Cognates are advantaged over non-cognates in early bilingual expressive vocabulary development

Abstract

Bilingual infants grow up with the unique experience of needing to learn two words for most concepts. These words are called translation equivalents, and translation equivalents that also sound similar (e.g., banana—banane) are called cognates. Research has consistently shown that children and adults process and name cognates more easily than non-cognates. The present study explored if there is such an advantage for cognate production in bilinguals’ early vocabulary development. Using longitudinal expressive vocabulary data collected from 47 English–French bilingual infants and toddlers starting at the ages of 16–20 months up to 27 months (a total of 219 monthly administrations in both English and French), results showed that overall children produced a greater proportion of cognate words than non-cognate words on the MacArthur-Bates Communicative Development Inventories. The findings suggest that cognate learning is facilitated in early bilingual vocabulary development. Just as in monolingual infants, these results suggest that phonological overlap supports bilingual language acquisition.

The role of translation equivalents in bilingual word learning

“…Specifically, we used data from Words & Sentences forms, which measure children's productive vocabularies. We used bilingual datasets from two origins: three datasets contributed to Wordbank (Frank et al., 2017; associated with Hoff et al., 2012, 2018; Legacy et al., 2018; Marchman et al., 2004), and one dataset (associated with Mitchell et al., 2022) openly available on the Open Science Framework, with a total of 2092 administrations from 518 children. These data were collected using the monolingual American English form and either the monolingual Mexican Spanish or the monolingual Quebecois French CDI form.…”

“…Specifically, TEs vary in the degree of phonological similarity between the two items in each language—for example, dog [eng] and perro [spa] have no phonological overlap, whereas animal [eng] and animal [spa] are very similar phonologically. Indeed, previous research has suggested that TEs with greater phonological overlap are more likely to be TE pairs (Grasso et al., 2018; Umbel et al., 1992), and bilingual children produce more TE pairs that have high phonological overlap than TE pairs that have low phonological overlap (Bosch & Ramon‐Casas, 2014; Mitchell et al., 2022; Umbel et al., 1992). TEs can also vary in the lexical categories of their items: They could label, inter alia, objects, actions, or attributes, which broadly correspond to nouns, verbs, and adjectives.…”

“…">Knowing a word predicts knowing its corresponding TE (due to previous results about TE knowledge, e.g., Anaya, 2021; Grasso et al., 2018); and Greater phonological similarity predicts greater likelihood of word production when the word's TE is known (based on studies suggesting that phonologically similar TEs are learnt earlier, e.g., Mitchell et al., 2022; Umbel et al., 1992). …”

“…Greater phonological similarity predicts greater likelihood of word production when the word's TE is known (based on studies suggesting that phonologically similar TEs are learnt earlier, e.g., Mitchell et al., 2022; Umbel et al., 1992).…”

Lexical development in an Indonesian-Balinese bilingual child

“…There are a number of studies report the production of TEs in the early lexicon of bilingual children involving various language pairs, Romanian-Italian and Nigerian English-Italian (Barachetti et al, 2022), French-English (Jardak & Byers-Heinlein, 2018;Legacy et al, 2017;Poulin-Dubois et al, 2018), German, Spanish-English (Floccia et al, 2020), Spanish-English (Gimenez-Arce, 2019Shiro et al, 2020), German-English (De Anda & Friend, 2020), Spanish-Catalan (Mitchell et al, 2022), German-Indonesian . However, most studies on early lexical development in bilingual children involve language pairs from the Indo-European language family.…”