Operant conditioning techniques were used to modify verbal behavior in 18 institutionalized, trainable mentally handicapped mongoloids with chronological ages from 4-6 to 7-10 and mental ages from 2.0 to 2.10. Two instruments were constructed to evaluate language: a language test and a speech rating scale. Project leaders had no prior knowledge of the language test. The six children in group 1 received five 15- to 40-minute group conditioning sessions daily. Shaping was used to condition object and picture naming, descriptions of action pictures in work combinations, and discrimination of colors and polar opposites. Operant procedures were used to lengthen attention span and increase verbal production. The six children in group 2 spent 7 weeks in the enriched experimental environment, but received no conditioning; the six in group 3 remained in the institution. Five of six members of group 1 significantly increased their scores on the Stanfore-Binet posttest; one from group 2 and none from group 3 increased significantly. On the speech rating scale, group 1 showed significantly greater improvement than group 2. Major changes were observed in the social behavior of the 12 subjects who resided in the research facility. (Author/RP)