I got into research in part because I was curious in lexical organization in bilinguals and in bilingual language impairment. Sometimes I feel like I’ve gotten distracted doing other kinds of work. So, it’s kinda fun to get back to something that I feel has gotten neglected.
How do children learn and organize their vocabulary? As children learn new words, they have to compare them with the words they already know. Words might sound the same but have different meanings (e.g., hoarse vs. horse). They also can compare words by category (e.g., chair, sofa) and function (e.g., cup, drink). These comparisons help children to make associations among words, and this helps children build their vocabulary knowledge. For bilinguals, it’s not so different, but the comparisons are made within languages and across languages. Across languages, children need to make connections among words that sound the same and have the same (e.g., velocity, velocidad) and different (e.g., contest, contestar (answer)) meanings. Bilinguals need to associate translation equivalents (e.g., dog, perro). And they need to keep word classes (e.g., nouns, verbs, adjectives) straight within each language. Children with language impairment have a difficult time learning new words, and making these associations between them. We were interested in how bilingual children with language impairment did this. We wanted to know if their patterns were different from typical bilingual children and in what ways. In this study, we examined Spanish-English bilinguals.
In this study (Sheng, Peña, Bedore & Fiestas), we used a repeated associations task, which my colleague Li Sheng had used with Mandarin-English bilinguals; and with monolinguals who had language impairment. In this task, children are given a word and asked to produce a word that goes with the one given. This is repeated 3 times. A child might hear, “tell me a word that goes with chair.” They might say, “couch.” Then, next time they might say, “sit” on a third repetition, they might say, “table.” This list of three words gives us a picture of whether children have the semantic depth in their vocabulary to retrieve an appropriately related word.
So, what happened? Consistent with the Sheng & McGregor (2010) study, children with language impairment produced fewer related responses than children with typical development. We used conceptual scoring, where a total score was derived based on the best language for each item. Children’s conceptual scores were better than either of the two single language (Spanish vs. English) scores. This means that children made associations in each of their languages and that a mix of the two languages gave the best representation of what they could do.
Children with language impairment made more errors in both languages compared to those with typical development. Children with typical development also more frequently (and successfully) switched to English during testing in Spanish, even though both groups performed better in Spanish. This may mean that typically developing children are starting to make associations between the two languages which will help them in developing English.
Based on the findings we report in the paper, we recommend use of conceptual scoring. Taking both languages into account resulted in the highest scores for the children and this score maximized the observed differences between children with and without language impairment, which is diagnostically useful.