The problem of this study is to see if there is more symbolic language transfer by autistic children in an untrained matching situation if there has been previous training on similar matching situations. The purposes of the study are twofold. The first is to see if subjects will make a motor sign response untrained to an object if the word for the object has been trained to the motor sign response and to see if they will make a motor sign response untrained to a word for an object if they have been trained to make the response to the object named. The second is to see, if transfer occurs, whether the order presentations of object and word are a factor.