Four hearing Ss (20- to 30-years old) and 4 deaf Ss (8- to 14-years old) trained in speech discrimination using a vocoder (a device which converts speech into tactual vibrations received through the skin). Hearing Ss (artificially deafened by white noise transmitted through headphones) received from 20 to 80 hours of training in isolated words presented randomly, words in connected sentences, inflection copying, and rhyming, all of which was presented primarily without face-to-face contact with the speaker. Deaf Ss received training similar to that for hearing Ss with a few differences such as use of a reinforcement system and work on articulation. Results supported conclusions such as the following: that deaf Ss can be taught to hear five speech discriminations through the tactual mode; that hundreds of corrected repetitions are required for either a deaf or hearing S to learn simple tactual discriminations; and that the S's memory and ability to discriminate increases as the number of words he has mastered increases. (LS)