I will have two central goals in the first part of this talk, which explores the relevance of latest research on mind-wandering for theories of consciousness. First, conceptually, and in opposition to what many philosophers following Descartes and Kant traditionally have liked to believe, I will argue for the claim that conscious thought actually is a subpersonal process, only rarely a form of mental action, but rather an unintentional form of mental behaviour, and demonstrably for more than two thirds of our conscious life-time. The paradigmatic, standard form of conscious thought is non-agentive, it lacks veto-control, and involves an unnoticed loss of epistemic agency and goal-directed causal self-determination on the level of mental content. Second, I present an empirical hypothesis: There will be a detectable self-representational blink (SRB), a small time window I which we are blind to ourselves, namely, when shifting from one phenomenal self-model or “unit of identification” (UI) to the next. Alluding to the well-studied phenomenon of the attentional blink (Raymond, Shapiro, and Arnell, 1992, Shapiro, Raymond, and Arnell, 1997), the notion of a “self-representational blink” refers to the fact that we are typically not able to consciously experience the actual moment of transition from mindful, present-oriented self-awareness to the identification with the “protagonist” of a daydream, the content of the self-model in autobiographical planning, etc. Phenomenologically, the SRB is characterized by a brief loss of self-awareness, followed by an involuntary shift in the phenomenal UI; functionally, we can describe it as a failure of attentional and/or cognitive self-control. The empirical prediction is that subjects should be blind to self-related stimuli during the SRB, and my main hope is that the audience can help in developing novel experimental paradigms to test this hypothesis.
If time allows, I will also take a closer look at the concept of “mental action” in the second part. Can we conceptually accommodate mental actions under a predictive processing approach? My main positive claim will be that mental action is the predictive control of effective connectivity, where what is predicted is the epistemic value of states integrated into the phenomenal self-model under counterfactual outcomes.