Within language development, we know that divergences based on socioeconomic status (SES) are evident early in life and well established by the time a child enters school. It is widely acknowledged that understanding SES effects on early language is critical for reducing achievement gaps in school readiness. In the area of grammatical development, higher-SES caregivers produce greater input quantity and more complex constructions relative to lower-SES peers. Yet, the exact consequences of these input differences remain unclear. Past research often rely on coarse-grained measures of language abilities (e.g., vocabulary size, length of sentences produced, etc.), which tell you where SES differences exist but provide limited insights into how they came about in the first place. Without a causal model, most interventions assume that frequent input (e.g., hearing more passive sentences) allows children to acquire linguistic representations. Children who hear lot of input are able to learn the grammar of their language, but those who don’t get a lot of input don’t learn.
Our research tests a different hypothesis. Maybe all children (across all SES families) learn grammatical structure with minimal input, but hearing more input allows children to retrieve their knowledge from memory more efficiently during comprehension. Thus, SES effects on development reflects not a failure to learn language but challenges with *retrieving* what you’ve already learned. We tested English-speaking 3- to 7-year-olds from varying SES families on their comprehension of passives, which are infrequent in general and particularly so among lower-SES groups. Yet, all children (across SES) used grammatical cues to correctly interpret passive when comprehension demands were minimal (e.g., “It is eaten by the seal” –> “it” is a fish). This demonstrates that input variation does not alter children’s ability to learn low-frequency structures. Yet, their comprehension was less accurate when an initial misinterpretation needed to be revised (e.g., “The seal is eaten by it” –> initially interpret “seal” as the predator, but can’t reinterpret it as the prey). This was particularly true for children from lower-SES families. Here, measures of real-time processing of the linguistic cues predicted the accuracy of final interpretation. Relative to higher-SES peers, children from lower-SES families were less sensitive to passive cues, and inefficient retrieval of this grammatical structure limited their ability to overcome their initial incorrect interpretation.
These findings also shed light on well-documented differences in vocabulary size. Current interventions assume that children’s failure to learn words reflects a lack of relevant input (e.g., not learning “stethoscope” if you’ve never heard “stethoscope”). Under this approach, closing the “word gap” in caregiver input will solve the achievement gap (e.g., 30 Million Words Initiative). Yet, our research suggests that even when input exists, learning can be challenging if children can’t accurately retrieval grammatical knowledge in order to comprehend sentences and learn from caregiver input. Consistent with this account, we found that children’s vocabulary size correlated with their likelihood of overcoming comprehension demands, but not with general measures of grammatical knowledge. Taken together, this study suggests that isolating why outcomes varies across populations requires identifying not just what children hear but how they use it.
For more information, check out the full paper in Cognition.