|
North American Congress on Biomechanics Canadian Society for Biomechanics - American Society of Biomechanics University of Waterloo Waterloo, Ontario, Canada August 14-18, 1998 |
|
The success of a dance performance depends on the ability of the dancer to communicate an idea or emotion to the audience. The dancer skillfully manipulates his/her movements to express a specific emotional or aesthetic intent which can be recognized by the audience. Although skilled dancers have an implicit understanding of the grammar of expressive intent, the physical language of expression has yet to be formally articulated. Knowledge of the quantitative relationship between movement and expression is important not only for better understanding of human interactions, but also for developing realistic simulations of human movement for scientific or entertainment purposes.
In this study, a qualitative dynamics approach was taken to examine expressive movement. We quantified body motions to identify invariant features associated with different emotional intents in a dance gesture.
Previous research using point light techniques has shown that subjects can distinguish the sex, age, traits, and emotional state of individuals by observing the motion of just a few points on their bodies (Brownlow et al., 1997; Montepare & Zebrowitz-McArthur, 1988). Among seven different emotions, anger and fear were most reliably identified by observers (Sogon & Izard, 1987). These analyses were based on subjective observations; the quantitative aspects of the motions associated with the emotional responses are not known.
Qualitative dynamics can suggest characteristic types of causal processes in movement. This method has been used to illustrate characteristic pathologies of hemiparetic gait (Winstein & Garfinkel, 1989). In dance, performers may employ similar strategies to express the same emotion in a gesture; presumably, these strategies can be detected as similar characteristics in the phase-plane portraits of their movements. Potentially, topological study of phase-plane portraits can be used to distinguish among different expressive intents.
Fifteen female students majoring in dance participated in the study. The subjects were asked to perform a whole body gesture under three different expressive conditions: (1) neutral, (2) fear, and (3) anger. The task began from a standing position with feet apart and parallel, and consisted of flexion of the knees, hips, and shoulders into a demi plie position, followed by extension back to the starting position. Subjects first performed two trials of the neutral task and then two trials each of the emotion tasks; the order of the emotion trials was varied. The subjects rated their own best performances; these self-selected best trials were used in the analysis. Reflective markers were placed over twelve body landmarks. Three video cameras (60 Hz) were used to capture joint motion. Kinematic data were filtered using a Butterworth filter with a 5 Hz cutoff frequency for the wrist and hand and 4 Hz for all other joints.
Preliminary results for the ankle, lower spine, and shoulder are reported here for five of the fifteen subjects. Subjects were similar in age (21.4±3.4 yrs) and height (1.7±0.09 m).
The phase-plane plots indicated that ankle kinematics changed with emotional intent (Fig. 1). The phase-plane plot for the neutral gesture was rectangular. Vertical regions indicated change of direction from dorsiflexion to plantarflexion, and horizontal regions indicated constant velocity. The anger and fear portraits were not rectangular, but had rounded regions indicating velocity change during plantarflexion and dorsiflexion. The fear portraits showed constant velocity during plantarflexion while the anger portraits showed rounded regions of changing velocity.
Figure 1(Neutral, Fear, Anger). Representative ankle phase-plane plots for one subject. Movements began with dorsiflexion.
Like the ankle, the topography of shoulder and lower spine phase-plane portraits clearly differed between the anger and fear and the neutral conditions. At these joints, however, no distinctive pattern emerged to discriminate between anger and fear.
Quantitative data indicated some differences between emotional intents (Table 1). Range of motion for the lower spine increased with anger and fear. Peak velocities differed between emotions, with the lowest velocities in neutral movements and the highest velocities in anger.
| A | Neutral | Fear | Anger |
|---|---|---|---|
| ROM | 25.2±0.2 | 24.7±7.2 | 26.2±4.9 |
| (+) Vel | 36.9±13.1 | 36.7±21.3 | 73.5±31.4 |
| (-) Vel | 25.6±7.3 | 60.7±8.8 | 59.6±34.4 |
| S | |||
| ROM | 54.4±12.4 | 46.6±26.4 | 66.3±17.8 |
| (+) Vel | 48.8±23.6 | 124.6±157.3 | 250.3±167.0 |
| (-) Vel | 45.0±9.8 | 101.6±86.9 | 334.9±155.9 |
| LS | |||
| ROM | 5.0±2.5 | 19.2±11.2 | 25.5±17.9 |
| (+) Vel | 18.7±3.0 | 30.3±18.3 | 85.3±59.2 |
| (-) Vel | 13.9±3.1 | 42.7±28.9 | 82.2±78.9 |
Table 1: Range of motion (ROM) and peak velocities for ankle (A), shoulder (S) and lower spine (LS).
In general, movements with emotional intent differed from neutral. Movements performed with anger and fear also differed, especially at the ankle, but the differences were less marked. These data may indicate common strategies used by dancers to communicate specific emotions in their movements. Although the data are preliminary, the study represents a first step towards quantifying different expressive intents in human movement.
Brownlow S. et al. Psych. Rec., 47, 411-421, 1997.
Montepare J.M. and Zebrowitz-McArthur L. J. Personality Soc. Psych., 55, 547-556, 1988
Sogon S. and Izard C.E. Jap. Psych Res., 29, 89-93, 1987.
Winstein C.J. and Garfinkel A. J. Motor Behavior, 21, 373-391, 1989.
