(FNRS 11-39551.93/11-049629.96)

Notwithstanding the increase of empirical studies of emotion, it remains difficult to transfer the results found in the laboratory to spontaneous real emotional experiences; experiences that are mainly evoked in dynamically changing interactive contexts. At least in part, this difficulty is caused by the artificial separation of intrinsically related psychological processes - like emotion, cognition, and behavior - in different research disciplines. A further problem is that current methods and experimental settings do not automatically enable us to cope with the complexity of human reasoning, feeling and acting. To overcome some of these shortcomings, we propose a more ecological approach, i.e., to analyze these processes jointly in an interactive computer game setting (a micro world scenario). The creation of the micro world scenario is based on theoretical predictions concerning emotion antecedent appraisal and emotion specific action tendencies as postulated by different cognitive emotion theorists (e.g., Scherer, 1984; Frijda, 1986).

This objective is part of the more general research goal towards modeling emotion, cognition, and behavior as situated processes in a dynamically changing environment. We refer to this as modeling a situated emotional problem solver. To accomplish this, the project’s research strategy is to take advantage of methods of Computer Science and AI. We think that AI methods such as computer simulation and synthetic modeling are instrumental to furthering the development of current psychological research, complementing traditional experimental methods. Consequently, this projects follows different branches of theoretical and methodological developments.

Theoretical objectives

• This research paradigm allows one to study the ongoing process of situated emotional problem solving as it occurs. Therefore, the analysis of this process should help to improve our knowledge about:

1. The relation between the cognitive evaluation of the situation (appraisal) and the subjective feeling.
2. The relation between the subjective feeling and the observable expressive behavior, especially facial behavior.
3. The relation between cognitive evaluation, facial behavior, subjective feeling, action tendencies, and ongoing behavior (game strategies).

• The main goal of the project is to use the empirical data to improve and develop concise and explicit theoretical models of the components that are involved in situated emotional problem solving.

Methodological objectives

• The Facial Expression Analysis Tool (FEAT) is a tool for the automatic coding of facial expressions.
• The Facial Action Composing Environment (FACE) is a tool for creating 3D animated facial expressions in real-time, including head- and eye-movements.
• The Geneva Appraisal Manipulation Environment (GAME) is a tool for inducing and measuring situated emotional problem solving in interactive computer games. This tool also provides modules for the computerized analysis of the subjective experience and the cognitive evaluation of the situation measured using on-line questionnaires. Additionally, the data of the ongoing behavior (game strategies recorded in a protocol file) and the automatically coded facial expressions are analyzed.
• The Geneva Appraisal Theory Environment (GATE) is a software package that allows the simulation of different appraisal theories as black box models. The purpose of black box models is to produce outcomes or decisions that are maximally similar to those resulting from the operation of naturally occurring systems, disregarding the processes by which these outcomes are attained as well as the structures that are involved.
• The Autonomous Agent Modeling Environment (AAME) is a simulation environment for process models. The purpose of process modeling is to attempt to simulate naturally occurring processes using hypothesized underlying mechanisms.

Staff