A cybernetical perspective on emotion-cognition interactions as self-directed adaptation in situated decision-makers

Emotion-cognition interactions are often (Simon, 1967; Sloman, 2002) understood as adaptational mechanisms that help to cope bounded decision-makers with constraints and dangers forced upon by their environments. In this work, emotions are regarded as emergent properties of interconnected systems of functional routines - decision-making, memory access, etc. - and associated internal monitoring and modulation systems. Couplings between these mechanisms create internal bidirectional feedback loops that can sustain globally synchronized responses, which have been referred to as “appraisal-emotion”-amalgams (Lewis, 2005). Although emotions have been suggested as crucial components for enabling biological agents to cope with difficulties of decision-making in complex, partially unknown environments (Hanoch, 2002; Muramatsu & Hanoch, 2005; Sloman, 2011), research on artificial emotion models lacks common frameworks to explore design spaces for particular classes of emotionally-influenced systems (Hudlicka, 2008; Sloman, 2002), rendering it difficult to identify basic principles or architectural constraints (Hudlicka, 2008; Sloman, 1999, 2002). This article presents a systems-level framework for modeling emotion-cognition interaction in domain-independent decision-making systems based on Optimal Control Theory as a particular form of self-modulation driven by continual evaluation of the relationship between system and environment.