A CORPUS-BASED ANALYSIS ON MULTILINGUAL CHILDREN’S NARRATIVES IN INDONESIAN CONTEXTS

Children’s narratives have served as an effective analytical tool for language teachers and researchers to dvelve into a wide variety of elements in language acquisition and development. Narrative production has been investigated along two dimensions of analysis: story memory and story quality, with the first focusing on children’s ability to reproduce units of information within the story, while the second emphasizing their understanding of story structures. Our current paper focused on the second, where we looked at two specific categories; orientations (person, time, place, connectors, etc) and evaluation (personal judgment). Our data were based, in part, on a limited corpus of multilingual written production elicited from 261 third graders in six primary schools in East Java, indonesia. This dataset contained the production of personal narratives, where our participants wrote their personally experienced event of school holiday, and fictional narratives, where they rewrote a short video story of a perfect father’s day. To generate the target expressions from our relatively large datasets, we utilized a corpus tool of Antconc. Our analysis has echoed several findings. In terms of ‘person’ orientation, our participants were consistent in orienting their personal narratives with ‘i’ and their fictional narratives with ‘dad/father’. In the case of ‘place’ orientation, the word ‘home’ appeared in the first place; it shows that ‘home’ earns a very special place in children’s point of view. The word ‘after’ and ‘then’ as connectors were largely found in the corpus, expressing a lack of lexical richness in their writings. With respect to personal evaluation toward the story, limited range of lexical diversities is also indicated. Together, the findings provide a significant contribution, especially, for foreign language pedagogy.


INTRODUCTION
Narrative production of children in multilingual settings embodies a rich and authentic resource for teachers as well as researchers in the area of language acquisition and development, that is naturally complex. Assessing the development of a language in multilingual learners is therefore a daunting task with regard to individual variations and a wide variety of factors. However, narratives can be an effective tool for it (Rodina, 2017). Beyond this practicality, narratives provide opportunities to better understand children's point of views about the worlds as reflected in their language (Ukrainetz et al., 2005). Our project, therefore, seeks to illuminate a critical importance of narrative production in language acquisition in formal settings.
Extensive investigations on the development of narrative as a skill have broadly addressed issues on how it relates to age factors. Children, as Drijbooms, Groen, and Verhoeven (2016) suggested, generally start to produce oral narratives from an early age and to compose written narrative by the time they attend elementary school. Other scholars pointed out that children are able to tell organized stories, including beginnings, settings, and outcomes by the age of five and to complete it with attempts and resolutions by seven (Morrow, 1985), to improve stories with internal state responses between six and 11 (Stein & Glenn, 1979), and to include inner states and affective responses at nine (Berman & Slobin, 1994). Narrative writing generally falls into two categories: personal and fictional, which is commonly occurred in either spontaneous or elicited speech settings. The analysis of these two sets is usually geared toward the dimension of microstructure and macrostructure. The first deals with the linguistic elements used in the narratives whereas the second pertains to a higher order organization of a text (Maviş, Tunçer, & Gagarina, 2016). In examining the microstructures, researchers generally look at vocabulary, grammar, and other linguistic aspects within the framework of morphology, syntax, and semantic analysis (Lucero, 2018;Westerveld & Gillon, 2010). The analysis is typically conducted on the basis of mean length of utterance, number of different words and number of total words (Ebert & Mikolajczyk, 2016;John Heilmann, Miller, Nockerts, & Dunaway, 2010). In the extent of investigating microstructure elements in bilingual production, phenomena such as code switching and code interference have attracted more attention (Altman, Armon-Lotem, Fichman, & Walters, 2016).
Macrostructure, on the other hand, sets to map text organization that includes episodic structure and story grammar components (J. . Specifically, macrostructure elicits GOA; the abbreviated form of Goals, Attempts, and Outcomes,in which the goals define what the characters aim to accomplish in the story, attempts describe the characters' efforts to achieve the goals, and outcomes explain whether or not the characters reached the goal (Maviş et al., 2016). In measuring GAO, Multilingual Assessment Instrument for Narratives (MAIN) is commonly utilized; it comprises elements of story structure, story complexity, and internal state terms (Gagarina et al., 2012). As far as narrative analysis is concerned, Trionfi and Reese (2009) offer different categories of narrative quality that include descriptors (adjectives and adverbs), dialogue, character names, temporal-locativecausal, and verbatim recall from text.
As abovementioned, previous works have heavily concerned on how narrative ability develops in children by underlining age factors. Maviş et al. (2016), for example, investigated the effects of age on Turkish narrative skills in 49 Turkish-German bilingual children aged 2;11-7;11 in two different experimental tasks; 'tell-after model' and 'tellno model' in the first study and 'tell-no model' and 'retell' in the second. The results pointed out significant age effects on story complexity and comprehension, but not story structure and internal state terms. These researchers had previously conducted similar study on Turkish monolingual and Kurdish-Turkish bilingual children proving that both monolingual and bilingual children could answer some comprehension questions correctly by about the age of four, were able to use internal state terms appeared by six, and that these macro-structural components and comprehension improved with age.
This current study aims to participate in this global discourse of narrative analysis. However, rather than conducting microstructure analysis as has been extensive in previous major works, we observe the macrostructure elements by looking closely at how children structure their stories under the sub-elements of orientation and evaluation. In the element of orientation, we investigate how person, time, place, and connectors of a story interplay, whereas in the element of evaluation, we analyze the use of evaluative devices employed in our participants' stories. By putting this project in place, we intend to articulate a potential significance of narrative analysis for pedagogical purposes, especially in measuring multilingual children's ability in story-structuring proficiency and linguistic skills (Lucero, 2018).

