How do logical inference rules help construct social mental models?

Starting from recent approaches in mental model research, it is argued that (1) logical inference rules are used in order to construct mental cliques from learned sentiment relations, and (2) social context cues (operationalized as primes) play a crucial role in activating such rules. Transitivity and antitransitivity are taken as examples, and are shown as core constituents of such models. In a first experiment, priming was achieved by announcing the sorting of fictitious persons in eithertwoorthreecliques. Thirty-one subjects studied eight sets of sentiment relations among these persons that either did or did not satisfy their primed clique expectations. They showed longer study times and more requests for additional information in the case of inconsistent fits between prime and set. Their sorting solutions also showed clear priming effects. A second experiment (n = 30) showed that when undergoing a recognition test after seeing the relation sets, subjects tended to confuse model-consistent distractors with information they had actually seen. In a third experiment (n = 30) the results from Experiment 1 were replicated using more realistic learning materials