AN INVESTIGATION OF THE COMMUNICATION OF SHAPE by

of the requirements for the degree of Master of Science. This thesis is an examination, through verbal protocol analysis, of the one-way transmission of shape information. A communication model was developed, and the hypothesized attributes of a "good " description were experimentally tested. The methodology consisted of several steps: Twenty-eight subjects viewed nine objects from a range of relatively simple 3D shapes. The subjects verbally described each object. The transcribed descriptions were analyzed for information content and syntactic features. Selected object descriptions were then presented to a second group of thirty subjects for reconstruction (i.e., drawing). The descriptions were evaluated with respect to the objects and subsequent reconstructions, and the reconstructions were evaluated with respect to the objects and their descriptions. It was found that the accurate positioning and placement of object parts is the most important attribute of a successful description. The describers ' use of scale, surface texture and color was infrequent. The majority of descriptions used static description techniques. Overall, it was found that the closer the descriptions were to the hypothesized attributes, the more likely they were to result in accurate and recognizable reconstructions