In 1974, when the Journal of Child Language, was founded, the claim that quantity or quality of linguistic input might be relevant to the course of language acquisition was highly controversial. For some, in fact, it was an absurdity to suggest such a thing. There had accumulated by the mid 1970s a modest body of work documenting that speech addressed to young children was generally grammatically simple and lexically redundant (Remick, Reference Remick1971; Snow, Reference Snow1972, Reference Kurland and Snow1977; papers published in Snow & Ferguson, Reference Snow and Ferguson1977), but no one had actually demonstrated that these adaptations made any difference. Furthermore, the much richer body of work documenting that speech to young children often included interactive features seemingly designed to support language development (repetitions, expansions, extensions, attempts at clarification, and so on; see, for example, Brown & Bellugi, Reference Brown and Bellugi1964) could similarly be relegated to description of a register or a parenting style, rather than being accepted as a claim about processes important to language acquisition.

Researchers who were inclined to believe in the impact of features of the language environment on children's language development adopted a pragmatic perspective, much influenced by Dell Hymes’ (Reference Hymes, Pride and Holmes1972) notion of communicative competence and by the ‘language as action’ perspectives of Austin (Reference Austin1962) and Searle (Reference Searle1969). An early example of this line of thought in the analysis of adult speech was the emphasis on the importance of semantic contingency in speech addressed to children (Cross, Reference Cross, Snow and Ferguson1977) – the claim that adult speech was effective input because it was semantically related to the prior child utterance or to the child gesture, action, or gaze. From the child language side of things, the pragmatic approach dictated the analysis of children's earliest semi-verbal utterances as legitimate attempts at communication (e.g. Carter, Reference Carter1975). That line of work led ultimately to a reconceptualization of the innate capacities for language acquisition as centered around social and pragmatic capacities rather than being syntactic structures or specifications (Ninio & Snow, Reference Ninio and Snow1996).

For about twenty years, then, child language researchers of the social-interactionist persuasion continued developing ever more sophisticated notions about how language input and interaction might relate to language development. Nonetheless, solid and convincing evidence that these features did relate to language outcomes remained elusive. Of course, no such evidence could be expected to emerge from the relatively small-scale small-n studies that characterized the child language field in its infancy. Recurrent findings from larger-scale studies of social class differences in amount of input and in child language outcomes (e.g. Snow, Arlman-Rupp, Hassing, Jobse, Joosten & Vorster, Reference Snow, Arlman-Rupp, Hassing, Jobse, Joosten and Vorster1976) could easily be dismissed because of the impossibility of distinguishing environmental from genetic and other biological influences. The generally disappointing impacts on language outcomes obtained from interventions such as Head Start reinforced the notion that social interaction could not be causally related to speed or ease of language development.

A shift in the zeitgeist occurred with the appearance in 1995 of the book by Hart and Risley entitled Meaningful Differences in the Everyday Experiences of Young American Children. Hart and Risley showed strong correlations between a very simple outcome measure (child cumulative vocabulary) and a very simple feature of the input: quantity of words heard. They commented as well on the presence of interactive language features and support for language in the homes of children with larger vocabularies, and the absence in those homes of occasions when child utterances were ignored or of punitive responses to their talk. Nonetheless, the simple message about the relationship between amount of input and amount of learning struck a chord. Hart and Risley turned the attention of many to questions about parent–child interaction in low-income families (e.g. Pan, Rowe, Singer & Snow, Reference Pan, Rowe, Singer and Snow2005; Rowe, Pan & Ayoub, Reference Rowe, Pan and Ayoub2005). The study had an immediate, strong impact in the field of early childhood education, with citations from researchers working in the field of language acquisition proper initially lagging behind.

Twenty years after the publication of the Hart and Risley book, the implications of its findings for our understanding of child language development are almost universally acknowledged. Hart and Risley's basic claim about quantity of input has been replicated and expanded upon through programs of research by Hoff (Reference Hoff2003) and by the work done at the University of Chicago under the direction of Huttenlocher and Goldin-Meadow (e.g. Rowe, Reference Rowe2008, Reference Rowe2012; Rowe, Raudenbush & Goldin-Meadow, Reference Rowe, Raudenbush and Goldin-Meadow2012), among others. Furthermore, the Hart and Risley work is cited as the direct impetus for a wide array of vocabulary interventions designed as preventatives for preschool-aged children (Leffel & Suskind, Reference Leffel and Suskind2013; Sanders, Reference Sanders1999; The Thirty Million Word Project, online: <tmw.org>) and as remediations for children in the primary grades (e.g. Beck & McKeown, Reference Beck and McKeown2007; Biemiller & Boote, Reference Biemiller and Boote2006; Coyne, McCoach, Loftus, Zipoli & Kapp, Reference Coyne, McCoach, Loftus, Zipoli and Kapp2009; Silverman, Reference Silverman2007).

Importantly, claims about the importance of input to language learners have been expanded beyond vocabulary. Huttenlocher, Vasilyeva, Cymerman, and Levine (Reference Huttenlocher, Vasilyeva, Cymerman and Levine2002) showed relationships of syntactic complexity in input to language comprehension. In experimental studies, Nelson and colleagues (Camarata, Nelson & Camarata, Reference Camarata, Nelson and Camarata1994; Nelson, Carskaddon & Bonvillian, Reference Nelson, Carskaddon and Bonvillian1973) showed effects of brief but intensive modeling of syntactic structures on children's comprehension and use of those same structures. Pine, Lieven, Theakston, and colleagues have shown effects of frequency and predictability of occurrence of closed-class elements on order and speed of acquisition of those elements (e.g. Lieven, Pine & Baldwin, Reference Lieven, Pine and Baldwin1997; Pine & Lieven, Reference Pine and Lieven1997; Theakston, Lieven & Tomasello, Reference Theakston, Lieven and Tomasello2003). In other words, though not yet universally accepted, the basic environmentalist/social-interactionist position on language acquisition has achieved respectability if not dominance. This position is, furthermore, the default justification for early childhood interventions such as home-visiting programs (e.g. the Nurse-Family Partnership: Olds, Holmberg, Donelan-McCall, Luckey, Knudtson & Robinson, Reference Olds, Holmberg, Donelan-McCall, Luckey, Knudtson and Robinson2014), and the basis for efforts to design high-quality early-childhood care settings (e.g. the Abecedarian Project: Campbell & Ramey, Reference Campbell and Ramey1994; Brookline Early Education Program: Hauser-Cram, Pierson, Klein Walker & Tivnan, Reference Hauser-Cram, Pierson, Klein Walker and Tivnan1991; Project Care: Wasik, Ramey, Bryant & Sparling, Reference Wasik, Ramey, Bryant and Sparling1990). So social interactionism has achieved unquestioned ascendance in early childhood practice even if not in language acquisition theory.