The development of narrative skills in learners of Basque as a second language

Abstract

The aim of this paper is to analyse some aspects of development of Basque as a second language (L2) in children for whom Spanish is their first language (L1) who attended immersion school in Basque in a Spanish-speaking sociolinguistic context. Data consist of oral story retellings produced in a classroom setting where the same children participated at ages 5, 8 and 11. Another group of children for whom Basque was the L1 and who lived in a strong Basque-speaking environment also took part in the study. Two aspects are analysed in the stories: lexical difficulties and the production of text organisers. According to the results, the Basque L2 children seemed to have acquired a linguistic competence quite comparable to L1 capacity: the lexical aspects studied show a clear development in L2, since lexical gaps were frequent at age 5, diminished at age 8 and were practically non-existent at age 11. The production of text organisers also shows a clear developmental pattern and, with increasing age, the children produced a higher variety of text organisers, providing precise temporal links to different segments of the story. It is concluded that Basque immersion seems to foster the development of Basque L2 in contexts where the use of Basque is quite reduced.

• Basque L2 development
• immersion
• storytelling
• lexical difficulties
• text organisers

Introduction

This article1 focuses on the development of Basque L2 competence through immersion educational programmes, in a context of language revitalisation of an endangered minority language. Basque did not have official status in the Basque Country until a few decades ago, and it shares its space with the majority languages Spanish and French. It was not until recently that the steady decline of native speakers was interrupted.

Our data were extracted from a corpus of oral story retelling in a classroom setting, where a group of Basque L2 children took part at ages 5, 8 and 11. They all acquired Basque in an early and total immersion programme in a Spanish-speaking environment, starting at ages 2 to 3 years. Their narrations are compared with those produced by children who spoke Basque as their first language (L1) at the same ages, using the same story. These children were educated in Basque in a Basque-speaking environment.

It has been shown that immersion programmes permit the development of an L2 and also the maintenance and development of the L1 (Cummins 1979, 2000; Genesee 2004; Tedick, Christian and Fortune 2011). Educational programmes in Italian/Romansh for majority-language German-speaking children in Switzerland show positive effects on both L1 and L2 development (Serra 2007). Immersion has also proved to be particularly efficient for the acquisition of minority languages by speakers of majority languages (Idiazabal et al. 2008). Given that immersion education implies schooling children in a second (or third, etc.) language, it seems important to us to point out that, in contrast to linguistic submersion (where children also attend school in a second language), immersion does not aim at replacing the L1 with the L2 but at developing both in a balanced way. If adequate support is given to a minority language, immersion programmes for majority language children can be successful, as has been shown both in Canada and in the United States (Genesee and Lindholm-Leary 2013; Tedick, Christian and Fortune 2011).

Nevertheless, it has been pointed out that although receptive skills can reach high levels through immersion programmes, productive skills may contain non-target-like forms that persist (Lyster 1987; Swain 2005). Likewise, the fear of language attrition in the L1 if the L2 becomes dominant (Schmid and Köpke 2007) and the question of whether L2 immersion is adequate for all kinds of students (see Genesee 2007, for a review) are also doubts that have arisen.

Therefore we believe that detailed analyses of specific aspects of L2 development may still provide interesting information on bilingual competence in immersion education, bearing in mind that in some aspects L2 development may not be identical to native L1 development (Ezeizabarrena, Manterola and Beloki 2009).

For the present purpose, our analysis focuses on lexical difficulties in the L2 and the production of text organizers that mark story phases, namely the arch-connectors (‘and’/‘and then’) and temporal text organizers.

Previous research about lexical difficulties shows that the lack of lexical items or lexical gaps in L2 causes children to interrupt themselves and ask questions like: “What does X mean?” or “How do you say Y?” (Gajo 2009). Bilingual children may also produce mixed utterances, incorporating items from their L1 into their L2. These are usually free morphemes like nouns or verbs that are not commonly used in the L2 in question (De Houwer 2009). However, speakers in bilingual communities may also use ‘bilingual speech’, integrating items from the other language when addressing bilinguals they know will understand them (De Pietro 1988; Grosjean 1982). This is a feature that clearly distinguishes them from monolingual speakers.

