Abstract
During 50-min sessions, 6 male human subjects could press either Button A or Button B available as nonreversible options. Button A presses were nonaggressive responses and earned points according to a fixed-ratio 100 schedule. Prior to the experiment subjects were instructed that every 10 (fixed-ratio 10) Button B presses (aggressive responses) subtracted a point from a fictitious 2nd subject. A random-time schedule of point loss was used to engender aggressive responding. The instructions attributed these point losses to the Button B presses of the subject's fictitious partner. Aggressive responding either escaped or avoided point loss by initiating an interval free of point loss. The duration of the interval was varied systematically across sessions. Avoidance contingencies maintained a high rate of aggressive responding over 30 sessions in the absence of point loss. Escape contingencies also maintained aggressive responding across sessions, with rates of aggressive responding corresponding to rates of point loss.