuanian cities were selected for data collection and data analysis, i.e., Vilnius, Kaunas and Klaipėda, with an approximate population size of 554,000, 350,000 and 190,000 inhabitants, respectively.Each of these cities is characterized by particular and distinct multicultural and multilingual populations, and primary school populations are good predictors of future language variation and language loyalty in each of these cities.
There are striking differences and similarities between Western European and the Baltic States in dealing with multilingualism and multiculturalism.What makes the situation in the Baltic States different, is their historical context of dependence on the Soviet Union and on Russian, and their recent accession to the EU (Hogan-Brun, Ramonienė 2005, Hogan-Brun et al. 2009).Having become a member-state of the EU, Lithuania has become part of a new public and political discourse on European vs. national identities and on European vs. national concepts of "integration".

Designing the questionnaire for a home language survey
A number of conditions for the design of the questionnaire needed to be met (Extra, Yağmur 2004, 112-114).The first prerequisite was that the questionnaire should be appropriate for all pupils and should include a question on the repertoire of languages used at home.For the most frequently mentioned languages, a home language profile will be specified.This language profile consists of four dimensions, based on reported language proficiency, language choice, language dominance, and language preference.
A second prerequisite of the questionnaire was that it should be both short and powerful.It should be short (no more than 20 questions in total) in order to minimize the time needed for pupils to complete it during school hours, and it should be powerful in that it should have an optimal and transparent set of questions to be answered by all pupils individually, if necessary -in particular with younger children -in cooperation with the teacher, after an explanation of the aims and the design of the survey has been given in class.
A third prerequisite of the questionnaire was that it had to be compiled in such a way that the answers given by the pupils could be scanned and verified in as automatized a manner as possible, given the large size of the resulting database.In order to fulfill this demand, both hardware and software conditions needed to be met.
The questionnaire was made available to schools and pupils, according to their own preferences, in three versions, i.e., in Lithuanian, Russian, or Polish.The 20 questions were distributed over five different boxes and were formatted for automatic data processing.Below, an outline of the questionnaire is provided in English.
Questions Focus 1-6 Name/code, age, gender, school/town, language of instruction at school 7-9 Birth country of pupil, father and mother 10 Ethnicity ("To which ethnic group do you belong?") 11-16 Language(s) used (most often) at home 17-20 Language learning (at school, before school, TV watching) The pilot and the main study

The pilot study
The rationale for conducting a pilot study was to test the content validity of the research instrument.The questionnaire was derived from the questionnaire used in other European cities (Extra, Yagmur 2004).Some questions were added that had not been used in previous surveys but that are relevant to the situation in Lithuania.One of these questions concerns the self-declared belonging to a particular ethnic group.Both the outcomes of sociological research (Kasatkina, Leončikas 2003) and our discussions with project experts from abroad provided sufficient ground for assuming that the importance of the ethnic dimension to personal identity is weakening in the contemporary world and that also in Lithuania the new tendency of a national/civic dimension is becoming increasingly evident (Leončikas 2007).But it was important to get proof on how ethnicity is understood by primary school pupils in Lithuanian cities, and whether it is possible to associate ethnic self-identification with language(s) used at home.
The question regarding the choice of languages when speaking to different interlocutors was also expanded.According to the results of previous research (Rytų ir Pietryčių Lietuvos gyventojų apklausa 2002, Vilniaus miesto gyventojų kalbų vartosenos įpročiai 2004), in the multicultural regions of Lithuania, in particular in the Southeast of Lithuania, significant changes concerning language behavior have occurred since the restoration of independence of Lithuania and the consolidation of Lithuanian as the official state language.When the status of the language changed, also the level of its knowledge and use in private and public sectors started to change.According to research in Southeast Lithuania and a survey conducted in Vilnius in 2004 (Rytų ir Pietryčių Lietuvos gyventojų apklausa 2002, Vilniaus miesto gyventojų kalbų vartosenos įpročiai 2004), the pattern of language choice in plurilingual and bilingual families is changing significantly.The official state language Lithuanian is more frequently used in talking to younger family members (children, grandchildren) and significant code-switching is occurring.Language behavior towards older family members remains more stable.With the aim of verifying whether these tendencies persist, we included an additional question in the questionnaire on language choice when communicating with grandparents.
Moreover, the Lithuanian questionnaire was complemented with a question on language and media use.The aim was to examine the prevailing tendencies of primary school pupils regarding their preferences for TV shows and the language these are in, depending on the dominant home language.The part concerning language teaching and language learning in the Lithuanian questionnaire was also modified in order to analyze the issue of globalisation and maintenance of Lithuanian.The increasing influence of English in Lithuania so far has been observed merely as a phenomenon (Vaicekauskienė 2009, Vaicekauskienė 2010).Therefore, one more question was added on preschool foreign language learning.
A school in Nemenčinė was prepared to participate in the pilot study.Nemenčinė is a multi-ethnic and multicultural town located some 25 kms to the north-east of Vilnius.The ethnic composition of its population reflects a "typical" multilingual town in Lithuania.
There are 2 gymnasiums in this town, a Lithuanian and a Polish one.Many Polish and Russian families tend to send their offspring to a Lithuanian school, but at home they usually speak Polish or Russian.In October 2007, a pilot survey amongst 33 grade-two pupils was conducted at the Gedimino Gymnasium (Lithuanian) of Nemenčinė.The pilot survey allowed us to improve the employed methodology, and the formulation of questions and multiple-choice answers was adjusted to the format presented in Section 2.

