3.Results

As mentioned earlier，the data used for this study—transcriptions of narratives elicited by a picture storybook without words，Frog，where are you？（Mayer，1969）—make it possible to examine the range of constructions used across speakers to present particular scenes and episodes.Following Slobin（1994），the study focuses on how bilingual children narrate two particular scenes in both languages.These two scenes come from the middle of the book.Something happens in a location among trees.Something happens to the boy，and something also happens to the dog.In these events an owl，bees，and a deer are involved.

Bee-chase Scene

As mentioned above，the basis for this research is the re-telling of the so-called frog story.The dog and bees have already appeared in scenes prior to the chasing scene（Pictures 11-12），where the dog knocks down a beehive hanging from a tree and runs away from a swarm of bees.

Topic，Agency，and Verb Types

The concern of this section is relative topicality of agent and patient as the critical pragmatic dimension.Narrators have a variety of options to choose to express causality.Specifically，one of the goals of this section is to explore how English-Japanese bilingual narrators make use of verbs，in particular in passive and alternative constructions，across the two languages.

English.In the narration of the chasing scene（Pictures 11-12） in English，25（62.5%） of the 40 bilingual narrators selected the bees as topic，and 20 of them（80%） used “chase” or “follow” as a transitive verb.Among the remaining five narrators，one used “follow（after the dog）” as an intransitive verb，one used “chase（after the dog）” as an intransitive verb，one used “go（toward the boy）”，and two used “run（after the dog）”.While “chase” seems to have a strong connotation（in the sense of attacking the dog），“follow” may not sound as strong as “chase”，and “go” and “run” do not sound strong at all.All of these narrators described the scene using either progressive/durative or simple past.Note that the use of the simple past tense signals that an action is，more or less，punctual or completed；on the other hand，the use of the progressive/durative signals that an action is continuative.

Table 1 Chasing Scene（Pictures 11-12）：English

续表

续表

Seven narrators（17.5%） selected the dog as topic.Of these，all except for one used either the passive or the progressive-passive as seen in Taizo's [Boy；Grade 4.7；Age 10 years，0 months] narration “The dog was being chased by bees”.In other words，six narrators selected a Cause-View；namely，the bees play a controlling role in the dog's actions，such as “The dog was chased by the bees”.On the other hand，one narrator，Hiro [Boy；Grade 2.6；Age 9 years，2 months]，selected the dog as actor and said，“The dog was running from the bees”.In this case，bees do not necessarily seem to be serving as controller even if the narrator's intention was to describe the bees as contributing some degree of agency to the event.Thus，while intransitive syntax does not necessarily encode a non-causal event view and may still indicate a hint that the locus of control lies outside the dog，the choice of the intransitive verb “run” signals a minimal degree of agency to the listener.

When returning to the issue of bees as controller，note that among those who used the passive voice，two narrators，Kaori [Girl；Grade 2.5；Age 8 years，2 months] and Kenta [Boy；Grade 3.7；Age 9 years，7 months]，opted for the so-called get-passive，“The dog got chased by bees”.While the ordinary passive voice（i.e.，the be-passive） and the get-passive both function in similar ways，the latter implies that the patients are negatively affected.That is，the get-passive strongly implies the Cause-View whereas the be-passive seems to have diminished agency.In this case，the dog is perceived as the victim，which probably conforms to the author's intent.

Regardless of their choice of active or passive voice，however，some narrators did not encode high agency.High agency is correlated with use of the transitive verb whereas low agency is correlated with the intransitive verb，r_s=.80，p＜.0001.[Note that the Spearman rank correlation coefficient was used because it is suitable for data of an ordinal level.] This is not a complete correspondence，because one narrator's [Ayako；Girl；Grade 1.5；Age 7 years，1 month] choice of a transitive verb “follow” is here categorized to have a relatively low degree of agency.If Ayako had added an adverb “angrily” and narrated，“The bees angrily followed the dog”，categorization would have been different and the correlation would have been a perfect 1.0.On the other hand，Eimi [Girl；Grade 4.8；Age 10 years，8 months] narrated using an intransitive verb，“The bees went straight toward him”.But if the narrator had added an adverb such as “angrily” and said，“The bees angrily went straight toward him”，the correlation would have been lower because even an intransitive verb can achieve high agency with the addition of an adverb.Thus，regardless of the verb types，agency is possibly heightened by associated adverbs.Additionally，the use of an intransitive verb “run” as in “the bees run after the dog” and “the dog runs from the bees” reflects a relatively low degree of agency.

