Representational Content and the Reciprocal Interplay of Agent and Environment

Abstract

Declarative modelling approaches in principle assume a notion of representation or representational content for the modelling concepts. The notion of representational content as discussed in literature in cognitive science and philosophy of mind shows complications as soon as agent and environment have an intense reciprocal interaction. In such cases an internal agent state is affected by the way in which internal and external aspects are interwoven during (ongoing) interaction. In this paper it is shown that the classical correlational approach to representational content is not applicable, but the temporalinteractivist approach is. As this approach involves more complex temporal relationships, formalisation was used to define specifications of the representational content more precisely. These specifications have been validated by automatically checking them on traces generated by a simulation model. Moreover, by mathematical proof it was shown how these specifications are entailed by the basic local properties.

Formalisation of Damasio’s theory of emotion, feeling and core consciousness

“…Also this may be analysed using similar methods. Two papers reporting other analyses of more complex cases of representational content are (Bosse, Jonker and Treur, 2005a) for representational content for a case of intensive reciprocal agent-environment interaction, and (Bosse, Jonker, Schut, and Treur, 2006) for collective represent ational content for a case of a society of agents.…”

“…The analysis approach that is applied in this paper to model Damasio's theory of consciousness, has previously been applied to complex and dynamic cognitive processes other than consciousness, such as the interaction between agent and environment (Bosse, Jonker, and Treur, 2005a). In a number of these cases, in addition to simulated traces, also empirical (human) traces have been formally analysed.…”

Specification and Verification of Dynamics in Agent Models

“…The main research goal in these projects was to analyse the behavioural dynamics of the agents involved (e.g., [6,8,9,10,11,12]). TTL was used to formalise dynamic properties of these processes at a high level of abstraction.…”

“…As a first example, from a paper about representational content for the mental state of an agent that intensively interacts with the environment [10] is discussed. This case study involves the processes to unlock a front door that sticks.…”

Agent-based simulation of episodic criminal behaviour1

“…In fact, the properties IP1a and IP1b express that having an episode has a backward representation relation (within duration e) to meeting a negative agent, and IP2a and IP2b that it has a forward representation relation to conditionally performing a criminal action as soon as (within duration e) an opportunity occurs; cf. [10]. The properties IP1a, IP1b, IP2a and IP2b can be refined further into more local properties describing the criminal agent's internal mechanisms.…”