Document Type



Human postures are handled by our natural language systems as if they were discrete entities: people “stand,” “sit,” “kneel,” “lie,” etc. Postural differences are treated as if they were varieties within types, rather than gradations between clusters. A sample of human postures was collected to test the adequacy of the discrete, “language” model. If variations intergrade a field model would be more apt. If variations do not intergrade the discrete model would be adequate. A sample of 590 poses showed that a field model has the better fit.