Human–Agent Symbiotic Cognitive Architecture (HASCA): A Pair-Centric Research Program for Human–AI Co-Reasoning and Cognitive Autonomy

Human–Agent Symbiotic Cognitive Architecture (HASCA) is a proposed, falsifiable research framework for longitudinal human–AI collaboration, cognitive autonomy, calibrated reliance, and governed agent memory. The framework comprises three research components: Layered Reasoning Scaffold (LRS), Pair Reflection Audit (PRA), and Governed Pair Memory. It proposes a structured approach to investigating whether recurring human–AI collaboration can support independent reasoning while preserving human decision authority. A Stage-1 feasibility study with 36 adult participants is proposed to compare human-only reasoning, conventional AI assistance, and LRS-assisted reasoning. The study has not yet been conducted, and no experimental superiority is claimed. This non-peer-reviewed preprint presents the conceptual architecture, its relationship to prior research, testable hypotheses, and proposed evaluation methods. Research originator and author: Goshtasb Gholami (ORCID: 0009-0009-7302-0380). AI-assisted research and drafting support are disclosed in the manuscript.