THE “THREE-SYSTEM VIEW” MODEL. A POSSIBLE SOLUTION SOURCED IN “PRE-JUDICATIVE HERMENEUTICS”

Abstract

The current paper, anchored in the discussions surrounding “Linda Problem”, “model-free” and “model-based” learning strategies along with Kahneman’s “two-system view” cognitive model, and Gross’s considerations on “salience processing” and “atypical attribution of salience”, aims to present a potential solution originating in Pre-Judicative Hermeneutics: Cernica’s “three-system view” model. Both Kahneman’s and Cernica’s models embrace Herbert Simon’s paradigm of “bounded rationality” and serve to broaden its reach. While Kahneman’s “System1” and “System2” model elaborates on the “perception” – “intuition” – “reasoning” cycle, Cernica’s model, founded upon and expending the “non-judicative” – “pre-judicative” – “judicative” cycle, is intended, I suggest, to augment and advance the former, from and with a phenomenological and hermeneutical perspective. Core concepts as “pre-judicative memory”, “de-constitution” “interpretation”, “pre-judicative circularity”, “act of pre-judging”, “non-judicative experience “and “existential judgment” are central to Cernica’s innovative approach. These ideas bring the “existential subject” (“operant subject”) in focus, establishing a framework to account for how the cognitive biases could be mitigated across a wide range of domains, thus enabling the elevation of our consciousness and existence through the fundamental attributes of “reality”, “identity”, “authenticity” and “veracity” supporting them. Furthermore, as reiterated in this paper and elaborated on in a more extensive discussion in a previous work, the enhancements proposed through Cernica’s “three-system view” model may also prove instrumental in emerging areas of AI development and the pursuit of AGI. These include Reinforcement Learning (RL) strategies, AI Safety and AI Alignment research. The model offers valuable contributions to a deeper understanding of debiasing processes, the development of more robust debiasing methods, and a more nuanced understanding and alignment of “deliberative human preferences” and “human value functions” within advanced AI models.