1.Introduction

Words are crucial for delivering and communicating one's intent to others.In the process of language development，children acquire more than a system of grammatical forms and semantic/communicative functions（Slobin，1991）.They are also faced with a variety of interlocking challenges through which they learn，from a range of available alternative options，to adopt a particular framework for schematizing experience.For example，children need to learn vocabulary，which includes nouns，verbs，adjectives，and adverbs，and use these parts of speech appropriately for schematizing their experience.

Narrative is generally defined as the manner of talking about a series of logically and chronologically related events（Ervin-Tripp & Küntay，1997；Labov，1972）.Berman and Slobin（1994） used an elicited narrative method to collect narratives depicting the same events from speakers of five different languages—English，Hebrew，Turkish，Spanish，and German.The pioneering work of these researchers helped later researchers（e.g.，Eriksson，2001；Manhardt & Rescorla 2002；Stavans，1996） to identify dynamics within the language acquisition process that come from the linguistic system of the language being acquired.For instance，McIntire and Reilly（1996） compared how English speakers’ strategies differed from those of American Sign Language（ASL） users，and found that stories told by ASL users contained more dramatic strategies with more explicit markers.

A schema（plural：schemata） is defined as a structure in semantic memory that identifies the general or expected arrangement of a body of information（Carroll，1986）.A story schema governs our processing of a story—what we expect to happen in that story（i.e.，cognitive structures for comprehension）.Studies of narrative schemata，in fact，are not new.Bartlett's（1932） work on recall of stories from unfamiliar cultures is considered the first of such studies；in his early work Bartlett attempted to reveal that remembering is not simple reproductive processing of a story，but that it rather represents the listener's reconstructing process based on an overall impression of the story.

If children acquiring different languages end up with different conceptual frameworks or schemata，the question arises about what happens in the case of bilinguals，particularly in relation to narrative.Take the use of verbs，for instance.In the case of English，intransitive verbs and their transitive counterparts，which usually take identical forms（e.g.，“Spring passes and summer follows.”vs.“Summer followsspring.”），sometimes establish the dichotomy of inchoative and causative.In Japanese，on the other hand，a great number of intransitive-transitive pairs share identical roots but different suffixes are attached（e.g.，“tsuzuk-u” vs.“tsuzuk-e-ru”=“follow”）.Previous studies on the acquisition of verbs in Japanese have shown that Japanese-speaking children，in their early stages of language development，use more intransitive verbs than transitive verbs（Nomura & Shirai，1997；Tsujimura，2006）.This is not a universal tendency，however.Children's early verbs are predominantly transitive verbs in some other languages including English（Fukuda & Choi，2009）.Then，as a result of acquiring two languages of different verb type bias（transitivity vs.intransitivity），it is interesting how bilingual English-Japanese school age children use intransitive and transitive verbs when narrating stories in both languages.

Recent studies include much more literature on bi- and tri-lingual narratives（e.g.，Gutiérrez-Clellen，Simon-Cereijido，& Wagner，2008；Simon-Cereijido & Gutiérrez-Clellen，2009）.Studies included in Verhoeven and Strömqvist（2001），for example，cover narratives told by children in a great number of languages，such as English，Norwegian，Finnish，Swedish，German，French，Dutch，Spanish，Hebrew，Turkish，Papiamento（a Creole language evolved from an Afro-Portuguese Pidgin spoken in West Africa），Moroccan Arabic，and American Sign Language.While narrative development among bilinguals was less studied compared to that among monolinguals，the studies included in Verhoeven and Strömqvist identified that regardless of the language，and irrespective of being monolingual or bilingual，interactive activities such as storybook reading positively affect children's narrative development.

With respect to the bilingual development of school age children，however，it is not yet clear what kinds of operating principles they use to separate the two languages（Verhoeven & Strömqvist，2001）.Research examining the interdependence of the grammar and communicative functions in older bilingual children，in particular，children who are learning English as a second language，is also limited and inconclusive，because of the measures used across studies.To address this gap，the current study follows the method developed by Berman and Slobin（1994），who compared ways in which speakers of different languages depict the same events in words.The current study，however，compares ways in which bilingual children describe the same scenes in two different languages，namely，English and Japanese.Rather than comparing ways in which speakers of different languages depict the same events in words，the study explores whether English-Japanese bilingual children，when narrating stories，are aware of how to encode certain aspects of their stories in order to signal their perspectives.In what ways do such factors as language，schematization，and grammar interact with each other in language production？Specifically，the following questions serve as a guide for the discussion that follows：

（1）To understand how a story is organized and to represent their perspectives accordingly（i.e.，to schematize a story，hereafter），how do bilingual children encode agency and causality in the two languages？

（2）Does the encoding of agency reflect the schematization of the scene in similar ways in both languages，or does it show disparities between the two languages？

（3）Do unanticipated patterns of encoding（i.e.，deviation from the book author's intent） indicate signs of language deficit or attrition？