The production of text organisers is included among the linguistic–discursive skills that are part of the language capacities needed for text production (Dolz and Schneuwly 1998). Previous studies have attested a massive use of the arch-connectors ‘and/and then’ at early ages, independent of L1 or L2 (Akinci and Jisa 2001; Berman 2001; Lambert 2003; Vion and Colas 2004). With increasing age, children produce more temporal text organizers, providing their stories with temporal nuances that mark the transitions between the different parts of the story (Akinci 2002; De Weck 2005).

Sociolinguistic context and educational research

Basque is spoken alongside Spanish and French in the Basque Country, situated in the north-west of Spain and south-west of France. It has been a minority and minorized language for centuries, without official status, without a standardized written language and without presence in schools. A slow revitalisation of Basque has taken place during approximately the last 40 years, mainly thanks to its incorporation in the educational system. There are, however, important administrative and legal differences. On the Spanish side of the Pyrenees, Basque speakers belong to two different autonomous communities: the Basque Autonomous Community and Navarre. On the French side, the three provinces where Basque is spoken do not form a community, but belong to the Département des Pyrénées Atlantiques. In this article, we will not refer to the linguistic situation in the French Basque Country.

Although great effort has been made during the last decades, contributing to an increasing social presence of Basque, it should be remembered that Basque shares its space with two majority languages, Spanish and French, in the area where it is spoken on either side of the Pyrenees. It should also be born in mind that although all adult speakers of Basque are Basque–Spanish or Basque–French bilinguals and fluent in both, inversely it cannot be assumed that all adult speakers of the majority languages also are proficient in Basque. This leads to a sort of asymmetric social bilingualism where Spanish or French are the dominant languages in most areas (Barnes and García 2011).

This is a common situation in minority language acquisition, where children also are exposed to the majority language from an early age, as reflected by O'Toole (2013) with reference to the Irish-speaking areas in Ireland, the Gaeltacht.

In the Basque Autonomous Community, Basque is one of the official languages according to the Statute of Autonomy of 1979 and has been compulsory in school since 1982, when the so-called linguistic models were introduced: the A model with Spanish as the language of instruction and Basque as a subject three to four sessions a week, the B model where both Spanish and Basque are used as the language of instruction, and the D model where Basque is the language of instruction and Spanish is taught as a subject three to four sessions a week.2 It is important to remember that it is up to parents to choose one of these models for their children.

Today, an estimated 50% of the 1,873,000 inhabitants of the Basque Autonomous Community have knowledge of Basque, even though only about 30% are active speakers (Basque Government 2013). However, although Basque is the L1 for an average of only 22.9% of the population, the overwhelming majority of parents choose the D or B model for their children in primary education, as reflected in Figure 1. It is important to note that the D model has increased since its establishment almost 30 years ago.

Figure 1. Choice of linguistic models in primary education in the Basque Autonomous Community. Source: Basque Government 2014

In Navarre, Basque has had official status since the 1982 autonomic agreement. Navarre is divided into three linguistic zones according to the 1986 law to regulate the use of Basque. In the northern part, considered Basque-speaking, Basque has official status alongside Spanish and it is compulsory in school. Parents may choose one of the three linguistic models: A, B or D. The central part, surrounding the capital of Navarre, Pamplona, is denominated the ‘mixed zone’. In this area, Basque is partially official and its presence in education is tolerated, but not particularly encouraged. And finally in the southern zone, which is considered Spanish-speaking, Basque has no official status and it is only incorporated into the educational system in some private schools run on a cooperative basis.

Nowadays, an estimated 80% of the population of Navarre has no knowledge of Basque (Basque Government 2013), and about 45% of the children in primary education are schooled in Basque or are taught Basque as a subject.

As can be observed in Figure 2, the G model (where Spanish is the only language of instruction and Basque is not present at all) has decreased during the last decade. On the contrary, there has been an increase in the number of children who are schooled in the D model. Furthermore, around 8% of children have taken part in partial immersion programmes in English since their introduction in 2007/2008.

