Receptive and Expressive Semantics: Does Bilingual Experience Matter?

The short answer is yes. But, the longer answer is more interesting I think. It’s well-known that we can understand more words than we can express. Generally though, there are strong associations between receptive and expressive language, the more words you understand the more words you can express. We see normal receptive-expressive gaps in early language development, later development, as well as in mature learners. As adults, there are words that we can recognize by context in reading for example, but don’t use them expressively or don’t consistently recall them. On standardized tests however, these inherent differences between the two kinds of tasks are controlled. We can compare performance on expressive and receptive tasks through use of standardized scores often using a mean of 100 and a standard deviation of 15. Large receptive-expressive discrepancies where receptive knowledge is much stronger than expressive knowledge can be an indicator of language impairment. How does this work in bilinguals?

From a logical perspective, it should work similarly. We would expect that bilinguals understand more words in each of their languages than they can express in each language. Often however, bilinguals start learning their second language after the first language. For children this is usually during preschool or kindergarten. So, they already have had 3 to 5 years of experience in one language before they start learning another. It’s known that the first language can help boost learning in the second language, so second language learners don’t start from nothing (like first language learners). These second language learners already have knowledge of words and their meanings in one language—it may not be too much of a stretch to attach another word to the concept they already have. Some languages (for example English and Spanish) have quite a number of cognates, so this may also help second language learners recognize words in a language they don’t know. On the other hand, second language learners may suppress their first language so that they can pay attention to the other language.

Gibson and colleagues noticed this receptive-expressive gap in a number of studies of bilingual children’s vocabulary performance. They analyzed dataset of kindergarten Spanish and English sequential bilinguals and found a larger receptive-expressive gap in children’s Spanish than English. Even when they divided children by those with higher and lower levels of English exposure, the gap in Spanish persisted. Why would this be? Presumably, these children have had more Spanish exposure over their lifetimes than English exposure. And most likely their English exposure was fairly recent. One intriguing possibility is that during early stages of learning a second language, children suppress or inhibit their first language. This is consistent with accounts of children’s “silent period” when first exposed to an L2. Doing this might help children to pay attention to the second language better. A study by Link, Kroll, and Sunderman with college students learning an L2 in different circumstances suggests that in immersion contexts second language learners inhibit access to their L1. We discuss this relative to assessment of L1 and L2 in a previous post.

In a more recent study, Gibson, Peña, and Bedore explored the receptive-expressive gap with a large group of children who were tested using the Bilingual English Spanish Oral Screener (BESOS). We looked specifically at the semantics subtest. The semantics subtest of the BESOS (and BESA for that matter) is not a single word vocabulary test it focuses on assessment of semantic relations. This way children can express what they know but have flexibility in what they say. For example, there are some items that require children to provide names of animals. It doesn’t matter which animals they name, just that they know at least a certain number and that they don’t make errors. Some of the items are receptive and some are expressive (at least the items for the 5 year olds) and so we were able to come up with subscores to compare. We had a large enough group of 5 year olds (about 750) to standardize the subscores. Then we compared children’s scores at 5 levels of exposure to English and Spanish. We found that for Spanish there were no differences between expressive and receptive scores for any of the 5 groups, from functionally monolingual English, dominant English bilingual, balanced bilingual and so on. For English however, there were significant gaps with receptive standard scores higher than expressive standard scores for children who had minimal exposure to English (note that because the scores are standardized, we would expect no real difference at the group level). The gap between standard receptive and expressive scores closed with increased exposure to English. This is a different pattern from what Gibson and colleagues had found previously– that there were receptive-expressive gaps in the Spanish of bilingual children. It may be that the differences in studies were due to task differences. We had developed the BESA and BESOS semantics tasks to allow for flexibility in WHAT kids named (as in the category task rather than having a specific word or picture that they needed to name). Whereas we might see more first language inhibition on tasks that require a specific name. It might be that in confrontation naming there is more competition between the two languages.

In these more recent findings, children with the least amount of exposure to English were those who had the largest gaps. So, they could understand more English than they could express. I think that this is important clinically. A relatively high score in receptive semantics may lead us to expect an equally high score on an expressive task in a child just beginning to learn a second language. But, this is not the case. Children seem to use knowledge of L1 to learn L2 and they can use their cognitive-linguistic skills to understand what is going on in a second language. But, at early stages of learning they look relatively weak in the L2. This could present as language impairment. So, we need to be cautious in interpreting these gaps and understand them with respect to children’s individual experiences.

Advertisement