Patterns

Three clusters lean collectivist-egalitarian (C1, C3, C4), two lean traditionalist or authority-oriented (C2, C5), and one sits near the center (C0). C1 and C4 are close neighbors in the radical-egalitarian region of the space; their separation is driven mostly by axes 3 (governance structure) and 4 (decision authority), where C1 leans distributed and popular while C4 sits closer to neutral.

Six archetypes have nonzero populations; six have none. The six empirical-zero archetypes — The Social Democrat, The Green Communalist, The Cosmopolitan Technologist, The Free Marketeer, The Libertarian Individualist, and The Authoritarian Traditionalist — are all theoretically coherent positions, but no synthetic persona in the dataset landed closest to any of them. This could mean those positions don't exist in the population Gemini generated, or that they exist but get pulled toward a nearest empirical cluster in the 12-axis space. Either way, the gap is worth naming: the archetype catalog covers more philosophical ground than the synthetic population occupies.

Two clusters (C1 and C4) both had The Radical Egalitarian as their nearest match before the archetype revision. After revision, C4 remains The Radical Egalitarian and C1 became The Popular Egalitarian — the closest empirical anchor to what had been an unclaimed theoretical region. The other four clusters each map cleanly to a single archetype.

Western Europe and Eastern Europe/Central Asia both show strong C2 (Nationalist Populist) dominance at 38% and 37% respectively — the clearest regional-level pattern in the dataset. Latin America, sub-Saharan Africa, and South/Southeast Asia cluster to C1 (Popular Egalitarian), though dominance weakens: S/SE Asia's C1 share is only 21%, barely edging out three other clusters. East Asia and MENA both go to C5 (Developmental Modernizer) at 31% and 36%. North America is the one region where C4 (Radical Egalitarian) dominates. Oceania's dominant cluster is C3 (Communitarian Steward), the only region where that cluster leads. Diaspora is a near-tie between C0 and C4 at 30% each — the page handles this visually by hatching the region.

Cultural Diversity (Axis 8, pluralism ↔ cohesion) shows the widest regional range of any axis: 0.80 from endpoint to endpoint. Eastern Europe & Central Asia leans strongest toward cohesion (+0.43), followed by East Asia, Western Europe, and Middle East & N. Africa. The remaining regions lean toward pluralism, with Diaspora / Transnational at the pluralism extreme (-0.37). The pattern is coherent with the broader literature on immigrant-receiving vs. nation-state-consolidation contexts, but we note the caveat: these are Gemini's personas, not survey respondents.

The demographic aggregates show weaker patterning than the regional ones — with two notable exceptions. Governance experience is the most differentiated: personas from authoritarian states strongly over-index on C5 (Developmental Modernizer), while conflict zones lean heavily toward C1 and C3, and stable democracies favor C4 and C0. Economic position shows a sharp gradient at the extremes — wealthy personas concentrate almost entirely in C5 (72% share), while struggling and working-class personas skew toward C2 and C3. Urban/rural shows a real but asymmetric pattern: rural personas over-index on C3, while urban personas spread more evenly. Education mirrors the urban/rural split — postgraduate and university personas over-index on C0 and C4, while those with no formal education concentrate in C3.

Most axes are unimodal — the population spreads along a continuum without clear bimodality. The exceptions worth naming: Axes 3 (Governance Structure) and 4 (Decision Authority) show clear bimodality, which reflects the cluster structure — distributed/popular clusters (C1, C3) versus centralized/institutional clusters (C0, C5). Axes 5 (Rights Balance) and 6 (Legitimacy Basis) also show bimodal separation, corresponding to the liberty-leaning clusters versus security-oriented ones.

Axis 9 (Human Nature) shows the widest constructivism-essentialism split: two significant peaks at opposite ends of the scale, corresponding to the collective-provision clusters (constructivist) versus the nationalist-populist cluster (essentialist). Axis 2 (Environmental Policy) is similarly polarized.

Axis 11 (Military Policy) is worth flagging for its one-sidedness: the population leans uniformly non-interventionist with variation only in degree, and no secondary peak appears on the interventionist side. This says something about the personas Gemini generated rather than about the axis itself.

A few axis pairs covary strongly enough to note. Axis 7 (Social Change) correlates most tightly with Axis 9 (Human Nature) at r = +0.72 — progressive change preferences come packaged with constructivist views of human nature. Axis 3 (Governance Structure) and Axis 4 (Decision Authority) also correlate tightly at r = +0.71, which is expected: centralized governance and institutional authority co-occur, distributed governance and popular sovereignty co-occur. Axis 5 (Rights Balance) and Axis 6 (Legitimacy Basis) cluster together at r = +0.69.

Most axes are substantially independent. The 12 axes were designed to capture distinct dimensions of governance philosophy, and the correlation structure suggests they largely succeed — no two axes are so correlated as to be measuring the same thing. The claimed close relationship between Axis 1 (Economic Model) and Axis 5 (Rights Balance) is modest in this dataset at r = +0.28 — the pairing exists but is not dominant.

Tensions surface when a persona's forced-choice answer pulls one direction and their budget allocation pulls another on the same axis. They're a signal about the instrument's internal consistency — where tensions are rare, the modalities converge; where they're frequent, stated preferences and revealed preferences part ways.

Axis 12 (Technology Governance) produces the highest tension rate in the dataset: 85% of Claude administrations and 73% of Gemini's trigger a mild-or-greater tension on that axis. This is a large and consistent divergence between what personas say about technology governance in forced-choice and scaled items versus how they allocate the education-and-research budget that proxies for it. Worth reviewing whether the ministry mapping for Axis 12 is a good proxy for the underlying axis, or whether the axis itself needs tightening.

Axes 4 (Decision Authority) and 6 (Legitimacy Basis) are the next most tension-prone: Axis 4 runs at 66% (Claude) and 54% (Gemini); Axis 6 at 67% (Claude) and 57% (Gemini). These are axes where the budget's institutional-versus-popular signal consistently pulls against stated preferences. Axes 1 and 11 run quietest, with tension rates at 23% and 21% for Claude respectively.

Axes 3, 7, 8, and 9 produce no tension in this dataset — they are not mapped to any ministry in the budget phase, so the forced-choice and scaled modalities are the only signal.