This publication reports on research which has investigated phoneme acquisition and carry-over in a programmatic fashion. This research began with investigations of the effect of an extension of stimulus control of new phoneme responses through a systematic operant procedure. After quantifying the effect of such procedures in terms of stimulus generalization, the program of research moved to the investigation of additional systematic procedures which were applied to such factors as phoneme position in words; phoneme responses embedded in sound-loaded sentences; and phoneme-responding in sentences under enforced time limits. Measurement of the effects of such training variables was made in terms of correct phoneme-responding in controlled samples of connected speech and in spontaneous conversation. The goal of the research has been to identify, empirically, the treatment variables which function to attain the carry-over of new phoneme learning into nonprogrammed conversational speech. The results of the research indicate that systematic procedures which treat certain stimulus and response variables result in high degrees of correctness in new-phoneme responses in conversational speech. These variables are: (1) antecedent stimulus variables ranging from imitative models to nonimitative evoking stimuli, such as pictures and intraverbal chains; and (2) response configuration variables, such as word-position of phonemes and sentence-production performance. (Author/DB)