A simple quantitative model for speech development is proposed. The model is developed from observations of two systems, one that of normal hearing children and the other that of those who suffer from congenital profound deafness. Comparison of the two systems provides clues to the type of controls needed to produce normal speech. The possible nonlinear controls which may be used during the development of both the suprasegmental and segmental aspects of speech are identified. The linguistic components of speech are ignored. The interdependence of perceptual systems during developmental phases is taken into account and the importance of the associative cortex in speech-motor control is formulated. Deaf persons may employ the tactile system in developing speech; the distinctive unit of code is called a tacheme.