Abstract
Within the methodologically diverse interdisciplinary research on the minimal self, we identify two movements with seemingly disparate research agendas – cognitive science and cognitive (developmental) robotics. Cognitive science, on the one hand, devises rather abstract models which can predict and explain human experimental data related to the minimal self. Incorporating the established models of cognitive science and ideas from artificial intelligence, cognitive robotics, on the other hand, aims to build embodied learning machines capable of developing a self “from scratch” similar to human infants. The epistemic promise of the latter approach is that, at some point, robotic models can serve as a testbed for directly investigating the mechanisms that lead to the emergence of the minimal self. While both approaches can be productive for creating causal mechanistic models of the minimal self, we argue that building a minimal self is different from understanding the human minimal self. Thus, one should be cautious when drawing conclusions about the human minimal self based on robotic model implementations and vice versa. We further point out that incorporating constraints arising from different levels of analysis will be crucial for creating models that can predict, generate, and causally explain behavior in the real world.
Highlights
The minimal self describes the immediate, pre-reflective experience of selfhood derived from sensory information (Gallagher, 2000; Blanke and Metzinger, 2009)
It has been subdivided into the sense of agency (SoA, “I produced an outcome with my voluntary action.”) and the sense of ownership
We have established that there is no complete mechanistic explanation of the minimal self yet – but why should mechanistic models be beneficial for further research on the minimal self? We see several benefits in striving for integrating evidence from different levels of description and thereby creating more mechanistic models of the minimal self: (a) It safeguards against overfitting to specific pieces of evidence, assumptions, or tasks, (b) it increases model comparability and the probability of model generalization, and (c) especially in clinical contexts, a causal understanding may help to find effective interventions fordisorders and interfaces with other theories (e.g., Schroll and Hamker, 2016; Neumann et al, 2018)
Summary
The minimal self describes the immediate, pre-reflective experience of selfhood derived from sensory information (Gallagher, 2000; Blanke and Metzinger, 2009). The same reservation as for Bayesian models applies – in our view, showing that an optimization scheme can be implemented through neural computation, while being necessary for a possible mechanistic explanation, is not sufficient as long as the more specific model does not capture relevant deviations from behavior predicted by computational constraints alone One such deviation yet unexplained by computational models may be the apparent dissociation of explicit and implicit measures of SoO in the rubber hand illusion under certain conditions (Holle et al, 2011; Rohde et al, 2011; Gallagher et al, 2021), which has been explained under the same framework of information integration (Apps and Tsakiris, 2014). Such an approach might require going to the edge of what is currently computationally possible (cf. Clune, 2020)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
