Abstract
This investigation, initiated on the basis of Jakobson's hypothesis, that is, his theory of a two-stage pattern of phonological development, was dedicated to the purpose of assisting a primary school girl with severe mental retardation to improve further in both phonological and semantic development. Successive sound patterns that she produced were divided into two topological features, that is, the first one was the tendency to adopt a four-rhythm-oriented strategy (i. e., "four-mora" illustrated in this study) toward the rhythmic property of the Japanese language, while another tendency of such a sort of incontiguous dissimilating as seemed strikingly to inhibit the cooccurrence of the same syllabic "i" went hand-in-hand with substitutions of other consonantal and vocalic front or back. For instance, it changed from [i〓〓: 〓i] to [i〓i: ro], but not to [i〓i: ri]. When phonological development as the logarithmic ratio of Type to Token in speech sounds, consonants, and vowels was compared, the correspondence to the developmental sequences between the subject and an infant of about one year of age without mental retardation was even greater. The findings support the hypothesis that, in a teaching program for language acquisition, a child with severe mental retardation would develop phonological and semantic utterance skills from consonantal back to front, and vocalic from front to back.
