The Values Spectrum Registry: From What We Value to What We Do

This paper introduces the Values Spectrum Registry, an addressable behavioural values ontology (a structured map of how values show up in behaviour) designed to place declared commitments and observed behaviour in a common representational space so the two can be compared on the same terms.

The registry arranges value behaviour into clusters of five axes each. Every axis runs from a counter-value pole (−1.0) to a value pole (+1.0) across 11 named positions, and each position is a Behavioural Value Expression (BVE) with its own label, address and hex code. The full registry holds 660 BVEs; a seven-cluster business register holds 385. Its central commitment is that counter-values are active states, not absences: the opposite of trust is not zero, it is fear.

The registry extends the Weighted Systems of Values Architecture (WSVA) and serves as the values layer of TILOS, an open architecture for detecting and tracking the Integrity Delta (the difference between what we say we value and what we do). The paper sets out what the registry is and what it does, its scope, and the uses it is explicitly not designed for. It then describes nine settings where it could support new work independently of TILOS, including coaching, team change, civic deliberation, education, AI evaluation, alignment in world models, qualitative research, public accountability and mediation.

Claims are marked throughout as designed, observed or proposed. The registry is a designed structure under early validation: pilot rater data shows strong agreement at the poles and more scatter in the middle of each axis. The paper addresses cultural variation directly. BVE order and addresses are held fixed across cultures, while placement, centres and hinge points are expected to vary. It closes with open questions and a validation path, including blind word sorts across all clusters and cross-country testing.

This work is intended for everyone: anyone curious about how values show up in what people, organizations and AI systems actually do. That includes researchers, practitioners, educators, system designers and the general public.

Keywords

Values ontology · Behavioural Value Expressions · Counter-values · Values Spectrum Registry · WSVA · TILOS · Integrity Delta · AI alignment · World models · Systems thinking · Values governance · Human-centred systems