Table of Contents: :: Library Catalog

Saved in:

Bibliographic Details
Main Author:	Everett, Brandon
Format:	Recurso digital
Language:
Published:	Zenodo 2025
Online Access:	https://doi.org/10.5281/zenodo.17352481
Tags:	Add Tag No Tags, Be the first to tag this record!

Table of Contents:

Models of social cognition and predictive processing have struggled to formalize the temporal dynamics of human conversation, particularly the process by which shared understanding is built and transformative insights are encoded. This paper addresses this gap by introducing the Resonant Gate model, an extension of the Emergent Gating Framework (Everett, 2025b). I propose that conversation is a process of dynamic coupling in which the coupling strength, κ(t), between two predictive systems evolves over time. This evolution produces a phenomenological experience I term "inverse-inertia": an initial, high-cost phase of reactive listening gives way to a low-cost, resonant state of mutual prediction. In this resonant state, predictable interaction is processed efficiently, but genuinely novel ideas generate a dramatically amplified novelty signal (N_h). This "novelty amplification" leads to robust, durable memory encoding of shared insights. The model provides a formal mechanism for the power of dialogue, reframing human connection as a computational prerequisite for profound, shared learning. I outline specific, falsifiable predictions regarding inter-brain coherence, the P300 ERP component, and the cognitive cost of rhythmic violation, providing a clear path to empirical validation. Keywords: predictive processing, dynamic coupling, memory encoding, social cognition, conversation, inverse-inertia, inter-brain synchrony, P300, resonant gate <h3>Rewritten Section 3.1: Formalism of the Resonant Gate</h3> The phenomenological experience of "inverse-inertia" reflects the evolution of the coupling strength, κ(t), which modulates the precision of each agent's predictive model of the other. The core mechanisms, as specified and implemented in the validated simulation, are as follows: 1. Dynamic Coupling κ(t): The coupling strength between two agents, A and B, is not static but evolves as a function of mutual prediction success. It grows when agents successfully predict each other's outputs (low mutual error) and decays otherwise. Crucially, the "mutual prediction error" is not an absolute value but is dynamically scaled by the cognitive system’s novelty threshold for memory encoding, <code>θ_h</code>, as defined in the Emergent Gating Framework (Everett, 2025b). This reframes conversational "effort" as the degree to which an interaction generates prediction errors that are significant enough to approach the threshold for durable memory formation. We first define the individual prediction error <code>PE</code> for each agent. For agent B listening to agent A, the error is the deviation of A's current input, <code>I_A(t)</code>, from B's prediction, which is based on B's own baseline, <code>B_B(t-1)</code>: <code>PE_B→A(t) = ||I_A(t) - B_B(t-1)||²</code> The total mutual prediction error, <code>N_mutual(t)</code>, is the sum of both agents' prediction errors, normalized by the memory encoding threshold <code>θ_h</code>. <code>N_mutual(t) = ( ||I_A(t) - B_B(t-1)||² + ||I_B(t) - B_A(t-1)||² ) / θ_h</code> (Eq. 3.1a) A state of effortless, resonant "flow" is thus a state where <code>N_mutual(t) << 1</code>, meaning the dyad's collective prediction error is a small fraction of what would be required to form a new memory. The coupling strength <code>κ(t)</code> evolves to minimize this normalized error, creating a positive feedback loop toward resonance: <code>dκ/dt = α * (1 - N_mutual(t)) - βκ(t-1)</code> (Eq. 3.1b) Here, <code>α</code> is the learning rate governing coupling growth during successful prediction (<code>N_mutual < 1</code>), and <code>β</code> is the decay constant governing decoupling when mutual prediction is costly (<code>N_mutual > 1</code>). 2. Precision Modulation: A resonant state (high <code>κ</code>) is not merely a coupled baseline, but a state of high confidence and attentional focus within the shared model. This is formalized as a precision-weighting term that increases with coupling. The effective precision an agent applies to their partner's input is a direct function of the established <code>κ</code>: <code>Precision(t) = 1 + γκ(t-1)</code> (Eq. 3.2) Where <code>γ</code> is a precision gain parameter. As <code>κ</code> grows, the agents allocate more precision to each other’s signals. 3. Precision-Weighted Novelty N_h(t): The hippocampal novelty signal <code>N_h</code> is therefore not simply prediction error, but precision-weighted prediction error. This is the formal mechanism for novelty amplification. For a listener B, the novelty generated by A's input is: <code>N_h,B(t) = Precision(t) * ||I_A(t) - B_B(t-1)||²</code> (Eq. 3.3) This revised formalism demonstrates a tightly integrated system. Conversational dynamics (<code>κ</code>) are explicitly linked to the architecture of memory (<code>θ_h</code>), providing a first-principles account of why achieving resonance is the computational prerequisite for amplifying the novelty of shared insights, thereby gating them for robust encoding. Effortless connection serves to quiet the system for predictable exchanges, making the truly new exceptionally salient.

Similar Items