Abstract
Temporal aspects of human movements are used to categorize affect (Paterson et al, ECVP 99), however, to what extent are such categorizations contingent upon form cues? 3-dimensional movement data were collected from 6 points on the arm and head of 2 actors as they performed knocking movements with 10 affects. Movements were phase-scrambled so that relationships among points would be disrupted while the motion of individual points remained the same. Data for each movement was also processed to obtain its kinematic properties. The normal and scrambled movements were each presented as point light displays in a forced choice recognition task. For phase scrambled movements recognition accuracy was worse than for normal movements, but still above chance. The averaged confusion matrices for each type of movement were converted to measures of dissimilarity and input to a multidimensional scaling algorithm. The first dimension from the resultant 2-dimensional psychological spaces correlated both with movement speed and each other. The second dimensions did not. That the first dimension could be accounted for by speed in both normal and scrambled displays, indicates that speed cues can be sufficient to distinguish aspects of affect from biological motion, regardless of form. However, since about half of the observers classified the scrambled movements as some kind of animate motion, it is possible that the use of kinematics for perception of affect is contingent upon the attribution of animacy to a display.