Figure 2. Choice of linguistic models in primary education in Navarre. Source: Government of Navarre 2013

The good results of education in Basque have been confirmed by research on the linguistic and academic outcomes of the Basque educational system. According to Cenoz (2009), evaluations reflect similar or even higher results than in majority language monolingual educational contexts. These achievements are important in order to explain the rapid and continuous increase in the choice of the D model.

Population and data collection

Basque L2 data collection for the present research was initiated during the school year 2003/2004 when the children were 5–6 years old and continued at 3-year intervals at ages 8 and 11. Age 5 corresponds to the last year of preschool education, which is mainly dedicated to the development of oral discourse skills in Basque. Spanish is introduced as a school subject at the primary school level.

The data were collected in a school in Estella/Lizarra, a small town situated in the mixed zone of Navarre, about 40 km southwest of Pamplona. Knowledge of Basque is limited in this area, and its use even more so. At most, an estimated 8% of the population of the mixed zone may use Basque in daily life (Basque Government 2013), and in the town of Lizarra estimates are around 6% (Dufur Otheguy 2012). That is why contact with Basque was restricted to school for our subjects. The 37 children who took part in our research were selected from Spanish-speaking homes, where parents made an active choice to have their children schooled in Basque. Starting at preschool age (two or three), all of them attended an early and total immersion programme in Basque at a school with a long tradition in these programmes, which are designed for the introduction of Basque as an L2 in early childhood. It should be mentioned that in these programmes Basque is systematically used as the only language by teachers in the classroom (except when Spanish and English are taught as subjects). The school in question recently celebrated its 40th anniversary (Soziolinguistika Klusterra 2010).

Basque L1 data used as reference were collected in a school in the small town of Zumaia, situated on the coast of the province of Gipuzkoa, near San Sebastian, in the Basque Autonomous Community. This is traditionally a strong Basque-speaking area, where transmission of the language from parents to children has never been interrupted. The 24 subjects who took part in our research came from Basque-speaking families and attended school in the D model. They were of course also exposed to the majority language, Spanish, in the sociolinguistic environment and as a school subject from primary education on and eventually became Basque–Spanish bilinguals.

Both parents and schools provided all the necessary permissions to carry out the activity and to use the data for research.

The activity

The activity chosen for our purpose, oral storytelling, is a text genre frequently used in psycholinguistic research with the aim of studying the development of discourse skills in children both in L1 and L2 (Berman and Slobin 1994; Hickmann 2003; Verhoeven and Strömqvist 2001). However, it is more than that: it is part of an age-old tradition, forging a common cultural heritage of people throughout the world, transmitting knowledge and experiences. It has also proved to be a powerful tool for didactic purposes, since “stories often hold a strange magical quality that can interest learners in a way few other materials or methods have” (Ellis and Brewster 2014, 2). It may well be used with children at different levels of language skills and learning needs and thus is adaptable to both L1 and L2 learners. Furthermore, it is a natural way of introducing children to continuous and coherent spoken discourse (Cameron 2001), helping them develop their language skills from dialogue interaction to monologue or self-sustained performance.

This kind of task is habitual in preschool curricula (Boiron 2001; Grossmann 1996). It is often reinforced in educational immersion programmes and it was familiar to our subjects. Before initiating the task, the children were told that their performances would be videotaped and shown to smaller children who did not know the story. The same story was used at ages 5, 8 and 11.

The story chosen is an adaptation of a traditional Basque fairy tale. It has a symmetric structure and the vocabulary items have been checked so as to be accessible for children from age 5. In the story “Mattin Zaku,” a little boy and his mother were very poor. Mattin Zaku told his mother he would go to the king's castle and ask the king for money. On the way, he met three helpers, but had to overcome three dangers presented by the king, which of course he did with the aid of the helpers. So he obtained the money, returned home and then lived happily ever after with his mother. As can be appreciated, the story corresponds to the classical structure of children's tales with a happy ending. The story's timeline spans several days, which requires the use of a variety of temporal text organisers, one of the points our research focuses on.

