Relationships between "combinatorial" abilities in language, symbolic play, blockbuilding, and non-semantic action sequences were explored in a study of 30 infants between 82 and 91 weeks of age. Subjects were observed in a laboratory playroom setting for approximately 45 minutes. During this time, a number of tasks were administered which were aimed at assessing the child's ability to combine his or her actions into sequences. For each of four target objects, two scenarios were modeled: one involved an appropriate exemplar of a target object (Doggie, Cup, Car, or Fiffin, a novel concept), and the other involved an inappropriate object which was substituted for the target object (e.g., a flashlight as a doggie). Additionally, the experimenter modeled meaningless actions and constructed four block structures. The children's utterances during the entire session were transcribed. Data on infants'"combinatorial" abilities were obtained in five domains: symbolic play with appropriate objects, symbolic play with inappropriate objects, non-semantic action sequences, blockbuilding, and language. Two independent transcripts were made of each child's utterances and play. Only those behaviors agreed upon by two raters were included in the analyses. Results are discussed. (Author/RH)