The Pre-Analytical Institutional Stack: ETI, ASLTP, and PJLS as a Structural Framework for AI-Mediated Institutional Reasoning

This paper introduces a pre-analytical institutional stack designed to structure reasoning processes in rule-dense institutional environments before interpretive or narrative analysis occurs. The framework consists of three integrated components: the Emotional Tension Index (ETI), the ASLTP routing protocol (Affect–Surface–Label–Transform–Present), and the Pre-Judgment Logic Shell (PJLS). The ETI layer encodes institutional events into structured signals that capture tension patterns within institutional environments. The ASLTP protocol then routes these signals through a standardized process that surfaces and labels institutional conditions before transforming them into structured representations. The PJLS layer functions as an admissibility gate, verifying logical coherence, evidentiary mapping, and path continuity prior to any reasoning stage. Within this architecture, AI or large language model reasoning is positioned as a downstream analytical layer rather than the primary interpretive mechanism. By separating signal encoding, institutional routing, and admissibility verification from subsequent reasoning, the framework aims to reduce narrative drift and hallucination risks commonly associated with generative AI systems. The proposed stack operates within an institutional tension field (ITF), conceptualized as the operational environment in which institutional signals, rule interfaces, and routing structures interact. The model therefore reframes AI-assisted analysis not as autonomous reasoning but as a constrained process embedded within institutional structures. This architecture is intended for applications in legal analysis, institutional governance, regulatory environments, and AI-mediated decision support systems where rule-based admissibility and procedural validity are critical.