The method

In each session, an adult (one of the researchers) narrated the story to a group of five children, selected at random from the participants. The storytelling was given good visual support by a block of 12 big wordless pictures representing the most important episodes, helping the children both to understand and subsequently to reproduce the story. The narrator showed the pictures one after the other as the story proceeded.

After listening, three of the children left the classroom and one child stayed on the narrator's chair, in order to retell the story to the second one. The child-narrator was allowed to use the pictures, but was told not to show them to the listener in order to avoid shared picture descriptions. Then he/she left the classroom and the child who had listened told the story to a third child, and so on. When these children had completed the task, the same procedure was repeated with the next group of five. It was considered that a group of five was an adequate size in order to motivate and actively involve the children in the activity, and that a group larger than five might make it difficult for them to remember and repeat the task. No substantial differences due to the order of the narrator were noted. The sessions were video recorded and transcribed for analysis.

Our method has similarities with the widely used research tool ‘Frog, where are you?’ Eliciting the same story makes it possible to control the content of the texts produced by the children and thus facilitates comparison across different groups of subjects or different languages. The main difference resides in our supplying an adult model, which is not the case with the Frog method (Berman and Slobin 1994).

In our opinion, storytelling by adults to children is not a one-way transmission of content, but also influences the children's own retelling competence (Bronckart and Dolz 1999; Dolz and Schneuwly 1998). In the first place, listening to stories is not a passive activity, but a sort of ‘shared social experience’ (Ellis and Brewster 2014, 7). If it is carried out in a way that motivates the children to participate, there is usually a response of laughter, excitement, anticipation, and so on. In fact, several examples of such spontaneous reactions from the children can be found in our storytelling sessions (1, 2):

(1) Adult: otsoa ere zakuan sartu zuen Mattin Zakuk

    Mattin Zaku also put the wolf in the sack

  Child: berriz?

    Again?

  Adult: bai

    Yes

  Child: ibaia/orain da ibaiarena

    The river/now it's the river

(2) Adult: zakua hemen baino haundiago zen e?/oso haundia zen/eta bidean

    The sack was bigger than here/it was very big/and on the way

  Child 1: orduan horrela igual?/etzanda

    Then perhaps like this/lying down

  Child 2: hau hau sartzen zuen

    He put this (one) in

  Child 1: igual sartuko zen horrela

    Perhaps he would put it like this

These examples reflect the way the children get involved in the story, anticipating what is going to happen (next, Mattin Zaku will put the river into his sack) and contributing with their own imagination (the possible ways of putting the big fox into a sack that seems to them to be too small).

It should be borne in mind that the aim of our activity was not only to understand the story but also to retell it. The narrator not only transmitted the plot, but also reinforced it by gestures, intonation, pointing at the pictures. This may have had positive effects on children's understanding of the communicative situation (where and when it happens, who the listeners are), the status of the speaker (a narrator, sitting before an audience) or the goal of the activity (to tell a story to children who do not know it). This was evidently not the same communicative context as for example a story-reading activity with parents at home.

In previous studies the impact of the communicative context has been shown to be a key factor when studying narrative development, since children's narrative performance may vary according to different communicative contexts (Gonnand and Jisa 2000). For instance, Brigaudiot (1993) compares the different communicative contexts of storytelling and shows that children produce the best organised and most coherent narratives in communicative contexts where children had previously listened to the story told by an adult.

Analysis

In the first place, our analysis focuses on how the children dealt with the lexical difficulties they were faced with during the activity. The strategies used to solve these problems are classified as follows:

1. The question about the lacking lexical item is formulated in Basque but with the introduction of the equivalent Spanish word.

