Attitude Representation: Attitudes as Patterns in a Distributed, Connectionist Representational System

How we think about the representation of attitudes has a profound impact on how we think about attitudes themselves, attitude change, and the attitude–behavior relationship. In this article, we briefly review the model of attribute representation in a distributed, connectionist memory system, which portrays attitudes as time–dependent states of the system rather than as static “things” that are “stored” in memory. This model is particularly well–suited to addressing some of the field's most pressing questions about the multiplicity of attitudes and their stability (or instability) over time. We address several of these questions from the distributed, connectionist perspective, concluding that the new model renders some questions meaningless, suggests straightforward answers to others, and hints at exciting new hypotheses about the answers to still others.