Finally，the remaining eight narrators did not narrate the chasing scene in English.They either entirely neglected to narrate the scene or paid attention to other things as seen in Eriko's [Girl；Grade 1.8；Age 7 years，2 months] narration “The owl flew out and scared him to fall right on the ground”.

Japanese.In Japanese narration of the chasing scene，26（39.5%） of the 40 narrators selected the bees（hachi） as topic.One child used “mitsubachi（honeybee）” whereas two narrators did not know how to say “bee” in Japanese and thus used “bee”，and，likewise，one used “hae”（fly）.Likewise，some narrators did not have Japanese equivalents of English verbs.One narrator [Girl；Grade 1.7；Age 6 years，6 months]，for instance，did not seem to have a Japanese equivalent of “follow” and thus simply said，“follow shi-te-i-ru（be following）”.The use of an English verb（e.g.，follow，chase）+a variation of a Japanese verb suru（which is a Japanese equivalent of an English verb “do”），which was sometimes observed [not only in the chasing scene but also in the deer scene]，indicates how bilinguals may need to stretch their limited resources for the communicative context in the other language.These variations are all included here（see Table 2）.

Overall，high agency is correlated with use of the transitive verb whereas low agency is correlated with use of the intransitive verb，r_s=.88，p＜.0001.The correlation was not perfect because “oikakekko o suru（do a chasing activity）”（which will be shown soon after） is judged to have a low degree of agency.

Table 2 Chasing Scene（Pictures 11-12）：Japanese

续表

Japanese words：Nouns：hachi（bee）；hae（fly）；inu（dog）；mitsubachi（honeybee）；oikakekko（chasing activity）；tsuku（to follow after）；Verbs：（r）areru（passive form）；hashiru（to run）；iru（to exist）；kowaku-naru（to become scary）；kuru（to come）；oikakeru（to chase）；suru（to do）；tobu（to fly）

The aforementioned 26 cases need more detailed descriptions，however.Fifteen narrators selected a transitive verb “oikakeru/ou（chase）”.By using “oikakeru /ou” high agency is encoded.On the other hand，while the remaining 11 narrators（25.5%） selected the bees as topic，high agency is not necessarily encoded.Eriko，for example，narrated “hachi ga i-mashi-ta（There existed bees）”and Leon [Boy；Grade 1.8；Age 7 years，5 months] narrated “motto takusan hachi ga ki-ta（More bees came）”.Likewise，the use of inchoative verb in Ayako's narration of “mitsubachi mo zenzen kowaku nat-chat-ta（The honeybees became very scary）” simply informs the reader of the state being described，or the narrator simply taking a Become-View with a verb of diminished agency.Thus，lexical choice reflects varying degrees of agency.Furthermore，Nicole's [Girl；Grade 2.9；Age 7 years，7 months] narration，“hachi wa inu to oikakekko o shi-te-i-mashi-ta（The bees and the dog engaged in a chasing activity）”，also does not imply the victimization of the dog.In addition to the aforementioned 26 narrators，one narrator（Kaori） narrated the actions of the bees and the dog as if those actions happened without any Cause-View；“hachi ga ki-te，inu ga hashit-te（Bees came whereas the dog ran）” does not imply the bees’ causal role.