2. The question is formulated in Basque but the child just points at the picture and asks what it is.

3. The problem is solved by incorporating Spanish lexical items with Basque morphology.

This issue refers to the transfer of lexical items from one language to the other. In bilingual communities this is a normal phenomenon, producing expressions that would not be used in monolingual contexts of the languages in question. Whether such an item should be considered a ‘loan word’ or a ‘code-switch’ is a question that goes a bit beyond the scope of this article. O'Toole considers ‘loan word’ any English word “which has been naturalized into the phonology, morphosyntax and everyday use of Irish” (2013, 87). In this context, it is clear that if one of the languages is dominant in the environment, transfer usually occurs from that language into the weaker one. In our case, it is usually Spanish that ‘lends’ words to Basque, even in predominantly Basque-speaking communities.

The production of text organisers is the second element of analysis. ‘Text organisers’ are defined by Bronckart (1996) as the linguistic units that mark the thematic progression of a text, establish a nexus between types of discourse and mark the transition between segments on a textual and enunciative level. They may include conjunctions, adverbs, noun phrases and prepositional phrases.

Previous studies have attested to the massive production of the arch-connector at age 5, which reflects a local–discursive control implying a simple succession of actions without links between episodes or narrative phases (Akinci and Jisa 2001; Berman 2001; Lambert 2003; Vion and Colas 2004). On the contrary, the increasing use of temporal organisers at ages 8 and 11 provides stories with temporal nuances that mark the transitions between the different parts of the story (Akinci 2002; De Weck 2005). In this way, the difference between story time and narration time is established.

For the present purpose, our analysis distinguishes between the following:

1. The production of the arch-connector eta/eta gero (‘and’/‘and then’)

2. The production of temporal connectors such as gero (‘later’), orduan (‘then’), orain (‘now’), -ean/-ela (‘when’), bat-batean (‘suddenly’), egun batean (‘one day’), hurrengo egunean (‘the next day’), handik aurrera (‘from then on’) and azkenean (‘finally’)

The total number of text organisers produced by each child was counted in order to obtain the average for each category. Given that comparisons are made between the 37 L2 and 24 L1 children, these data are converted into percentages in Figure 3.

Figure 3. Percentages of arch-connectors and temporal connectors at ages 5, 8 and 11 in Basque L1 and L2.

Results

Lexical difficulties

Lexical gaps appeared in Basque L2, although the children were already fairly fluent in their storytelling activity at age 5. Lack of specific lexical items leading to interruptions were noted in 24 out of 36 narrations (one of the children did not say anything at all in Basque), as reflected in Table 1.

Table 1. Number of stories in Basque L2 containing interruptions due to lexical gaps

Age 5	Age 8	Age 11
24/36	22/37	2/37

Some examples where the equivalent of the lacking lexical item was introduced in Spanish were noted (3):

(3) Child: e: nola esaten da a- zorro?

    e: how do you say [in Basque] ‘fox’ [in Spanish] (O.H., 5, L2)3

More frequently, the children just pointed at the picture they had before them, formulating the question in Basque (4, 5):

(4) Child: nola esaten da?

    how do you say [this]? (J.E., 5, L2)

(5) Child: zer da hau?

    (pointing) what is this? (I.F., 5, L2)

At age 8, these lexical problems still appeared in 22 stories (Table 1) and, although the children mainly used the same strategy as at age 5, they now introduced the unknown word in Spanish more frequently:

(6) Child: lobo bat nola esaten da?

    how do you say a [in Basque] ‘wolf’ [in Spanish] (L.V., 8, L2)

(7) Child: nola esaten da esposa?

    how do you say ‘handcuff’?(J.V., 8, L2)

By age 11, lexical difficulties causing the children to interrupt themselves and ask for the missing item had practically disappeared: they were only noted in two stories (Table 1).

