The most crucial requirement for the study was to identify several congenitally blind children who were normally developing and who had no concomitant handicaps. This was imperative since the purpose of the investigation is not to describe blind children's language behavior but to examine the possible roles of vision in language development more generally and to try and understand better the impact of certain cognitive and social factors on language development.

The principal data base is media-recorded samples of spontaneous language produced in the familiar home environment. These are supplemented by parental records and individualized experiments. There are four major reasons for using this technique. First, experimental paradigms for gathering language data from infants and toddlers are not generally feasible. Secondly, we do not have adequate information about how blind children can be expected to respond in natural, much less experimental, situations. Thirdly, most of our understanding of early language acquisition is drawn from naturalistic studies of sighted children and this will provide an important basis for comparison. Fourthly, individualized experiments that are tailored to an individual child's vocabulary and experience maximize the possibility that a young child will comprehend and respond to experimental probes. These probes then are an important supplement to the spontaneous data that can illuminate findings and justify drawing certain conclusions.

This chapter describes each of the six subjects in some detail and then discusses the general methodology for data collection. Specific coding systems and analyses are described in the chapters presenting results.