Unified theories of cognition

Unified theories of cognition (UTCs) offer an alternative to the modal ‘divide and conquer’ methodology within cognitive science and attempt to address the full range of cognitive activity within a single theoretical framework. These theories, also termed ‘cognitive architectures’ are generally computational in nature and are intended to model, at some degree of fidelity, human cognition in a broad range of tasks. This style of research has numerous advantages, not the least of which being that the actual human cognitive system is itself an integrated system and many important tasks require bringing integrated capabilities to bear. There are also drawbacks, particularly dealing with the incompleteness of the knowledge base in cognitive science and the difficulty of evaluating such theories. Three architectures are profiled, each one representing a different ‘home’ discipline: from AI, Soar; from cognitive psychology, Adaptive Control of Thought‐Rational; and from neuroscience, Leabra. Future directions for UTCs include expansion into branches of cognition not already well represented, such as spatial cognition, and increasing attention to cognitive moderators such as emotion and fatigue. Overall, this is a powerful research strategy that is likely to remain an important part of cognitive science for the foreseeable future. WIREs Cogn Sci 2012, 3:431–438. doi: 10.1002/wcs.1180 This article is categorized under: Computer Science > Neural Networks