The main study
A large team of project researchers and research assistants from three Lithuanian universities was put together in all three cities in order to conduct the main study in 2008.Special seminars were organized and guidelines were given to all research assistants.Departments of Education in every municipality were contacted by the project researchers, meetings with principals of schools were organized and each school was invited to participate in the project.In schools, the project coordinators and research assistants explained the aims and the rationale of the project, handed out and collected the questionnaires and carried out the administration process.Parental consent forms were provided asking the parents' (or "carers") consent for their child to take part in the survey.The forms were given to the children to take home with them and to be signed by their parents, and were later processed by the teachers.The project was generally greeted with enthusiasm by schools, teachers, parents and pupils alike.In each school, the questionnaire was administered with the support of research assistants during classtime.Completed questionnaires were personally collected by the research assistants and subsequently handed over to the project coordinators.
Data processing was conducted at Tilburg University in the Netherlands.Given the anticipated future size of the database in the main study, an automatic processing technique based on specially developed software and available hardware was developed and utilized (Extra, Yağmur 2004, 116-118).Because the answers to some of the items in the questionnaire were handwritten by the pupils, additional verification of these items had to be done using character recognition software.After scanning and verification was completed, the database was analyzed using the SPSS program.The schools referred to are different.Some of them are primary schools, some are secondary schools with a primary school division incorporated in them.As Table 1 shows, there is an almost complete1 coverage of schools in the sample.

City
There were slightly more boys than girls in the sample.The aim of the survey was to administer the questionnaire to pupils from grade 2 (age 8) to grade 4 (age 10).There were younger and older pupils in these grades but with a markedly lower frequency than those in the age range of 8-10 years old.The total number of pupils in the sample is 23,686, almost half of them from Vilnius.
The total number of reported home languages is 37.Only 23 home languages were reported more than three times, and out of these 23, only 9 languages were reported at least 90 times.Although Poles are the largest minority group in Lithuania, Polish is the third, not the second most often reported home language in the survey, apart from English.The reason for this lies in the Russification policy in all former Soviet republics.Every language was under the influence and pressure of Russification and in many domains Russian was the dominant language.Many non-Russians (e.g., Poles, Belorusians, Ukrainians) were assimilated by being completely submerged in the Russian culture and language.Not only Russians but also many non-Russians used Russian at home and declared Russian to be their native language (Hogan-Brun et al. 2009, 34-47).As a result, Russian is the second most often reported home language in the survey.
How can the popularity of English at home be explained?After restoration of the independence of Lithuania in 1990, English has become very popular.The changed geopolitical orientation and the accession to EU and NATO in 2004 have influenced the language constellation in Lithuania.English now is most popular as foreign language at all levels of education, as language of international communication, in the media and in entertainment.It is also the language of prestige among urban youngsters (Vaicekauskienė 2010) and has in a sense "invaded" the home language survey for this reason.

Crosslinguistic perspectives on language profiles and language vitality
Here, we present language profiles in crosslinguistic perspectives for the top-9 of reported home languages in the age range of 8-10 years old.The concept of "language group" in the tables below is based on the pupils' answers to the question which language(s) is/are used in the home.On the basis of their answer patterns, pupils may belong to more than one language group.For each language group, four language dimensions will be presented in terms of reported language proficiency, language choice, language dominance, and language preference.In the analyses, the outcomes for each of these dimensions are compared in terms of proportional scores, i.e., the mean proportion of pupils per language group that indicated a positive response to the questions under consideration.Given the possible non-responses to any of these questions, all tables are presented and interpreted in proportional values.
From the analyses on the basis of the four language dimensions mentioned above, we eventually construct a cumulative Language Vitality Index (LVI) for each of the 9 language groups under consideration.The LVI is based on the mean value of scores obtained for the four language domains.This LVI is by definition an arbitrary index, in the sense that the chosen dimensions with the chosen operationalisations are weighted equally: -Language proficiency: the extent to which the language under consideration is understood by the pupils; -Language choice: the extent to which this language is commonly spoken at home with the mother; -Language dominance: the extent to which this language is spoken best; -Language preference: the extent to which this language is preferably spoken.