As to the incorporation of Spanish lexical items with Basque morphology, examples were found in both Basque L1 and Basque L2 stories. Such was the case of akabatu, from the Spanish word acabar, “to finish (with), to kill.” This lexical item was present in most productions at all three ages in both groups, as reflected in Examples 8 and 9:

(8) Child: zaldi guzti guztiak akabatu in zuen (G.A., 5, L1)

    he killed all the horses

(9) Child: eta akabatu zuen denak oiloak

    and he killed all the hens (A.Z., 5, L2)

In the previous examples, the Basque perfective verb suffix –tu is attached to the Spanish verb. In the following examples from Basque L2 the same verb is used with the Basque imperfective verb suffix –tzen (10) and the suffix for the future –ko (11):

(10) Child: oilo erraldoiak akabatzen dute

    the giant hens kill (him) (J.E., 5, L2)

(11) Child: zaldiak akabatuko dute

    the horses will kill (him) (A.S., 8, L2)

Specifically, in Basque L2, there are several examples of items incorporated from Spanish that are not used in Basque L1. These examples are attested at the three ages as reflected in Table 2.

Table 2. Number of stories in Basque L2 containing lexical incorporations not present in Basque L1

Age 5	Age 8	Age 11
14/36	19/37	14/37

In Example 12, the Basque suffix –tu is attached to the Spanish verb encontrar (‘to meet’):

(12) Child: bidean joan zen ibiltzen eta enkontratu zuen azeri bat

    he went walking on the road and he met a fox (U.A., 5, L2)

In Example 13, the Spanish item caballo (‘horse’) carries the Basque dative case mark -ri, and in Example 14 the Spanish verb responder (‘answer’) is marked with the Basque suffix –tzen:

(13) Child: caballori/eta eraman zion

    to the horse(s)/and they took him (J.L., 5, L2)

(14) Child: eta ez zuen inor responditzen

    and nobody answered(C.A., 8, L2)

In Example 15, two Spanish nouns are used: madera (‘wood’) and paja (‘straw’), to which the Basque uncountable determiner asko (‘a lot’) is added.

(15) Child: hartzeko madera/hartu zuten paja asko/

    to take wood/they took a lot of straw(A.G., 11, L2)

These resources of course increase the degree of fluency of the narrations in Basque L2, since the children do not stop and ask about lacking lexical items, but use the resources the ‘bilingual speech’ offers them. There are, however, also examples that show that Basque L2 students distinguished quite well which item belongs to which language. In Example 16, the Spanish word camino (‘road’) does not seem appropriate to the listener, who changes it into the Basque word bidea.

(16) Child 1: hurbil dagoela/eta/kaminoan/e

     it is near/and/on the road/e(J.L., 8, L2)

   Child 2: bidean/

     on the road(A.S., 8, L2)

In Example 17, the narrator seems quite aware that enkontratu is not the Basque word for ‘meet’, but he decides to use it in order to continue narrating.

(17) Child: azeri bat/e:/enkontratu zen/beno enkontratu

     a fox/e:/he met/well met/ (I.F., 11, L2)

Text organisers

With one or two exceptions, the stories produced at age 5 contain chains of isolated actions linked together by eta/eta gero (‘and’/‘and then’), forming a simple line of events. In fact, the arch-connectors represent more than 80% of the organizers taken into account in the productions. No distinction is marked between narration time and story time: the story lasts as long as it takes to narrate it (Example 18):

(18) Child: kaixo Mattin Zaku/nora zoaz?/eta esan zion Mattin Zaku/

   erregearen etxera/nahi al duzu nirekin etorri?/eta esan zion/e azeria/bai     bai/eta esan zion Mattin Zaku/

    hello Mattin Zaku/where are you going?/and Mattin Zaku told him/to the king's house/do you want to come with me?/and the um fox said/yes yes/and Mattin Zaku told him/ (A.L., 5, L2)

In contrast, if we compare these data with Basque L1 data at the same age, there are no statistically significant differences: the predominance of the arch-connector seems rather to be an age-related feature of development, showing very similar values in L1 and L2 (Figure 3).

At age 8, the productions showed not only a larger variety of forms, but also development in marking connections between text segments above the level of enouncement, that is to say, between episodes or phases of the story. Arch-connectors represented around 70% of text organisers analysed in Basque L2, as reflected in Figure 3. The increase in the use of temporal text organisers was mainly due to the introduction of orduan (‘then’) and -ean (‘when’), as in Example 19, where orduan (‘then’) marks the onset of a new episode or set of actions.

