Carol Miller and Ji Sook Park*
In every kindergarten classroom, there are usually one or two children who have difficulty learning spoken language. It can be hard to identify them, because they seem to be typically developing in other ways. Their hearing is adequate, and they do not have autism, intellectual disability, or neurological disorders. The terms specific language impairment (SLI) and developmental language disorder (DLD) are used to describe such children. Although we still have a lot to learn, we have many tools to help us identify children with DLD and choose an appropriate intervention—if they are monolingual English speakers, that is. Identifying DLD in bilingual children is more challenging for several reasons.
One of the reasons it can be difficult to identify DLD in bilingual children is that we don’t yet have the right assessments. Imagine arriving in a new country where you’re just beginning to learn the language, and being asked to take a language test in this new language. Do you think that test would give an accurate representation of your general language abilities? You probably answered no to that question, yet many schools are forced to use this method because assessment tools are unavailable in many of the languages spoken by children from diverse home language backgrounds. But even when assessments are available in the child’s home language, they often prove ineffective. This is because they are typically designed for monolingual speakers, so they do not account for important differences between bilinguals and monolinguals, such as levels of exposure to a language, and different contexts for its use (e.g. home vs. school). In addition, typically developing bilingual children may make grammatical errors that look similar to errors made by a child with DLD, simply because they are still learning to distinguish between the rules of their languages.
Because of these challenges, we have asked whether we could use non-linguistic tasks to identify children with DLD. Why non-linguistic? We know that children with DLD have difficulty with some non-linguistic cognitive tasks. It is not clear if these cognitive issues cause their language impairment, or are caused by it. Nonetheless, if we find that children with DLD perform consistently worse than their typically developing peers on certain non-linguistic tasks, we might be able to use those tasks to identify DLD. If we can find tasks that will identify children with DLD independently of whether they are learning multiple languages, then we can use these tasks to help educators more accurately identify bilingual (and monolingual) learners with DLD.
Our research group has explored two non-linguistic tasks. One is a type of procedural learning task. Procedural learning happens when individuals’ performance speeds up after a certain amount of exposure to sequences with a similar structure. Typing is a good example of this. As you practice, you get faster and more accurate, and if you’re a good typist, you actually do it better if you don’t think about it too much. In our procedural learning task, there are 4 boxes on a computer screen, with 4 corresponding buttons to press. A cartoon monster appears in one of the boxes (see Figure 1), and the child has to press the matching button. Sometimes the order in which the monster appears is random, but sometimes there is a repeating sequence. If procedural learning is happening, children will get faster in pushing the buttons during the repeating sequences, because they are anticipating the next location—even though they aren’t aware of it.
Our other task looks at how children control their attention. A row of five fishes appears on the screen (see Figure 2). The child’s job is to indicate which way the middle fish is facing. The surrounding fish may be facing the same way (congruent) or the opposite way (incongruent). As you might imagine, children tend to be slower when the surrounding fish are facing the opposite direction. However, children vary in how much they are slowed by this incongruence between the target fish and its neighbors. And sometimes we provide clues to help the child respond more quickly, but it only helps if they can control their attention enough to take advantage of those clues.
We tried these tasks with bilingual and monolingual typically developing children and children previously diagnosed with DLD. For the procedural learning task, we found that children with DLD did not get faster during the repeated sequences, whereas the typically developing children did. Importantly, monolingual and bilingual children in each group learned equally well, which suggests that our task was successful in distinguishing bilingualism from DLD.
For the task that requires attention control, we found that children with DLD were slower than typically developing children when the fish surrounding the target fish were pointing in the opposite direction, while again, monolingual and bilingual children’s performance was the same.
It is important to note that before any task can be used as a test for DLD, it has to meet several scientific criteria, and we do not yet know whether our tasks meet these criteria. But the results point us in a promising direction toward developing better tools to identify bilingual children who have DLD, and therefore, toward providing these children with appropriate treatment and intervention as early as possible.
*This research was conducted by Ji Sook Park and Elina Mainela-Arnold as principal investigators, along with collaborators Carol Miller, Teenu Sanjeevan, Janet Van Hell, and Daniel Weiss. A portion of this work was published in 2018 in the Journal of Speech, Language, and Hearing Research: https://pubs.asha.org/doi/10.1044/2017_JSLHR-L-16-0409.
Rueda, M. R., Fan, J., McCandliss, B. D., Halparin, J. D., Gruber, D. B., Lercari, L. P., & Posner, M. I. (2004). Development of attentional networks in childhood. Neuropsychologia, 42, 1029–1040.
Tomblin, J. B., Mainela-Arnold, E., & Zhang, X. (2007). Procedural learning in children with and without specific language impairment. Language Learning and Development, 3, 269-293.