Seven narrators selected the dog as topic.Of these，all except for one used either the passive or the progressive-passive（e.g.，“inu wa hachi ni oikake-rare-te-i-mashi-ta”（Taizo）] The dog was being chased by bees”）.[Note that，interestingly，Taizo used the progressive-passive form in both English and Japanese.] One narrator（Tadashi [Boy；Grade 2.9；Age 8 years，6 months]） selected the active voice，“inu ga kyuu ni hashitte kite（The dog suddenly came running）”，and thus the bees in his narrative do not play any causal role.

The remaining six narrators did not narrate the chasing scene at all.They either entirely ignored the scene or paid attention to other things.Like English，therefore，there is a great amount of variation in Japanese when it comes to the degree of agency，while choosing the dog as topic and the locus of effect，bees as controller，and a Cause-View.

Similarities and Differences Between English and Japanese

Recall that，setting aside the degree of agency，eight narrators did not describe the chasing scene in English，and six narrators did not describe the scene in Japanese.Melissa [Girl；Grade K.7；Age 6 years，1 month]，Eriko，and Leon，for instance，did not do so in English but they did refer to the chasing scene in Japanese.In contrast，Koharu [Girl；Grade 1.9；Age 7 years，0 months] referred to the chasing scene in English，but she omitted the narration in Japanese.The remaining five narrators did not refer to the chasing scene either in English or in Japanese（see Table 3）.

Table 3 Chasing Scene（Pictures 11-12）：English and Japanese

续表

Degree of agency.With regard to the degree of agency，recall that Eriko and Leon did not encode a high degree of agency in their Japanese narrations.Likewise，Melissa narrated “（hachi ga） ippai wanwan ni tsuite（A lot of bees stuck to the doggie）”，which does not involve a Cause-View.Thus，except for Koharu（who referred to the chasing scene in English but not in Japanese），all narrators behaved in more or less similar ways in both languages.As seen in Table 4，overall，20 narrators encoded a high degree of agency in both languages.Five narrators encoded a high degree of agency in English but encoded a low degree of agency in Japanese，whereas one narrator encoded a high degree of agency in Japanese but a low degree of agency in English.Five narrators encoded a low degree of agency in both languages，χ2（1，31）=8.88，p＜.003.Thus，there is a greater degree of encoding high agency in narration of both languages.

Table 4 Chasing Scene（Pictures 11-12）：English and Japanese（Degree of Agency）

Moreover，those narrators who provided a high degree of agency in English tended to do the same in Japanese；in like manner，those who encoded a low degree of agency in English tended to do the same in Japanese，r_s=.54，p＜.002.It thus appears that beyond the language of narration，bilinguals schematize the scene in similar or even identical ways in both languages.

Voice.As seen above，Taizo，for example，narrated the scene using the passive-progressive form in both languages.As seen in Table 5，however，23 narrators chose the active voice in both languages whereas four narrators chose the passive voice in both languages.Two narrators chose the active voice in English but used the passive voice when narrating the same scene in Japanese.Two narrators used the passive voice in English but used the active voice when narrating the same scene in Japanese，χ2（1，31）=10.67，p＜.002.Thus，there seems to be a greater tendency to use the active voice when narrating the chasing scene in both English and Japanese.

Table 5 Chasing Scene（Pictures 11-12）：English and Japanese（Active Voice or Passive Voice）

In addition，those narrators who used the active voice in English also tended to do so in Japanese；likewise，those who used the passive voice in English also tended to do so in Japanese，r_s=.59，p＜.001.The selection of the voice，which signals who is the topic，shows perspective or schematization differences between individuals，but not necessarily within individuals.Thus，we may conclude that regardless of the language of narration，bilinguals tended to describe the scene from the same perspective.

