Input and output in code switching: a case study of a Japanese–Chinese bilingual infant

Abstract

Code switching (CS) (or language mixing) generally takes place in bilingual children's utterances, even if their parents adhere to the ‘one parent – one language’ principle. The present case study of a Japanese–Chinese bilingual infant provides both quantitative and qualitative analyses on the impact of input on output, as manifested in CS. The database consists of the transcription of 12 hours of audio recordings, selected out of nearly 130 hours of recordings of an infant from the age of 2;1 to 3;0. The quantitative analysis exhibits some influence of parental input on the subject's output in regard to CS. In Qualitative Analysis A, the parental discourse strategy toward the child's CS does account for the overall high ratio of output CS. However, the parental adjustment of input strategies contradicts the present data of output. Qualitative Analysis B concerns the notion of language dominance in terms of intra-sentential CS. Our findings show that, in spite of the absolute Chinese dominance evident in the parental input, the child maintains a strong Japanese-dominant output. Hence, as far as CS is concerned, this study obtains little evidence of qualitative influence of parental input on the child's output. Instead of parental input, language dominance can better account for the increased trend of the child's output CS.

Keywords:

• bilingual first language acquisition
• code switching (mixing)
• input and output
• parental discourse strategy
• language dominance

Notes

1. As a matter of fact, the distinction between ‘code switching’ and ‘mixing’ concerns a long-term debate. ‘Code switching’ tends to be adopted more frequently in adult bilingualism, to refer to a certain skill of the bilingual speaker. It requires grammatical and pragmatic competence in both languages. On the other hand, ‘language mixing’ is associated more with young bilingual children, to indicate their limited lexicon, and lack of grammatical and pragmatic competence.

2. Among bilingual and multilingual studies, it is generally believed that the language with social community support tends to become more dominant in the children, rather than the parents’ native languages. It has been proved by many researchers, such as Hoffman and Widdicombe (1999), Yamamoto (2001), etc. Hence, active bilingualism must be actively cultivated by the parents.

3. The selection of the data for transcription follows two principles. One is to make the time distribution average one hour per month. The other is an intention to select those with comparatively better quality of recording, and those that contain quantitatively more utterances from the target child.

4. Note that the present database of 12 hours transcriptions does not contain any Japanese-dominant conversation recording. Hence, the analysis on language input here is merely limited to the Chinese-dominant input from the parents at home. A sample of the transcription segment can be found in the Appendix.

5. Since the input environment is Chinese-dominant led by the parents, the base language for all the recordings is regarded as Chinese. Hence any Japanese utterances will be considered inter-sentential CS in both input and output.

6. Origin is a kind of software for graphing and data analysis. The general trend of different types of utterances in both input and output can be indicated by the value of the slope in the affiliated tables, which can be automatically generated through the operation of linear fitting in Origin.

7. The series of number that appears after every example corresponds to the name of the recording file, indicating the recording date. Note that in all the examples, Japanese words will be in italics and bold, demonstrating the CS.

8. The rest of 56 (98–42) tokens of input intra-sentential CS utterances usually include some explanations from the mother to the subject. For instance, *MOT: zhè jiào xiăo-mántou, manjuu. (07081300) ‘This is called a steamed bread, steamed bread.’

9. Since the diary data have only contributed to the calculation of types, not to tokens, for qualitative claims, we have based our analysis on the type of data.

10. For a detailed examination of language dominance, please refer to Meng and Miyamoto (2009): CS in a Japanese–Chinese bilingual infant: a study on language dominance.