Three groups of four preschool children who were profoundly deaf or had severe hearing impairment were given training to produce constant–vowel (CV) syllables. One group palpated a vibrotactile vocoder‐type display of the acoustic signals. One group used an analogous visual display. Children in the third (control) group were given reinforcement in the form of slide or puppet displays. In theinitialsessions, all three groups of children were trained to vocalize in response to a ’’light out’’ signal. No displays were used. The children were then trained to pay attention to the displays and to model CV syllable series presented by the experimenter. In thefinalsessions all three groups of children were given a model CV syllable series and encouraged to respond by vocalizing when given the microphone and a ’’light out’’ signal. At the same time they had to attend to the tactile, visual, or nonspeech display. The ratio of CV syllables to all syllables, (CV+V), was calculated for each child for the initial and for the final sessions, and the difference score obtained. The paired observationt‐test indicated a significant increase in the ratio for the experimental groups and no significant change for the control group. We conclude that the acoustically driven nonauditory displays were an essential factor in the observed change in CV production.Subject Classification: [43]70.75; [43]65.82, [43]65.64, [43]65.70.