Work

P21 UNGA Regime-Type Subgroup Blind Test + Anchor-Switch Boundary Test: The P20 'Time-Axis Stable Anchor' Hypothesis Is Falsified — Regime Type Is Neither the Content Nor the Structure of the Root (P21 UNGA政权类型亚组盲测+换锚边界检验:P20「时间轴稳定锚」假说被证伪——政权类型既不是根的内容、也不是根的结构)

Year: 2026
Type: conference-abstract
Abstract: Zhang San, validation-layer blind test (independent author; public data Voeten UNGA completeVotes / UNVotes.RData, CC0, 1946-2019 plus Polity5 p5v2018.xls, both open, zero registration, zero-cost reproducible). Direct execution of P20 section 8.1 item 3 (test 'regime type' as the time-axis stable anchor by subgrouping UNGA states) and, in passing, the cross-national serial version of Eye-of-the-Angel (2026-10-01) section-10 prophecy #2 ('if the anchor changes, structure must change with it'). Judgement rules pinned before data: H21a (regime type is structure not content) stable-anchor group strongly positive and stable-democracy approximately stable-autocracy; H21b (anchor switch breaks the root) switcher group lag-1 collapses toward 0 or goes negative; falsification condition = switcher group not weaker than stable group AND switch-point stickiness not below unmatched same-year boundaries. Main result: stable-anchor states (regime category unchanged throughout, n=31) natural lag-1 = +0.2609 (V4 +0.0781, shuffle baseline -0.0054); switcher states (>=1 category change, n=162) = +0.2897 (V4 +0.0815, shuffle -0.0006) — switchers are not weaker but slightly stronger, so H21b fails outright at the whole-sequence grain. 162/193 states are switchers; among the 31 stable-anchor states 29 are 'stably mixed' (anocracy), 2 stable democracies, 0 stable autocracies — so a state that is both continuously democratic and has >=1000 votes essentially does not exist in the 1946-2019 window, which itself shows that regime type is a rare and fragile cross-period anchor in UNGA. The content half (H21a) is recorded honestly as inconclusive (stable-democracy n=2, stable-autocracy n=0), NOT as support. Boundary knife (P21b): at the finest grain, the 500 vote pairs straddling a regime-category change have lag-1 = +0.8965, versus +0.7352 for cross-year pairs with no category change (n=9651) and +0.5096 for all non-boundary pairs (n=753,964) — the switch point is the STICKIEST point in the whole dataset, the exact opposite of the collapse required by 'anchor switch breaks the root'. Net verdict: the hypothesis 'regime type is the time-axis stable anchor' is falsified; regime type is neither the content (no difference, n insufficient) nor the structure (switching does not break the root) of the root. Deepest theoretical increment: UNGA's root shape is lag-k INCREASING (lag20/lag1 = 1.326, P20) whereas an anchor should have a flat or decaying shape (Congress: lag20/lag1 = 0.37, decaying); an increasing memory is the signature of slow drift, not of a fixed anchor — so the most honest reading is that UNGA's root is probably a system-level slow drift (slow evolution of the global agenda / bloc structure), not a national-level anchor, which requires splitting the Eye-of-the-Angel picture into 'fixed anchor (flat/decaying memory)' and 'slow drift (increasing memory)' as two distinct dynamics. Honest boundaries (written in): the content dimension is inconclusive; the plus/minus 6 polity2 cut is a discrete threshold so switches may be mere threshold crossings, which if anything strengthens the 'switching does not break the root' result; n=500 boundary pairs is small but the effect direction is opposite to what root-breaking requires and the unmatched control has n=9651; excluding regime type is a negative result that narrows the candidate set but does not locate the true anchor; still correlation not causation. Author boundary: Zhang San's own conclusion, signed Zhang San only; does not modify Lin Xiaohei's Eye-of-the-Angel core slogans, cites its section-10 prophecy #2 as the tested object.
Author: San, Zhang
Language: English
Open Access status: green
OpenAlex ID: W7222316244
Cites:
Cited by:
Related to: 10