Completion or continuation.Kazuo [Boy；Grade 6.8；Age 12 years and 1 month] considered all actions of the boy，the dog，and the bees to be continuing and thus narrated，“Mark was still searching for his frog” and “All the bees were chasing Mark's dog，and the dog was running hard”.Perhaps recognizing the durativity of searching，he also narrated，“Maaku wa kaeru o sagashi-te-i-mashi-ta（Mark was searching for his frog）”，but he narrated “nanbyappiki no hachi wa zenbu Maaku no inu o oikake-mashi-ta（“Hundreds of bees all chased Mark's dog”） as if it were punctual and completed.The same individual thus might see the same action to be perfective and durative（imperfective） in different languages.

Fourteen narrators described the scene as a completed action in both languages.On the other hand，four narrators chose the progressive form in both languages（see Table 6）.Eight narrators who chose the progressive in English selected the completive in Japanese.[Note that one narrator（Melissa） used the progressive form in Japanese to describe the scene，but she did not refer to the scene in English.] While duration of an action was observed in the two languages as seen above，the “start + to infinitive or gerund” form，which indicates that a particular action begins，showed a great difference between the two languages.Five English narrators chose the “start + to infinitive or gerund” form，such as “The bees came out and started to chase the dog”（Nicole）.This form indicates the moment of a particular action taking place and that action continuing afterward for a certain duration of time.While such a form is also available in Japanese，it was not observed at all in Japanese narrations.When the “start + to infinitive or gerund” form is also counted as duration，then，four narrators described the scene as a continuing action in both languages whereas 14 described the scene as a completed action in both languages.

Thirteen narrators described the scene as a continuing action in English，but they described the same scene as a completed action in Japanese.Yet there was no narrator who described the scene as a completed action in English but as a continuing action in Japanese，χ2（1，31）=3.78，p＜.06.[Note that，as reported above，although Melissa described the scene as a continuing action in Japanese，her narration was not included since she did not refer to the scene in English.] Thus，there are two predominant pairings：（1） using the completive in narrations of both languages，and（2） using the completive in Japanese narration but the continuative in English narration.One possible interpretation is that bilingual narrators tended to focus more on the process or continuation in English narration，but tended to focus on the completion or even the result of an event in Japanese narration.In other words，bilingual children，when narrating in English，may adopt a process-focused point of view whereas they adopt a result-focused point of view when narrating the same scene in Japanese.Another interpretation is that the language influences the form of expression even though the bilingual narrator perceives or schematizes a certain event in similar ways.

Table 6 Chasing Scene（Pictures 11-12）：English and Japanese（Completive or Continuative）

It has turned out，however，that those narrators who used the completive in English also tended to use the completive in Japanese；similarly，those who used the continuative（or durative） in English also tended to use the continuative in Japanese，r_s=.35，p＜.06.These seemingly mixed results suggest to some degree that while the choice of language may influence the preference of form，bilingual narrators conceptualize and schematize the scene in similar or even identical ways—whether complete or continuing—in both languages.

The Deer Scene

Prior to the deer scene（Pictures 15-17），the boy climbs onto a rock，holding onto some branches，which unexpectedly turn out to be a deer's antlers.This event is accidental and surprising both to the boy and the deer，and the deer picks the boy up and throws him into a pond.

English.The deer scene turned out to be much more complex than the chasing scene.Thirty-nine narrators narrated Picture 17 in which the deer throws the boy into a pond.Using transitive verbs，such as “throw”，“dump”，“knock”，and “drop”，22 narrators，more or less，selected a Cause-View with the deer as topic.Seventeen narrators，in contrast，did not encode an agency，such as “The boy and the dog fell”（Kaori）.Note that one narrator，Mika [Girl；Grade 2.8；age 7 years，9 months]，narrated，“The reindeer fell the boy to the pond”，but this use of verb “fall” should be interpreted as a retrieval error.It should be interpreted that she wanted to say “The reindeer dropped the boy into the pond”.High agency is correlated to the transitive verb whereas low agency is correlated to the intransitive verb，r_s=.95，p＜.0001.[Note that the Spearman rho（ρ） or r_s is a special case of the Pearson product-moment correlation coefficient.] The fact that the correlation was not perfect is attributable to one narrator's [Jon；Boy；Grade 3.9；Age 8 years，11 months] choice of a transitive verb “bring”，which was judged to have a lower degree of agency.