(19) Child: otsoa zaldi guztiak jan zituen/orduan/erregea/e/herriko plazan sua/e/enborrak biltzeko eta sua pizteko esan zuen eta Martin sua/e/notatzen zegoenean/ibaiari deitu zion/

     the wolf ate all the horses/then/the king/e/fire on the square/e/to gather wood and to light the fire he said and Mattin the fire/e/when he was feeling the fire/he called the river/(A.E., 8, L2)

It is interesting to observe that in 15 of the productions the narrator now managed to separate story time and narration time by the use of hurrengo goizean (‘the next morning’) or hurrengo egunean (‘the next day’), as in Example 20.

(20) Child: e: hurrengo egunean erregearen laguntzaileak joan ziren

     e: the next day the king's assistants went (there)(M.N., 8, L2)

In Basque L1, the percentage of arch-connectors is higher (77%) than in Basque L2, and in this case the difference proves to be statistically significant (t(57) = 4.32; p<0.001).

At age 11, arch-connectors represented a lower proportion than formerly as compared to temporal text organisers: around 61% in Basque L2 (see Figure 3). By this age, all the children distinguished background frame and foreground events by means of -ean; -ela,-ala (‘when’, ‘while’), as reflected in Example 21:

(21) Child: bidean zihoala/e:/azeri bat aurkitu zuen

     while he was walking on the road, he met a fox(J.A., 11, L2)

Almost all the children produced temporal text organisers marking links between the phases and episodes (hurrengo egunean, ‘next day’; egun batean, ‘one day’; azkenean, ‘finally’). They now made a clear difference between internal story time and the time it took them to tell it (22, 23).

(22) Child: azkenean ireki berriro eta oilloangana/e:/bidali zioten/…/eta oilloek akabatu behar zuenean azeria poltsotik atera eta oillo guztiak akabatu zituen/orduan/e:/e:/hurrengo egunean joan ziren ikustera

     finally they opened again and sent him/to the hens/…/and when the hens were going to kill him he took the fox out of his bag and he killed all the hens/then/e:/e:/the next day they went to have a look (M.N., 11, L2)

(23) Child: egun batean esan zuen Mattin Zaku/ba joango naiz e:/erregearen gaztelura/dela oso aberatsa eta eskatuko diot/eskatuko diot dirua/orduan hurrengo goizean abiatu zen/

     one day Mattin Zaku said/well I'll go e:/to the king's castle/he is very rich and I'll ask him/I'll ask him for money/then the next morning he left (A.G., 11, L2)

In Basque L1 the percentage of arch-connectors is similar (63%). Temporal organisers represent 37% and in this case the difference as compared to Basque L2 (39%) is statistically significant (t(56) = −2.14; p=0.037). However, this quantitative difference is compensated for by a greater variety of temporal organisers, which are not present in Basque L2: hortikan aurrera (‘from that moment on’); halako batean (‘at one of those moments’).

The increase in the use of temporal organisers is clearly reflected in Figure 3, and in both groups the age-related differences are statistically significant (for L1: F(2.38) = 34.28, p<0.001; for L2: F(2.70) = 32.28; p<0.001).

Discussion and conclusions

The aim of this work was to provide a detailed analysis of some aspects of the development of Basque L2 in Spanish L1 children who attended school in a Basque immersion programme. The task they were asked to carry out was an activity they were familiar with from a classroom context: to retell a story previously narrated by an adult.

Our results may be interpreted from age-related as well as from language-related points of view. The Basque L2 children participating in our study seemed to be acquiring a linguistic competence quite comparable to L1 capacity, at least as far as the discursive skills analysed here are concerned. The lexical aspects studied show a clear development in L2. Lexical gaps due to lack of vocabulary were frequent at age 5 but diminished at age 8 and were practically non-existent at age 11, denoting growing command of the L2.

