Predicting Individual Variation in Language From Infant Speech Perception Measures

Abstract

There are increasing reports that individual variation in behavioral and neurophysiological measures of infant speech processing predicts later language outcomes, and specifically concurrent or subsequent vocabulary size. If such findings are held up under scrutiny, they could both illuminate theoretical models of language development and contribute to the prediction of communicative disorders. A qualitative, systematic review of this emergent literature illustrated the variety of approaches that have been used and highlighted some conceptual problems regarding the measurements. A quantitative analysis of the same data established that the bivariate relation was significant, with correlations of similar strength to those found for well-established nonlinguistic predictors of language. Further exploration of infant speech perception predictors, particularly from a methodological perspective, is recommended.Research over the past five decades demonstrates that infants tune their speech perception abilities to language-specific properties of the speech they hear before their first birthday (for a recent review, see Werker & Gervain, 2013). Although most of this past work has sought to pinpoint milestones by using group averages, increasing attention is being paid to correlations between behavioral and neurophysiological measures gathered within the 1st year of life, on the one hand, and individual differences in language development, on the other (e.g., see the collection in Colombo, McCardle, & Freund, 2008;. The study of individual variation in infancy is far from new. For example, habituation and dishabituation measures have been used to assess the fit of different models of the architecture of cognitive skills (e.g., Bornstein, 1998) and to predict language, IQ, and educational outcomes longitudinally, even after several other variables have been controlled for (e.g., Bornstein et al., 2006; see also the Discussion section). Here, we discuss the potential of infant speech processing measures to predict variation in language outcomes over this backdrop of the larger field of cognitive development.Evaluating this emergent strand of research is important for two reasons. First, from a theoretical perspective, focusing on individual differences in speech perception may provide crucial insights into the relative merit of competing theoretical models. For instance, there are two large classes of theories bearing on the emergence of phonology in the 1st year of life (see review in R€ as€ anen, 2012). According to one, infants begin to learn phonology by determining which speech sound categories (which we call phones) their ambient language uses, and they later employ these units to build word formsWe are grateful to our colleagues at the Laboratoire de Sciences Cognitives et Psycholinguistique, and at both the Psychology of Language and the Genetics of Language Departments at the MPI for insightful discussions. We thank particularly the reviewers for their extremely constructive comments. Some of this work was done wi...

Semantics guide infants’ vowel learning: Computational and experimental evidence

“…In these two studies infants' vocabulary knowledge was assessed immediately, but another study shows that audiovisual integration at 6 months is related to infants' vocabulary at 12 months (Altvater-Mackensen & Grossmann, 2015). A recent meta-analysis (Cristia, Seidl, Junge, Soderstrom, & Hagoort, 2014) summarizes ample studies relating infants' measures on linguistic tasks with their future vocabulary development. This meta-analysis shows that infant performance on tasks tapping three levels of auditory-only speech perception (phones, word forms and prosody) correlates positively with future vocabulary size.…”

Individual Differences in Language Acquisition and Processing

“…Both the ability to make phonetic distinctions and segment words from running speech develop within the 1st year of life, but are subject to significant interindividual variation. Notably, longitudinal studies show that children who master these skills early have better vocabulary knowledge months or even years later compared with children who master them at an older age [34][35][36].…”

Can infants learn phonology in the lab? A meta-analytic answer

“…Despite the fact that experimental developmental psychology methods are relatively stable within a strand of research, the meta-analyses presented here, and those discussed elsewhere Cristia et al, 2014;Tsuji & Cristia, 2014) reveal above all the variability in results found across papers. This variability may in part be due to the fact that either direction of results is acceptable, both in peer-reviewed journals and other venues.…”

“…• the correlation in individuals' looking times across the two trial types within each experiment Additionally, I coded papers on a number of other dimensions that have been found to be relevant in previous meta-analyses of infant laboratory results Cristia, Seidl, Junge, Soderstrom, & Hagoort, 2014;Tsuji & Cristia, 2014), and which would be useful in a meta-meta-analysis (Lewis et al, 2015). A complete explanation for all fields can be found in https://osf.io/3hdk2/.…”