Language proficiency
In Table 3, we present a crosslinguistic and pseudolongitudinal overview of the first language dimension, i.e., the extent to which the languages under consideration are understood by the pupils.The top position of Lithuanian as dominant language in society does not come as a surprise.By comparison, the scores obtained for Polish and English are rather high, and those for French, Latvian and German are rather low.Across age groups, there is an increase in reported understanding of English, Belarusian, Ukrainian and French, and a decrease in reported understanding of Russian.
In Table 4, we present a comparative perspective on reported average language skills per language group in the age range of 8-10 years old.As expected, a decreasing level of reported skills emerges for most language groups, apart from Lithuanian and French, as we go from left to right, from understanding, speaking, reading to writing.Relatively high literacy scores (reading and writing) are reported for Lithuanian, English, and Polish.Much lower literacy scores are reported for all other language groups.The reported language skills in Lithuanian are very high and similar across skills.

Language choice
In Table 5 we present a crosslinguistic and pseudolongitudinal overview of the second language dimension, i.e., the extent to which the languages under consideration are commonly spoken with the mother.The patterns of language choice presented in Table 5 are much more differentiated than the patterns of language understanding presented in Table 4. Apart from Lithuanian, only Polish and Russian are chosen rather frequently in interaction with the mother.Across age groups, there is a decrease in the use of Polish and Russian, and an even stronger decrease in the use of Ukrainian as common language of interaction with the mother.

Age group
Table 6 offers a comparative mirror-like overview of the reported choice of Lithuanian as common language of interaction with different interlocutors at home.
First of all, Table 6 shows a strong choice in favor of Lithuanian in interaction with most interlocutors across most language groups.Lithuanian is chosen least in the Polish, Ukrainian and Belarusian language groups.Across interlocutor types, the choice in favor of Lithuanian occurs most strongly in all language groups in interaction with best friends and is least apparent in interaction with older and younger siblings.

Language group
Total pupils

Grandparents
Mother Table 6.Choice of Lithuanian as common language of interaction with different interlocutors at home (in %)

Language dominance
In Table 7, we present a crosslinguistic and pseudolongitudinal overview of the third language dimension, i.e., the extent to which the languages under consideration are spoken better than Lithuanian or as good as Lithuanian.The top-4 of languages in Table 7 shows the same hierarchical ranking as the top-4 in Tables 4-6.Lithuanian clearly emerges as the dominant language for all language groups and even for all age groups.

Language preference
In Table 8 The patterns of language preference in Table 8 are more differentiated than the patterns of language dominance in Table 7.The high status effect for English is very apparent in the language settling itself firmly in the top-4 of languages.The ranking of dominance vs. preference in the two tables for Polish, Russian, and Ukrainian remains the same.Again, Lithuanian clearly emerges as the preferred language of all language groups and even all age groups.Across age groups, there is an increasing preference for Lithuanian (and French), and a decreasing preference for Russian.
In Table 9 we compare the reported patterns for language preference and language dominance.Selected are those pupils who report different languages for the questions on language preference and language dominance.
The total number and proportion of mismatches in Table 9 show that there are many pupils for whom the preferred language is not the dominant language.Most mismatches in almost all language groups result from dominance in Lithuanian and preference for another language, and not from the reverse.The least mismatches between language preference and language dominance emerge for Ukrainian and Belarusian, the most mismatches emerge for English.

Language vitality
As was mentioned at the beginning of Section 5, this section would be concluded with the construction of a cumulative Language Vitality Index (LVI) for all 9 language groups on the basis of the four analyzed language dimensions, i.e., language proficiency (understanding), language choice (in interaction with the mother), language dominance, and language preference.The LVI (the rightmost column in Table 10) is based on the mean value of the scores presented for each of the four language dimensions referred to.
As was pointed out at the start of Section 5, this LVI is by definition an arbitrary index in the sense that the chosen dimensions with the chosen operationalisations are weighted equally.Table 10 gives a crosslinguistic overview of the language vitality per language group and language dimension.
A clear hierarchical ranking emerges in the language vitality of the language groups under consideration.The highest cumulative score (LVI) is obtained for Lithuanian, the lowest for German.Across age groups, a decreasing language vitality emerges most clearly for Polish and Russian.
Finally, Table 12 gives a crosslinguistic overview of the language vitality per language group and per generation.The three generations were operationalized as follows: -G1: pupil + father + mother born abroad; -G2: pupil born in country of residence, father and/or mother born abroad; -G3: pupil + father + mother born in the country of residence.
First of all, Table 12 clearly shows that very few pupils are represented in the first generation and that most of them are found in the third generation.Only the Belarusian and Ukrainian language groups are more strongly represented in the second generation than in the third one.
Secondly, it can be predicted that directionality will emerge in the obtained patterns of language shift in the sense that dominated languages will shift to the dominant language in society (Lithuanian), not the other way around.This prediction is confirmed by the data.A clear intergenerational increase of language vitality emerges for Lithuanian.A clear decrease of language vitality in the third generation emerges for Russian, German, Belarusian, Ukrainian and Latvian.However, the strongest decrease emerges for Russian and Latvian.Apart from Lithuanian, the highest vitality in third generation groups emerges for Polish.