METHOD
Our data were based on a limited corpus of multilingual written production of 261 third graders in six primary schools in East Java, Indonesia. The field sites were (1) SD Laboratorium UM Malang, (2) SD Laboratorium UM Blitar, (3) MI Al-Akbar Surabaya, (4) SD Muhammadiyah Manyar Gresik (5) SD Muhammadiyah Ikrom Wage Sidoarjo, and (6) SD Laboratorium UNESA Surabaya. This narrative corpus contained 154.496-word tokens, which is significant in enriching CBLING (Corpus of Bilingual Learners' Languages); the corpus we initially built in 2016 under the support of Universitas Negeri Malang (UM). The whole body of CBLING is basically designed to collect both written and spoken production in multilingual children's repertoire; Indonesian, Javanese, and English. However, for the purpose of current paper, we only presented and analyzed data from the English datasets. The data elicitation was carried out in an elicited setting, where our participants were prompted to write personal and fictional narratives in two different time frames. In the first attempt, they wrote about their personal experience of school holiday, while in the second they rewrote a short video story of a perfect father's day. The two types of narrative were chosen to represent two different elicitation environments; in the personal narrative, participants were in a spontaneous mode of writing; unlike in the second, the story retelling, they were situated in a more controlled task. These writing practices were all conducted in classroom under a specific time constraint, with their homeroom teacher acting as no more than only a time keeper. It means that teachers were not allowed to provide any assistance with regard to the process of writing, such as providing vocabularies, creating sentences, etc. By putting this restriction, we wanted to keep our participants' writings as authentic as possible to really measure their own language development.
To generate targeted expressions from our relatively large datasets, we utilized a corpus tool of AntConc.

FINDINGS AND DISCUSSION
Our analysis has echoed several findings. In terms of 'person' orientation, our participants were consistent in orienting their personal narratives with 'I' and their fictional narratives with 'dad/father'. The first person singular pronoun 'I' (2965 times) appeared to be the most frequent word used in personal narratives, followed by the first person plural pronoun 'we' (451 times). This frequency informs us that these eight to nine years old children have already gained knowledge on how to orient a personal story with a very strong nuance of first-hand experiences. By using 'I' (and also 'we'), they emphasized the sense of agency as the core idea of personal narrative. Second person singular/plural 'you' appeared less frequently with 56 times. Some of them were used in common expressions, such as 'thank you' and 'I want to tell you about my school holiday'. Some others were in direct quotation, such as 'This is simple, but I hope you will like this' and 'We have the surprise for you' in which the pronoun 'you' was being referred to the writer him/herself.
Results from fictional narratives, on the other hand, offered different insight on how children positioned themselves as a 'story teller'. Therefore, they tended to behave as an outsider, as the noun 'father' and 'dad' -one of the main characters in the story-were at first to appear with 1383 and 1123 occurrences, respectively. In this manner, they excluded their self-centeredness by avoiding the use of first person singular 'I'. The pronoun 'I', moreover, was only to refer to the main characters in direct quotation practices (See Example 1). In this respect, children have shown a well-developed narrative skill, particularly in choosing an appropriate stance and point of view.

Example 1
Suzi saw ballons. "Dad, may I get a balloon?" (1A18_Eng) "Can I drive it?" Said Dad (6B15_Eng) "I like the red one" He said (2A24_Eng) Father said "I like your surprise" (4A14_Eng) Time orientation is another critical component in a story making. With regard to the construction of 'time', different type of experimental tasks (story telling vs. story retelling) may determine the writers' cognitive behaviour. Story telling is arguably more difficult because the writers should generate their own story without the presence of prior text model, whereas story retelling tends to be more complex due to the involvement of auditory verbal memory in getting to see and listen the story first before rewriting it (Maviş et al., 2016). In relation to 'time', we assumed that children should have undertaken more efforts in remembering terms of 'time' appeared in the video story presented to them (if any), while they should be more 'time specific' in narrating personal experience of their own school holidays. Our corpus suggested lack of past time representation in the story retelling, such as yesterday, last time, last week, and so forth. Instead, they most frequently used 'now' in directly quoting the dialogue between Sussie and her dad (See Example 2).

