Semantic ambiguity and memory.

• A normed measure of the ambiguity of words' emotional valence has been developed. • With that measure, emotional ambiguity displays characteristic empirical effects. • We developed a similar normed measure of the ambiguity of other semantic attributes. • The ambiguities of other attributes displayed the same empirical effects as emotional ambiguity. • People process the ambiguity as well as the intensity of semantic attributes. The emotional ambiguity hypothesis posits that as items are encoded, people process the ambiguity as well as the intensity of their valence. The hypothesis predicts three signature effects, all of which have been reported: ambiguity-driven declines in valence-arousal correlations, a quadratic law relating perceived valence to valence ambiguity, and improvements in episodic memory as a function of increases in valence ambiguity. After reviewing evidence on these effects, we evaluated fuzzy-trace theory's proposal that the ambiguity hypothesis should apply to a broad range of semantic attributes other than valence (e.g., categorization, concreteness, meaningfulness). According to that proposal, all three effects should be observed for other attributes. They are. In Experiment 1, ambiguity-driven reductions in attribute correlations were identified for seven other attributes. In Experiment 2, 16 other attributes obeyed the quadratic law. In Experiments 3–5, three other attributes displayed ambiguity-driven improvements in recall. It appears that people process the ambiguity as well as the intensity of many semantic attributes, and hence, the memory effects of such attributes can be due to either or both. [ABSTRACT FROM AUTHOR]