Conclusions
The findings of this project have generated a wealth of hitherto hidden evidence on the distribution and vitality of a range of languages used at home and at school by primary school pupils in the three largest cities of Lithuania, i.e., Vilnius, Kaunas and Klaipėda.The total number of participating pupils was 23,686, the distribution being 45% in Vilnius, 39% in Kaunas and 16% in Klaipėda.Most of the pupils were 8-10 years old.Most pupils and parents were born in Lithuania, followed at a considerable distance by Russia, Poland and Belarus.With respect to the latter three countries, parents were more commonly born abroad than pupils.There was a wide range of other birth countries of pupils and parents.In most cases, however, reference was made to former Soviet republics or former socialist countries in Eastern Europe.These reported demographic patterns strongly diverge from the demographic patterns known from Western European countries (Extra, Yağmur 2004).Ukrainian, Russian, French, Latvian and German.The focus was on the four dimensions of reported language proficiency, language choice, language dominance, and language preference.From the analyses on the basis of these four language dimensions, we eventually constructed a cumulative Language Vitality Index for each of the 9 languages under consideration.
Apart from reported home language profiles, this study also focused on reported school language profiles.Lithuanian and English co-occur with similar and highest frequencies as reported languages of schooling, leaving all other languages far behind.The decreasing ranking of Russian, Polish and German as languages of schooling does not come as a surprise.Apart from these five languages, eleven other languages are mentioned more than once as languages learnt at school.Both the reported spectrum and the reported frequencies of languages that pupils would like to learn at school come as a surprise.
The same holds for the reported variation of languages in which TV programmes are watched.
The findings show that making use of more than one language is a way of life for an increasing number of pupils.Mainstream and non-mainstream languages should not be perceived and interpreted in terms of competition.Rather, the data show that these languages are used as alternatives, depending on such factors as type of context and interlocutor.The data also reveal that the use of other languages at home does not occur at the cost of competence in the mainstream language.Many children who addressed their parents in another language reported being dominant in the mainstream language.
The findings on multilingualism at home and those on language needs and language instruction at school reported by the pupils in this study should be taken into account by both national and local educational authorities in any type of language policy.A full-size and detailed report on the presented outcomes of this project is under preparation.

Figure 1 .
Figure 1. Outline of the Multilingual Cities Project

Table 1 .
Number of schools per city and schools and pupils in the sample

Table 2 .
Table 2 gives an overview of the latter.
Top-9 of repor, Hogan-Brun et al. 2009)ian is the official state language and the home language most often reported.It is not only used in native Lithuanian families but also by other ethnic groups, in particular since the rise of Lithuanian as a national and European language of prestige.There has been a marked rise in the level of Russian-Lithuanian bilingualism amongst the Russian community, prompted by an increasing desire to integrate in the overall population (Hogan-Brun,Ramonienė 2005, Hogan-Brun et al. 2009).Polish is the native language of the Polish community in Lithuania and in the Soviet period was used more commonly in rural areas than in cities.Although not all Poles in Lithuania used to speak Polish in Soviet times, Polish is now gaining ground again in Polish families and it is used for communication with different interlocutors at home.

Table 4 .
Oral and written skills per language group (in %)

Table 5 .
Language choice in interaction with the mother, per language group and per age group (in %)

Table 7 .
Language dominance per language group and per age group (in %)

Table 8 .
, we present a crosslinguistic and pseudolongitudinal overview of the fourth language dimension, i.e., the extent to which the languages under consideration are what pupils prefer to speak.Language preference per language group and per age group (in %)

Table 9 .
Language dominance vs. preference of pupils for whom dominance is different from preference

Table 10 .
Language vitality per language group and per language dimension (in %, LVI in cumulative %) Table 11 gives a crosslinguistic and pseudolongitudinal overview of the language vitality per language group and per age group.

Table 12 .
Language vitality per language group and generation (in %)