In our recent article in Child Development we sought to test the claim of Hart and Risley (1995, hereafter HR) of a 30-million-word gap between the amount of vocabulary heard by our nation’s poorest and most affluent children. In the 23 years since it was published, the original HR study has become a touchstone for many scholars, policymakers, and practitioners attempting to improve educational outcomes for disadvantaged children. Our new study’s findings challenge this established narrative, and—to our surprise—triggered a flurry of reactions from many corners weighing in on both sides of the debate.
Given our 40 years of work in low-income communities, it did not make sense to us that all impoverished children were deprived of language input from their primary caregivers during the early years of life. Our reservations were supported by ethnographic studies of young children’s language from interdisciplinary fields of study, especially developmental cultural psychology, language socialization, and education. By contrast, the literature on vocabulary development and subsequent school achievement rarely acknowledged the real lives lived by children and their families.
Our new study is a hybrid, drawing insights and methods from both ethnographic studies and studies of vocabulary. We analyzed longitudinal language data of young children and their families from five diverse American communities. Three communities were urban and two were rural. Two were poor, two were working class, and one was middle class. One was African American, and the others were European American. Each community was distinctive with its own beliefs and attitudes about what family life should be like, how children should be raised, and how language should be practiced in their homes. Our participants did not live in single-parent households exclusively (as did HR’s Welfare children); many lived in large families with several siblings and frequent comings and goings of other family members and friends.
To capture that richness, we analyzed three different definitions of the verbal environment. First, we counted the number of words spoken to children in directed speech by their primary caregivers, in parallel with HR’s definition. We then analyzed a second condition, namely the number of words spoken to children in directed speech by all caregivers (mothers, fathers, siblings, other relatives). Finally, we counted the number of words spoken around them but not addressed to them, words that they might overhear. An example would be grandmother telling the child’s aunt a story about what the child did the day before.
Our analyses demonstrated that our poorest communities did not fall out along income lines, unlike HR’s findings. While the middle-class and working-class communities were similar to their counterparts in the HR study, children from South Baltimore (a poor community) heard 1,048 words per hour from primary caregivers compared to the Welfare group children in HR who heard 616 words per hour. Most astonishing, however, was the result that children from our Alabama African American community heard 1,838 words per hour spoken by their primary caregivers, nearly three times the number of words heard by the Welfare group children in HR. This result cannot be overemphasized; both our Alabama group and the HR Welfare group were similar not only in terms of SES, but also ethnicity. These results suggest that the HR Welfare group was an outlier.
Our other two definitions similarly demonstrated that differences in the number of words children hear cannot be predicted by socioeconomic standing alone. In each community, when we counted the number of words spoken to the child by all family members, the number of words increased between 17 and 58 percent. When speech around the child was counted, the total number of words in the child’s ambient environment increased between 54 and 210 percent. The variability in these conditions was not predicted by family income, as the advocates of the word gap would suggest. These results challenge the unspoken assumption in many studies of language-learning children that only one person, the mother, routinely talks to the child.
In a recent blog post on the Brookings Institution forum Education Plus Development, our work was criticized on several counts. First, the authors noted that our study was not a complete replication of HR’s famous study. This criticism is accurate—we did not have an extremely highly educated group of parents similar to HR’s Professional group. Of course, the absence of this group neither takes away from our essential point that our poorest groups outperformed HR’s Welfare group nor erases the possibility that the lower end of the number of words directed to poor children is not as low as HR reported.
A second criticism related to our focus on overheard speech: the authors suggested that overheard speech has no significant role in language learning for “children just beginning their language-learning journey.” This point has been disproven by decades of ethnographic research (e.g., Brown & Gaskins, 2014; Duranti, Ochs, & Schieffelin, 2012; Miller, Cho, & Bracey, 2005; and in a growing number of laboratory studies (e.g., Akhtar & Gernsbacher, 2007; Scott & Fisher, 2009; Shneidman et al., 2009). Whether overheard speech is the most efficient vehicle for children just learning to talk is debatable; however, our participants were on average 2 years old when our studies began. Overheard speech may gain in importance as children get closer to school age. After all, what parent has not at one time or another said an unfortunate word to her spouse only to hear her child repeat it at an equally unfortunate time?
Although the intent is not malicious, when one focuses exclusively on “fixing” the language of poor, working-class, and minority parents and children, one tacitly engages in deficit thinking. Interventions based on deficit thinking imply that school success requires that such children must “catch up” even before they enter kindergarten or first grade. Are there other alternatives? Yes. We advocate the adoption of school-based curricula that focus on all children’s language strengths. One such strength is oral narrative, a highly sophisticated speech genre. In our earlier research, we found that two year olds from our poor and working-class communities participated precociously in co-narrations (directed speech) and in overheard speech (stories to which they were bystanders). Through pedagogical approaches rooted in an understanding of the sociocultural variation inherent in language, no child will feel demeaned or devalued if their family does not follow mainstream practices at home.
By seeking a more nuanced understanding of social class, one that takes into account cultural and community differences within and across income levels, we hope to renew interest in the many literatures that demonstrate the verbal strengths of diverse children. Language is richer than vocabulary alone; language environments are richer than a primary caregiver’s directed speech. We posit that educational approaches to language achievement must be similarly rich.
The Brown Center Chalkboard launched in January 2013 as a weekly series of new analyses of policy, research, and practice relevant to U.S. education.
In July 2015, the Chalkboard was re-launched as a Brookings blog in order to offer more frequent, timely, and diverse content. Contributors to both the original paper series and current blog are committed to bringing evidence to bear on the debates around education policy in America.