Marc D. Lewis
University of Toronto, Canada




If the dialogical self is indeed embodied, then we should address the role of the brain, and especially the emotional brain, in dialogical acts. As a step in this direction, I develop aspects of a neural model of internal dialogue. The model is premised on Hermans' idea of voicing and its implications for subjectivity and action. This is important for overcoming the problem of multiple voices in a unified mind subserved by a coherently functioning brain. To solve this problem, I specify two kinds of voices - one that anticipates and listens and one that is spontaneous or impulsive. These voices correspond in cognitive style to two distinct attentional systems in the cortex. According to the model, the anticipating/listening voice, subserved by the orbitofrontal cortex, may take the familiar I-position in an internal monologue. This monologue is generally anxious, and it is fueled by gist-like perceptual expectancies of another's response, mediated by temporal regions. The spontaneous/impulsive voice, subserved by the anterior cingulate cortex, may occupy a second I-position that is often angry or critical. These two cortical systems are partly independent, and they can compete for control based on changes in emotional content and intensity. Thus, switching activation between them may account for semi-autonomous I-positions in a coherent brain. This model highlights a basic dialogical structure that may permit the elaboration of multiple positions over developmental time.