DISCUSSION: Dr. Theodore Shapiro noted that Dr. Rubinstein avoided the problems inherent in predicting particular events by requiring only the confirmation of representatives of a class of events, which makes the rules of confirmation less stringent. Analysts, however, must limit the universe of acceptable derivatives if confirmation is to depend upon the usual laws of probability. He suggested that what Dr. Rubinstein considered causal relationship might better be viewed as empirical entailments. With regard to postdiction, Dr. Shapiro felt that analysts have a reconstructive bias which leads to the assumption that a past event has caused a present derivative. But the data may reflect contingent simultaneous