Why does parental language input style predict child language development? A twin study of gene–environment correlation

Philip S Dale,Maria Grazia Tosto,Marianna E Hayiou-Thomas,Robert Plomin

doi:10.1016/j.jcomdis.2015.07.004

Abstract

There are well-established correlations between parental input style and child language development, which have typically been interpreted as evidence that the input style causes, or influences the rate of, changes in child language. We present evidence from a large twin study (TEDS; 8395 pairs for this report) that there are also likely to be both child-to-parent effects and shared genetic effects on parent and child. Self-reported parental language style at child age 3 and age 4 was aggregated into an ‘informal language stimulation’ factor and a ‘corrective feedback’ factor at each age; the former was positively correlated with child language concurrently and longitudinally at 3, 4, and 4.5 years, whereas the latter was weakly and negatively correlated. Both parental input factors were moderately heritable, as was child language. Longitudinal bivariate analysis showed that the correlation between the language stimulation factor and child language was significantly and moderately due to shared genes. There is some suggestive evidence from longitudinal phenotypic analysis that the prediction from parental language stimulation to child language includes both evocative and passive gene–environment correlation, with the latter playing a larger role.Learning outcomes: The reader will understand why correlations between parental language and rate of child language are by themselves ambiguous, and how twin studies can clarify the relationship. The reader will also understand that, based on the present study, at least two aspects of parental language style – informal language stimulation and corrective feedback – have substantial genetic influence, and that for informal language stimulation, a substantial portion of the prediction to child language represents the effect of shared genes on both parent and child. It will also be appreciated that these basic research findings do not imply that parental language input style is unimportant or that interventions cannot be effective.

Highlights

Since the seminal work of Snow (1972), an extensive body of research has been conducted on the characteristics of speech directed to, or produced in interaction with, young children in the early stages of language acquisition, typically below three years of age
Numerous earlier studies found a positive relationship between quantity of input and rate of language development, this result was made more vivid to a wider population by the work of Hart and Risley (1995), who documented wide variations in the amount of speech to young children, and the extent to which these were reflected in vocabulary differences in the children
A child with a strong genetic endowment for mathematics, or for music, or for athletics does not became exceptionally skilled without years of practice; the genes may have their largest effect by inclining the child to spend his or her time in that practice

Summary

Introduction

Since the seminal work of Snow (1972), an extensive body of research has been conducted on the characteristics of speech directed to, or produced in interaction with, young children in the early stages of language acquisition, typically below three years of age (see Hoff, 2006; Rowe, 2012, for reviews of this work). Beyond the quantitative differences, numerous qualitative differences have been shown to predict child language both concurrently and longitudinally, generally with weak-tomoderate effect sizes. These qualitative, facilitating aspects of CDS are found in all the subdomains of language (phonology, lexicon, syntax, semantics, and pragmatics), and include greater vocabulary diversity, appropriately scaffolded mean length of utterance (MLU), increased repetition, exaggerated prosody, promotion of joint attention, proportion of conversationeliciting speech as opposed to behavior regulation, semantic contingency, decontextualized language use such as narratives, ‘grammatical tutorials’ such as sentence recasts and expansions, and others (Hoff, 2006; Rowe, 2012)

Methods

Results

Discussion

Conclusion