Learner-controlled instruction (LCI provides a model for developing the self-directed learners modern society requires. LCI allows the student to specify goals, to control significant resources to attain them, and to choose learning strategies, thus enabling him to learn how to learn. It appears that the environment needed to support this model must rely heavily upon computer-assisted instruction (CAI) and computer-managed instruction (CMI). Time-shared terminals and sophisticated hardware are required, and it is necessary to reorganize content fields for learner-controlled manipulations. Additionally, research needs to identify those who are unlikely to succeed in LCI, to discover how to remedy this problem, and to determine how to structure CAI so that it equalizes the achievements of poorer students without penalizing the better ones, rather than merely amplifying the advantages of the gifted. The TICCIT system at Brigham Young University is being used to field test an LCI program in which learners control the pace, sequence, and mode of instruction, specify the depth and detail of instructional interaction, and determine access to support facilities and advice. Evaluation after two years will determine the success of the system in making operational the theoretical model. (LB)