A Unified Taxonomy of Architectures for Quantifiable Moral Ethics A Field-Defining Prior-Art Note for Computational Moral Quantification

Efforts to quantify moral or ethical behavior in artificial intelligence systems have historically lacked unified taxonomies, conceptual boundaries, and operational clarity. This publication introduces a classification of all known, viable architectural families capable of supporting quantifiable moral ethics—computational systems capable of generating measurable, repeatable, and numerically grounded evaluations of moral content. To the best of the author’s knowledge, no prior publication has established a complete taxonomy describing the architectural solution space for computational moral quantification. This document fills that gap by defining three and only three distinct families: (1) internalized moral-loss architectures, (2) rule-based or heuristic constraint architectures, and (3) external, information-theoretic governance architectures. By delineating these families at the conceptual level, this note establishes foundational prior art for the entire field, preventing monopolization through future patent claims while preserving open academic exploration.