The Role of Emotion in an Architecture of Mind

The question for many designers of intelligent systems is no longer whether to incorporate emotion and motivation, but how to do so. This question has strong consequences for the autonomy and flexibility of a system and its ability to respond to novel, unpredictable or changing environments – in other words, the degree of intelligence it will display. Autonomous agents that learn through situated embodiment in a rich and changing environment may require a very different approach to emotion and motivation than systems in which procedures are written to cover all contingencies and choices are made among known alternatives, albeit using a sophisticated decision process. Here, we dispute Sloman’s suggestion that emotion is an emergent control state and thus not a component of a system architecture. We also propose ways in which an emotion mechanism can enhance intelligence in autonomous agents.