ASA/CAA '05 Meeting, Vancouver, BC

The "Forbidden Experiment" in Language Development

Mario A. Svirsky - msvirsky@iupui.edu
Dept. of Otolaryngology-HNS, Indiana University School of Medicine
Depts. of Biomedical and Electrical Engineering, Purdue University

Rachael Frush Holt
Dept. of Otolaryngology-HNS
Indiana Univ. School of Medicine

Popular version of paper 1aSC1
Presented Monday morning, May 16, 2005
Joint ASA/CAA Meeting, Vancouver, BC

It is generally accepted that there is a "sensitive period" for language development in the first few years of life when the ability to develop language and learn its rules is optimal. According to this hypothesis, the innate human ability to learn language diminishes without early language exposure. However, the ultimate experiment that would establish the existence and the characteristics of sensitive periods is one that cannot be done. The "forbidden experiment" would involve withdrawing all language input from children, and then measuring their ability to develop language after different periods of deprivation.

This, of course, would be highly unethical.

Language is a uniquely human ability, so we do not have appropriate animal models to study the effect of early deprivation. Given that the "forbidden experiment" cannot be done, how can we learn about sensitive periods for language development? One possibility is to examine several sources of converging but indirect evidence. One such source of evidence comes from the tragic cases of feral or isolated children. The most famous case may be that of Genie, the child who was kept isolated by her parents, locked in a room and tied to a potty chair most of the time until she was discovered by authorities at the age of thirteen. Genie was able to learn some words but she never came close to developing normal language and was able to handle only very simple grammatical constructions. However, the evidence in support of biologically determined sensitive periods from the cases of Genie and other children like her is not definitive because their difficulties may be due to other cognitive deficits or to deprivation of experiences other than language.

In this study we provide a new source of converging evidence that is relevant for the investigation of sensitive periods in language development. We studied the speech perception and language skills of children who were born profoundly deaf and whose hearing was partially restored with cochlear implants after a period of sound deprivation, ranging from six months to four years. The sound deprivation period extended from at least the time it took to determine that the children were indeed deaf to the point when the parents decided that they wanted to proceed with surgery. Thus, the examination of these children's
speech and language skills represented an imperfect version of the forbidden experiment.

Results obtained using a standard test of language development were consistent with the existence of a sensitive period that starts within the first four years of life. Improvement in language skills was faster for children who received auditory input from a cochlear implant early in life than for those who received it later. In contrast, a simple test of speech perception showed that children's performance improved at about the same rate after implantation, regardless of whether they received cochlear implants in the first, second, third, or fourth years of life. A sensitive period for speech perception may well exist, but
it may start later than age four.