When it comes to lexical difficulties, it is noteworthy that at age 5 the children mostly used the strategy of pointing at the picture of the missing item instead of mentioning the unknown word in Spanish. Why this should happen is of course difficult to know. One possible interpretation could be that due to the Spanish-dominant sociolinguistic environment, it was strongly stressed by the teachers that Basque was the language to be used at school, and this fact might have made these children reluctant to introduce words in Spanish when they were in the classroom. However, at age 8 the children seemed to have become more confident in their competence in Basque and did not avoid the use of unknown words in Spanish so consistently.

Lexical incorporations from the Spanish L1 as a resource of bilingual speech (De Pietro 1988) were frequent at age 8 and persisted at age 11. It should be noted, however, that the Spanish lexical items always carried Basque morphology. This feature of bilingual speech is also found in some cases in the productions of Basque L1 speakers. However, we found specific examples in the stories produced by Basque L2 children that were not attested in Basque L1.

Concerning the production of text organisers, a larger variety of organisers was attested with increasing age. At age 5, our analysis reflected a pattern that is common at that age: the repertoire of organizers was very limited, as pointed out for instance by Berman (2001). At the age of 8 we found that the overuse of the arch-connector tended to decrease while the production of temporal organisers gradually increased, providing much more temporal nuances to the actions, episodes and phases of the story. This production of temporal text organisers may reflect the development of the ability to distinguish between the story time and the narration time. The most remarkable change between ages 8 and 11 is that the use of temporal text organisers can be found in almost all children's stories. It is important to note that this age-related development was found in both Basque L2 and Basque L1 children.

These results seem to confirm that immersion educational programmes foster the development of the immersion language (Cummins 2000; Genese 2004). In our opinion, what makes our study particularly interesting is that the language of immersion is a minority language with quite limited presence and use out of the school context.

Our study may also provide interesting data on the issue of similarities and differences between L1 and L2 development. Ezeizabarrena, Manterola and Beloki (2009) analysed the production of the ergative case in Basque. This is a feature that is not present in Spanish and often causes difficulties to L2 speakers. Ezeizabarrena et al. studied the same corpora of stories analysed in our present work. At the age of five, differences were found between Basque L1 and L2. Whereas Basque L1 children produced the ergative case in morphosyntactic contexts where it is required, the Basque L2 children systematically omitted its production. In another study, Ezeizabarrena (2012) found that these differences between Basque L1 and Basque L2 still persisted at age 8.

Bearing these and our study's results in mind, it seems that different aspects of Basque L2 show different developmental patterns: on the one hand, lexical limitations considerably decrease from age 5 to ages 8 and 11. The ability to produce bilingual speech by incorporating Basque morphology with Spanish lexical items appears by age 5 and is also attested at ages 8 and 11. And finally, the increasing production of temporal text organisers reflects age-related skills related to establishing temporal links and nuances between the actions, episodes and phases of the story. These skills seem to follow the same patterns in Basque L1 and L2. The acquisition of these discourse features undoubtedly requires more specific teaching than the acquisition of lexical items, a fact that illustrates the efficiency of immersion teaching.

Other features of Basque may show different developmental patterns and, although these are not analysed in the present study, undoubtedly constitute a very interesting field for future research on Basque L2 development.

Additional information

Notes on contributors

Ibon Manterola

Margareta Almgren is a former teacher of foreign languages in secondary education and of Swedish at the University of the Basque Country (UPV/EHU). Since 2002 she has been collaborating with the Language Acquisition and Use research group (UPV/EHU), with a focus on language acquisition in Basque–Spanish bilinguals.Ibon Manterola holds a PhD in linguistics at UPV/EHU. He is member of the Language Acquisition and Use research group and the UNESCO Chair of World Language Heritage (UPV/EHU). His main research interests include child bilingualism, immersion education and language didactics.

Notes

1. This work has been carried out thanks to financial support granted by the Basque government (IT676- 13) and the Spanish government (FFI2012-37884-C03-01). We are grateful to Eva Lindgren and to the anonymous reviewers for their contributions to previous versions of this work.

2. Since the letter c is not included in the Basque alphabet, there is no C model.

3. In the examples, initials corresponding to the child's name and surname are given, followed by age and language status.