Example 2
SuShi asked her Dad. "DaD now are father's day" (4A12_Eng) "OK! We gonna go home now" Dad said (1A25_Eng) "I think we have to go home now" He said (2A24_Eng) "dad we now go to hous for 2 surprise" suzie say (3A10_Eng) Storytelling reveals different results where the adverb of time 'yesterday' (35 times), 'last week' (4 times), and 'last month' (2 times) have been found quite frequently, as we may expect. These words were employed to express events happened in the past which really fit into the context of last school holiday. However, the adverbs of 'now' and 'tomorrow' were surprisingly following them with 25 and 19 times of appearance, respectively. In this way, we observe that the sense of 'here and now' was still very intense, even when children were prompted to tell past experience as illustrated in the following sample (Picture 1). Picture 1. Sample of Personal Narrative (4A07_Eng) The conception of narrating personal experience does not seem to be able to detach the writers from the feeling that the events were not actually in progress. Another possibility, I strongly assume, is that learners have not completely acquired lexical and grammatical knowledge of adverb of past time in English. Assuming this to happen, the findings of this study can be useful for language pedagogy. In other words, this piece of evidence should be helpful, especially, for language teachers to give more focus on the use of common linguistic features in telling personal narratives that include the use of past verbs and past time adverbs.
Beside 'time' orientation, expression of 'place' also holds a key feature in narratives. Our participants utilized the word 'home' (363 times) most prominently in their personal narratives that it has interestingly provoked the importance of 'home' as a place to start and end the activities in children's point of view (See Picture 2). To put it differently, 'home' is particularly close to the heart of their life, so much so that the word appeared most frequently to indicate 'place'.
Picture 2. Samples of concordance on 'home' Words of place in fictional narratives revolve around the word 'restaurant' (86 times), 'house' (53 times), 'garden' (47 times), and 'living room' (17 times) that refer to the setting of the story. On the basis of this evidence, we can see that these children have acquired certain degree of language comprehension about the story being presented to them as well as been able to retrieve importance details including places or settings.
In analyzing narrative quality, we borrow Ebert and Mikolajczyk's (2016) argument in stating that the use of cohesive devises in supporting other elements such as plot and characters is, by no means, essential in making the story comprehensible. On that, we focused to see how children make use of connectors. This is one of the cohesive devices that also earn a space in narrative writing, as to ensure every single event flows not only logically but also nicely. The connector 'after' and 'then' were robust in our dataset. The heavy appearance of these two words, however, requires our attention as it may signal lack of input and exposure to a wide range of English connectors being possibly used in narrating events. At some points, teachers can introduce other connectors, such as 'following', 'as soon as ', 'afterwards', 'later', 'next', 'after that', and so forth to enrich students' lexical knowledge.
As far as lexical knowledge is concerned, enriching students' vocabularies is a crucial aspect in language learning as it strongly affects other linguistic competence (Canga Alonso, 2015). With respect to vocabulary learning, we follow the social interactionist framework in maintaining that communicative input from caregivers, in the context of family, and significant others in the environment including adults as well as peers through social interactions is pivotal within the process (K.A., V.C., M.L., & N.J., 2013;Quinn, Wagner, Petscher, & Lopez, 2015). Furthermore, Kim (2017) provides support by underlining that one of the contributing factors for vocabulary acquisition is exposure frequency and quality being enforced in children's surrounding. In this respect, school should provide rich resources for vocabulary development, supplementing family as the primary learning resources. Students are estimated to acquire 88,700 different words between kindergarten and Grade 12, putting an average of 8-11 new words per day (Graves, 2006).
The last feature we observed is evaluation. The feature expresses external judgment of the story or the characters (Trionfi & Reese, 2009). There are 10 categories of evaluative devices consisting of direct speech, emotive terms, intellectual terms, perceptual terms, negative qualifiers, hedges, modal verbs, figurative language, evaluative comments, and intensifiers (Drijbooms et al., 2016). Of these devices, we highlighted the use of evaluative comments expressing writers' opinion about an event or a person; it is commonly in the form of adjectives. . On that, we locate evaluative words articulating children's feeling about their activities. Surprisingly, we found lack of lexical diversities where only 'happy' (179 times), 'tired' (7 times), and 'special' (6 times) were used to convey the overall impression of school holiday. It is, yet, another teacher's responsibility to familiarize students with a bunch other choices. The word 'happy', for example, can be substituted with glad, joyful, delighted, excited, fortunate, lucky, and fun. Results from story retelling suggested that our participants mostly relied on lexical varieties used in the story. That being said that, they did not do much of lexical modifications. Evaluative words such as 'perfect', 'best', and 'fun' were all originated from the video story. This particular evidence should, again, be treated as fruitful feedback for language teachers, especially in finding effective strategies to enrich learners' vocabularies.

CONCLUSION AND SUGGESTIONS
To conclude, our findings demonstrate that children under investigation have reached certain level of acquisition in narrative writing, where they have already had a clear person, time, and place orientation. They were also able to distinguish the personal narrative from the fictional ones by taking relevant forms of orientation. However, lack of lexical diversities was identified from the use of monotonous and homogeneous words in their writings. This should particularly put forward significant contribution for the quality improvement of language pedagogy.
To suggest, in the near future, further investigations could focus on lexical richness comprising lexical diversity, density, and sophistication. It should be done to better assess children's ability in the story making, and more particularly to evaluate their language acquisition and development.