Synthetic agents: synthetic minds?

There are a plethora of agent definitions. These range from descriptions based on a functional analysis of how agents are used in technology to far more ranging expositions based on different interpretations of the role and objectives of artificial intelligence and cognitive science. It is possible to establish an ontology within which agents (and their applications) can be characterised, allowing agent definitions to be compared and providing an ontological framework within which the design requirements for synthetic agents, and by extension synthetic minds can be addressed. The contention here is that by developing sufficiently cogent models of (human) minds that are capable of acting as specifications for a synthetic mind, we cannot only address the strengths and shortcomings of those models (or theories) through the development of computational models but develop synthetic agents that could be said to exhibit qualities associated with having a (synthetic) mind. Irrespective of what dialectic we use to analyse the behavioural and cognitive qualities associated with a mind, there are a number of underlying questions that need to be addressed, including: what sort of computational architecture will enable this phenomenon? It is suggested that there is no one architecture and the rest of this paper considers a few alternatives. The discussion is based on experiments with computational agents that address questions related to the architecture, the range of control states, and the behavioural and cognitive capabilities associated with a mind