Instead of HEXACO at the core with everything hanging off it, the five theories sit as equal nodes. They connect only where the science says they relate and the person's own scores show it too. Play with a person below, then read exactly how it works.
Pick a sample, or open a theory and move its dials. Nodes resize by how pronounced you are; links light up where the science and your scores agree.
The earlier model put HEXACO at the centre and grew the other four theories off it as appendages. This one treats all five as co-equal nodes. Each keeps its own internal glyph drawn from its own sub-scores, so the richness lives inside the nodes and the relationships live on the lines between them.
Honest from the start. The evidence is not symmetric. HEXACO is the most validated and most-connected of the five, so we do not force it to the centre, but we do not pretend the five are equal in evidence either. The layout places HEXACO (it lands well-connected on merit), base node sizes are equal, low-evidence links draw faint and dashed, and nothing here is a public personality claim until a psychologist signs off the relationship map.
The research surfaced one thread shared by every framework, which gives the constellation a real coordinate system instead of arbitrary positions. Use two axes (kept 2D, not collapsed to a line). They are, not by accident, the two higher-order axes of the Schwartz value circumplex, which the other frameworks line up onto.
Approach / growth vs anxiety / threat. Growth: BAS, attachment security, openness-to-change values, need satisfaction, Extraversion and Openness. Threat: BIS, attachment anxiety, conservation values, need frustration, Emotionality.
Self-transcendence / communion vs self-enhancement / agency. Communion: Honesty-Humility, Agreeableness, universalism and benevolence values, relatedness. Agency: power and achievement values, low Honesty-Humility, BAS drive.
Toggle "organising axes" on the prototype to see them behind the constellation.
Every pair of theories has a documented relationship strength (the bar) and an evidence confidence (how well-studied it is). Both come from a five-axis literature sweep. The matrix drives where the nodes sit and how strongly a link can ever draw. Confidence is shown honestly, so a thin or proxy-only link can never masquerade as a solid one.
Representative figures calibrated from the literature; exact coefficients pending full-text verification and psychologist sign-off. Strongest single links found: Honesty-Humility to power values (-.60), Agreeableness to benevolence (+.61), Openness to self-direction (+.52), Emotionality to attachment anxiety (+.35), attachment security to need satisfaction.
Every node starts at the same base size, then nudges up or down a little by salience (how far this person sits from the average on that theory), clamped to a tight band. A vivid theory reads slightly stronger, none can dominate, and a flat one keeps a floor.
A link between two theories is the product of three gates: prior × confidence × personal-coexpression. The coexpression term only fires when the person's own standardised scores sit along the documented direction of the link (high-high for a positive correlation, and so on). So a line appears only when the science supports it and the person actually shows the pattern. Its thickness and brightness scale with the result; low-confidence links draw faint and dashed.
The relationship map is calibrated from real research but is designed, not validated for this product. A psychologist signs off before any public "this is your measured personality" claim. Three things a careful reader should know:
Privacy carries over. The shareable still is drawn from bands (low, mid, high), never exact scores; the exact, interactive version stays private; a per-person seed means two identical profiles never render the same. Note for the build: the coexpression maths needs population norms to standardise scores, so the quiz needs a norm sample (even borrowed published ones) before the links fire correctly.