TWENTY HEAD START PRESCHOOL CHILDREN WERE GIVEN THREE LANGUAGE TASKS DESIGNED TO MEASURE THEIR LANGUAGE DEVELOPMENT--(1) A PRODUCTION TASK REQUIRED THE CHILDREN TO ENGAGE IN FREE SPEECH. THE CHILDREN WERE ASKED TO ANSWER A QUESTION, TO DESCRIBE A SERIES OF PICTURES, AND TO RETELL A STORY. THE FREE SPEECH OF THE CHILDREN IN RESPONSE TO THESE SUB-TASKS WAS ANALYZED IN TERMS OF THE RANGE AND FREQUENCY OF SYNTACTIC STRUCTURES USED, THE NUMBER OF KERNEL AND TRANSFORMED SENTENCES USED, AND THE COMPLEXITY OF THE SYNTAX. THE DATA FROM THIS TASK IS NOT YET AVAILABLE. (2) AN IMITATION TASK REQUIRED THE CHILDREN TO REPEAT 20 SENTENCES USING 10 DIFFERENT SYNTACTIC STRUCTURES. EACH STRUCTURE WAS USED IN TWO SENTENCES. OF A POSSIBLE 400 CORRECT RESPONSES, 281 CORRECT RESPONSES WERE MADE. THE LENGTH OF THE SENTENCE WAS FOUND TO BE NEGATIVELY RELATED TO NUMBER OF CORRECT RESPONSES FOR ONE SET OF THE STRUCTURES. THERE WAS EVIDENCE OF MODIFICATION OF SENTENCES BY THE CHILDREN TO CONFORM THEM TO THEIR OWN LINGUISTIC SYSTEM. (3) THE SAME SENTENCES USED IN TASK NUMBER 2 WERE USED IN A COMPREHENSION TASK THAT REQUIRED THE CHILDREN, AFTER HEARING A SENTENCE, TO POINT TO THE PICTURE CORRESPONDING TO THAT SENTENCE FROM A THREE-PICTURE DISPLAY. CORRECT RESPONSES NUMBERED 266 OUT OF A POSSIBLE 400. LENGTH OF SENTENCE WAS NOT FOUND TO BE RELATED SIGNIFICANTLY TO THIS TASK. A RECOGNIZED PROBLEM IN INTERPRETING THE DATA FROM THESE TASKS IS CHOOSING CRITERIA THAT WILL RELIABLY INDICATE WHEN THE CHILD REALLY HAS CONTROL OF A PARTICULAR SYNTACTIC STRUCTURE. TOTAL CORRECT-RESPONSES, AS A CRITERION, DOES NOT INDICATE WHETHER THE CHILD CONSISTENTLY RESPONDED CORRECTLY ON A PARTICULAR SYNTACTIC STRUCTURE IN TASKS NUMBER 2 AND NUMBER 3. IF THE CONSISTENCY OF CORRECT RESPONSE CRITERION WERE USED ON THE DATA, THE TOTAL SCORE OF THE CHILDREN WOULD BE 64. (WD)