Abstract: A robotic head for modeling the articulation of speech sounds, including a three-dimensional head section, a moveable tongue, upper and lower lips, and jaw, and at least one actuator for moving them. A memory stores a plurality of motion command sets, each being a predetermined set of commands for moving one or more of the tongue, lips, and jaw in a manner that corresponds to generation of a specific speech sound. In response to receipt of a speech sound input command that identifies a requested speech sound, a processor transmits the commands of the motion command set corresponding to the requested speech sound to the at least one actuator to move the tongue, lips, and jaw in a manner that corresponds to generation of the requested speech sound. The three-dimensional head is made of transparent material such that the tongue is viewable by a patient when moved.