Abstract
We present the extended Mental States Framework (EMSF) for artificial cognitive agents able to include a static and a dynamic definition of mental states at large. These (individual and collective) states are defined in terms of basic components (attributes, laws and controls) organized on a table. Every state is viewed according to two perspectives, the declarative one, containing definitions (attributes, laws), and the prescriptive one having associated control mechanisms for triggering, filtering and selecting. At the same time we introduce a theory of artificial minds, where behaviour is based on the interplay between the space of mental states and the space of agent architectures. Every social action (and also of an individual agent) emerges as the behaviour generated by the causal relations among the corresponding (individual and collective) mental states of that agent, and also by the dynamics of their interactions. The ongoing research around EMSF crosses several disciplines and has been supported by listening and observation (via Clinical Psychology) and by the practice of extensive experimentation (Agent Based Design). Recently, progress was made around the role of attributes and the side-effects we can envisage within the context of Agent Software Engineering.
Get full access to this article
View all access options for this article.