with regard to operant conditioning, a process wherein the probability that a reaction will be supported relies on the reaction exceeding a value based upon a distribution from former reactions. Generally, the distribution is based upon some group of the most recent reactions and it is brought up to date with each reaction