When viewing Pictures 15-17 as a sequence，we have a better sense of the narrators’ thought processes（see Table 7）.Story characters can be described in various ways：（1） using a full noun phrase，a reduced noun phrase（e.g.，a pronoun），or zero-marking/zero anaphora（ellipsis/omission of an overt reference term）.If pronouns are used excessively，for instance，the listener may have difficulty determining their antecedents，as in “Then，he got stuck on the deer's horn.The deer took him to a high mountain，and he dropped him down to the lake”（Taichi [Boy；Grade 2.7；Age 8 years，0 months]）.A more relevant issue here is encoding of agency.Describing Pictures 15-17，for example，Maria [Girl；Grade 1.7；Age 6 years，6 months] narrated，“The big deer came and picked him up，and threw him down from the big cliff”.In this example，Maria encoded a relatively high degree of agency by using the verbs “pick up” and “throw”（in contrast to the diminished agency of the deer as seen in the use of a neutral verb such as “get”：“Then，he（=the deer） gets him（=the boy） on”（Takuto [Boy；Grade 1.7；Age 6 years，11 months]）.As a matter of fact，four narrators chose “pick（up）” in either Picture 15 or 16，and seven narrators chose the verb “throw” in Picture 17.

In the sequence of these three pictures，12 narrators chose the deer as topic or subject of a verb of motion throughout.In contrast，only one narrator chose the boy as topic or subject of a verb of motion，namely，“He went on the reindeer and he fell in the swamp”（Hanayo [Girl；Grade 5.7；Age 11 years，2 months]）.In this example，we do not see any Cause-View，and virtually no agency is attributed to the deer.However，this does not necessarily mean that we cannot expect high agency when the boy is used as topic.With the boy as topic in Picture 15，one narrator，Misaki [Girl；Grade 2.7；Age 7 years，9 months]，used a simple active verb of posture，“He held onto something，which was a moose”.

Twenty-three narrators used a mix（the boy and the deer as actors）.Four of them included passive constructions at least somewhere in one of the three pictures，and three were in the form of the get-passive：“The boy got stuck on the deer's horn”（Hiro，Taichi，Picture 15） and “He got carried”（Mike [Boy；Grade 1.7；Age 7 years，7 months]：Picture 16）.While a be-passive appears only once [“He was picked up by something”（Taizo：Picture 15）]，none of them used passive constructions—either a get-passive or a be-passive—when describing Picture 17，such as “The boy was thrown/dropped/dumped by the deer”.These narrations give us a Cause-View.On the other hand，narrations such as “It was a deer's horns，and the deer went to the cliff，and the boy fell off his head”（Naoto [Boy；Grade 4.7；Age 10 years，0 months]） and “The deer ran and suddenly stopped，and the boy and the dog fell”（Keigo [Boy；Grade 4.7；Age 9 years，7 months]） do not necessarily give us a Cause-View even if the narrators intended to imply it.

Table 7 Deer Scene（Pictures 15-17）：English

续表

Japanese.In Picture 17，the event can be seen from the perspective of the boy as subject of a verb of motion.When this happens，unless the passive voice is used，the boy simply falls into a pond.That is，those narrators who treat the boy as topic and actor，unless they are using the passive form and thus maintaining the boy's volitionality，describe the scene as if the boy（and the dog） simply fell from the cliff.As a matter of fact，more than half of the narrators（21） expressed diminished agency by choosing the intransitive verb “ochiru（fall）” or its variations（okkochiru，ochichau），and thus did not take any Cause-View.With the deer as topic，on the other hand，the narrator does not fail to take a Cause-View.Kazuo，in fact，used both active and passive voices，“Maaku o soko ni omoikkiri otoshi-mashi-ta（ø threw Mark away very strongly to that place）” “inu mo isshoni otos-are-mashi-ta（the dog was thrown away together with him）”.

