This master's thesis begins by discussing the differences between 3-D computer animation of solid three-dimensional, or monolithic, objects, and the animation of characters, i.e., collections of movable parts with soft pliable surfaces. Principles from two-dimensional character animation that can be transferred to three-dimensional character animation are outlined and advantages of working with computer generated animation are discussed. The current means of animating characters at Ohio State University are illustrated by describing the process of creating three separate character animation pieces and the advantages and disadvantages of various techniques are outlined. Topics discussed include shape interpolation; Amorpha and the Fractalites animation; Twixt, a keyframe animation system; the runner animation; the uneven bars animation; Twixt-3, an event driven animation system; the skeleton animation system; the Snoot and Muttly animation; a future low level character animation system; and higher level character animation systems. It is concluded that if a new animation language is to improve the quality of animation, it must enable the animator to: (1) build the original character; (2) specify its hierarchy; (3) define the original motion; (4) easily change the motion; (5) create individualized parameterized models; (6) build the original surface description of a character; (7) manipulate the surface of its individual parts; (8) manipulate the surface as a whole; and (9) view the surface as needed during the animation process. Both a 19-item reference list and a 29-item bibliography are provided. (MES)