Independence and correlation for coherent lower conditional probabilities

Coherent lower conditional probability is the main tool in decision models representing the ambiguity aversion of the decision maker, such as in the pessimistic synthesis of probabilistic judgments, under different (also unexpected) scenarios. Aim of this paper is to put under the right perspective the concepts of independence and correlation in the framework of coherent lower conditional probability, taking suitably into account also events whose lower probability is zero or one.