The event can be seen from the perspective of the boy as subject of a verb of motion.In order to depict it that way，there is a need to choose the passive voice.As mentioned above，three narrators used the passive voice with the boy as subject.By doing so，they defocused the agent（i.e.，the deer） with concomitant topicality of the patient（i.e.，the boy），but they still maintained a Cause-View perspective.But here，as also seen above，we need to simultaneously consider the Japanese agentless passive，and in order to understand its function we need to see the entire sequence（Pictures 15-17）.[Note：Ten narrators used a copula when narrating Picture 15，but those linking verbs are not listed in Table 8.]

As mentioned earlier，story characters can be described in various ways.From a syntactic perspective，Japanese，in contrast to English，can be categorized as a so-called pro-drop language due to frequent zero pronouns，and thus，the narrator's referential choice could be a serious issue leading to possible confusion on the part of the listener.However，even if the narrator uses zero-marking，narration may be clear to the ear of the native speaker of Japanese as long as the subject stays the same.One clear example of this is Kenji's [Boy；Grade 6.7；Age 12 years，6 months] narration：“soshitara soko ni shika ga ita node，sono ue ni nokke-rare-te，ike ni otos-are-mashi-ta（then，because there was a deer there，ø was carried on its back and thrown into a pond）”.If the subject changes and ellipsis is used at the same time，on the other hand，it will be very difficult to identify the subject.

Here high agency is perfectly correlated to transitive verbs whereas low agency is perfectly correlated to intransitive verbs，r_s=1.000，p＜.0001.Caution is warranted，however.As we have reviewed，agency can be heightened by lexical choice of the verb（or verbs） as well as associated adverbs.

Table 8 Deer Scene（Pictures 15-17）：Japanese

续表

Japanese words：Nouns：doobutsu（animal）；inu to kono hito（dog and this person）；kodomo to inu（child and dog）；mono（thing）；otoko no ko（boy）；otoko no ko to inu（boy and dog）；shika（deer）；shika no tsuno（deer's horn）；tonakai（reindeer）；wan chan（doggie）

Verbs：agaru（to rise）；aruki-hajimeru（to start moving/walking）；au（to meet）；butsukatte iku（to bump into）；dete kuru（to appear）；hairu（to enter）；hanasu（to release）；hasama-reru（to be sandwiched） [passive voice]；hashiri-dasu（to start running）；hashiru（to run）；hashitte iku（to run forward）；hikkakaru（to hang on）；hoori-koma-reru（to be thrown into） [passive voice]；hoori-komu（to throw into）；hoori-nageru（to throw away）；iku（to go）；iru（to exist）；itchau（to have said）；kuru（to come）；mitsukeru（to find）；mochi-ageru（to raise）；motsu（to have/get）；motte-iku（to take）；mukau（to go toward）；noru（to get on）；noseru（to put something on）；notchau（to have got on）；notte iru（to be riding）；ochichau（to have fallen）；ochiru（to fall）；ochi-saseru（to let someone fall） [causative]；ochisoo ni naru（to be about to fall）；oikakeru（to chase）；okkochiru（to fall）；okkochite shimau（to have fallen）；okoru（to get angry）；orosu（to put down）；otosa-reru（to be dropped） [passive voice]；otosoo to suru（to try to drop someone）；otosu（to drop）；suru（to do）；tobu（to fly）；tobu（to take）；tsurete iku（to follow）

Similarities and Differences Between English and Japanese

This section examines whether or not English-Japanese bilingual children schematize the deer scene in similar ways in both languages.We will look specifically at whether children choose appropriate verbs，produce active and passive sentences，and encode agency accordingly.

Degree of agency.Examination of how the narrator views Picture 17 tells us whether he or she takes a Cause-View regardless of the language of narration（see Table 9）.

