Some important empirical findings and theoretical positions that interest contemporary students of learning and memory are discussed in the context of Luria's (1979) memory coherence perspective. This approach assumes that an important challenge to memory function is remaining closed to the influence of associations that are extraneous to the demands of the task at hand. We argue that the ability to support configural associations between representations of the joint occurrence or conjunction of two or more stimulus elements and a target memory is an important feature of a closed system but that an open system can support only elemental associations with the target. Successful performance in many tasks used to study memory can be achieved by elemental associations, but other tasks require the formation of and retrieval by configural associations. From this perspective, the pattern of spared and impaired performance often seen in animals and people with brain damage to the hippocampal system or in amnesic people of various etiologies results because their memory systems are open and can support only elemental associations.