Avoidance of Token-Loss by Pigeons

Abstract

The response-dependent delivery and removal of tokens can reinforce and punish the responding of humans and nonhuman animals. The responding of humans can be maintained by the avoidance of token-losses, but this has not been demonstrated with nonhuman animals. Across three experiments, it was demonstrated that the responding of pigeons could be maintained in a token-loss avoidance arrangement. Tokens were delivered response-independently at the beginning of a token-loss avoidance period. Tokens were scheduled to be removed on a variablecycle schedule. Responding canceled scheduled token-losses. Those tokens that remained were exchangeable for food during programmed exchange periods. Several parameters of the tokenloss avoidance arrangement were manipulated to assess their effects on responding. In Experiment 1, responding was a function of the frequency of scheduled token-losses. In Experiment 2, responding was a function of the frequency of exchange periods. In Experiment 3, no consistent changes in responding occurred whether a single token produced multiple reinforcers or multiple tokens were required to produce a single reinforcer. Across all experiments, local response rates were an inverse function of the number of tokens present. The similarity of these results to those obtained in previous experiments, the possible functions of tokens in the token-loss avoidance arrangement, and the implications of the methods used in the present series of experiments are discussed.