Table 9 Deer Scene：English & Japanese（Picture 17）

续表

续表

As seen in Table 10，in both English and Japanese，14 narrators consistently took a Cause-View（these narrators interpreted the scene in the way probably intended by the author）.In contrast，12 narrators did not take a Cause-View，regardless of the language.And 13 used a mixed view；that is，eight narrators took a Cause-View when narrating in English but did not do so in Japanese，whereas five narrators took a Cause-View when narrating in Japanese but not when narrating the same scene in English.We can see here two predominant pairings：（1） high agency in both languages，and（2） low agency in both languages，χ2（1，39）=4.50，p＜.05.

Table 10 Deer Scene（Pictures 15-17）：English and Japanese（Degree of Agency）

The above observation is also supported by the choice of verb type（see Table 11）.Although no differences approached statistical significance at the.05 level，there seem to exist two predominant pairings：（1） transitive verbs in both languages，and（2） intransitive verbs in both languages，χ2（1，39）=3.31，p＜.07.

Table 11 Deer Scene（Pictures 15-17）：English and Japanese（Verb Types）

Overall，those who encode a Cause-View in one language also tended to encode a Cause-View in the other language，whereas those who did not take a Cause-View in one language tended to behave in the same way in the other language，r_s=.34，p＜.05.Again，although the association is relatively weak，relationships seem to exist between the two languages in terms of the choice of verb type；those who chose a transitive verb in one language tended to select a transitive verb in the other language whereas those who selected an intransitive verb were also likely to do so in the other language，r_s=.29，p＜.08.

The above correlations seem to suggest that regardless of the language used for narration，those who interpret the scene as causal（assumingly intended by the author） encoded high agency by using a transitive verb，whereas those who interpreted the scene with less or no causality encoded diminished or no agency by using an intransitive verb.Overall，therefore，regardless of the language of narration，bilinguals tended to describe the scene from the same perspective and chose an appropriate verb accordingly.

Here，however，we need to think further about why inconsistency occurred.First，as to the 12 narrators who did not take a Cause-View perspective regardless of the language（see Table 10），there are two possible interpretations：

（1）They did not believe the boy's falling from the cliff was caused by the deer.The boy（and the dog） simply fell from the cliff（i.e.，the issue of schematization discussed above）.

（2）They did understand that the deer caused the boy to fall from the cliff，but they simply did not know what to say in either language（i.e.，the issue of language weakness/deficit）.

Second，as to the mixed use，there are at least two possible interpretations，both of which may be attributable to the issue of language weakness/deficit：

（1）They knew how to describe the scene appropriately in one language，but did not know how to encode it in the other language.

（2）They simply described the scene arbitrarily.

Voice.Regardless of the language of narration，the active voice was predominantly used in Picture 17（where the deer throws the boy into a pond），but language-specific tendencies were still observed.In English，four narrators used get- and be-passives in either Picture 15 or 16，but none used the passive voice when describing Picture 17.All narrators used the active voice when narrating Picture 17 in English.The choice of the active voice here means：（1） The event was seen from the perspective of the deer as subject of a verb of motion（by using such transitive verbs as “throw”，“drop”，or “dump”），thus with the suggestion of the nature of causality（i.e.，high degree of agency）；or（2） The event was seen from the perspective of either the boy or the deer as subject of a verb of motion，with minimal or virtually no degree of agency，as can be seen in Hiro's narration “The deer stopped at the cliff，and the boy fell into the pond”.

On the other hand，in Japanese narrations the passive form，although still infrequent，appeared exclusively in Picture 17.In addition to Kenji，whose narration was mentioned above，Asuka [Girl；Grade 5.7；Age 11 years，1 month] and Kazuo used the passive form.Interestingly，Asuka was a fifth grader and Kenji and Kazuo were sixth graders at the time of the data collection.Certainly，age or grade may not correlate with the development of bilingual skills；this is because as children age or advance in school，either one of the two languages may become the primary language and the other may decline.But the use of passive in this case may make a sophisticated impression on the Japanese listener.