Full knowledge of any specific-language grammar is not biologically programmed; the theory of Universal Grammar does not propose that it is. The input of specific language data must interact with whatever biological programming exists within children.

In this chapter we first review empirical evidence that experience is necessary, but suggest that widely varying forms of experience allow language acquisition (6.1). We then review evidence for the infant's remarkably rapid knowledge about first language input (6.2). We consider evidence for the claim that a specialized form of input is provided to infants as a “Baby Talk Register” (6.3–6.4), then evidence regarding the nature of children's relation to the input to which they are exposed (6.4–6.5), concluding (6.6) that children's use of this data is indirect, selective and reconstructive. The structure of the organism – the child mind – continually determines and mediates (a) what data are used and (b) how they are used by children in language acquisition.

Is experience necessary?

The “royal” experiments

What would happen if children were given no experience of a specific language? Herodotus describes an experiment conducted by two kings who isolated two infants from all language input to discover which language, Phrygian or Egyptian, would emerge as “the first of all languages on earth” (Feldman, Goldin-Meadow and Gleitman 1978, 354). Although the royal experiments cannot, thankfully, be conducted anymore, several alternative forms of scientific evidence exist now regarding the role of experience in language acquisition and its precise nature.