Quantity, lexical diversity, and syntactic complexity in bilingual picture books and shared reading interactions

• In Event: Opportunities for learning from picture books and shared reading in monolingual and bilingual environments

Abstract

Bilingual books, which present text in two languages, have been increasing in popularity (Domke, 2018), and can provide rich opportunities for language learning (Read et al., 2021). Yet, little is known about how different types of bilingual books present text in two languages, which is crucial for understanding dual language learning opportunities. We examined the text in Spanish/English Translation books, which present the narrative text in both languages, and Codeswitching books, which present English and Spanish intermixed (See Figure 1). We first explore the use of English vs. Spanish text in Translation books, and then compare Translation books to an already-existing, age-matched corpus of Codeswitching books.

Method. Expanding on previous research (Benitez et al., 2022), we coded 45 Translation Spanish-English bilingual picture books targeted at ages 0-9 years. Books were transcribed and analyzed in CLAN for the number of words (tokens), number of unique words (types), number of utterances, mean length of utterance (MLU), and type/token ratio.

Results. Translation books provided relatively equal amounts of text in English and Spanish, but even with direct translations, some differences were present. The type/token ratio was significantly higher in Spanish than English, [t(38)=2.53, p=0.02], suggesting more diversity of Spanish words. Number of utterances, word types, tokens, and MLU did not differ significantly by language (p’s>0.05).
We then compared the text in Translation and Codeswitching books (N=29 in each age-matched corpus), both overall and by language. Book types did not differ significantly in number of total word tokens, types, and utterances nor by type/token ratio or MLU, suggesting similar overall richness of input. However, Translation and Codeswitching books differed in the amount and complexity of English vs. Spanish text. Codeswitching books contained predominantly English text (averaging 85% English), whereas Translation books presented equal amounts of English and Spanish, although, Translation books were twice as likely to present English text first [71%, χ²(1)=17.64, p<0.001], possibly reflecting asymmetry in language prominence. Compared to Codeswitching books, Translation books presented more Spanish tokens [t(43.46)=-5.24, p<0.001], types [t(48.31)=-5.3, p<0.001], and utterances [t(58.63)=-5.68, p<0.001] and longer MLU [t(59.14)=-10.03, p<0.001] (See Figure 2). Codeswitching books presented more English word types [t(45.99)=3.46, p<0.001] than Translation books.

Discussion. We found that Translation books offer potentially similar learning opportunities in English and Spanish, while Codeswitching books present significant imbalances of English vs. Spanish input, offering less diverse and complex Spanish text than Translation books. Thus, language content is presented in distinct ways across Codeswitching and Translation books, which may have implications for how parents read these books to their children. Work in progress is exploring how bilingual parents read Codeswitching and Translation books to their 3- to 4-year-old children. This research aims to understand how the richness of text in books may influence parents’ language use.

Authors

• Marissa Anne Castellana, Arizona State University
  Presenting Author

• Matthew Guerra

• Christine Potter, University of Texas at El Paso

• Viridiana L. Benitez, Arizona State University