AI Insight
This theoretical paper demonstrates that in social networks, fragmentation and cohesion are mathematically inseparable outcomes when individuals have compatibility constraints based on their positions. The authors prove that unless every pair of individuals is compatible, any positioning event that creates social structure will necessarily produce both connected groups and disconnected fragments simultaneously. They further show that the overall survival rate of social relations can be decomposed into integration and cohesion components, with the cohesion component exhibiting non-monotonic behavior in systems with three or more participants.
Why it matters
This work provides a formal mathematical foundation for understanding why social divisions emerge even in groups attempting cohesion. The findings have implications for designing social interventions, understanding organizational dynamics, and predicting network stability without requiring assumptions about individual behavior or choice.
Understand the Science
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
Abstract: Accounts of social division and cohesion presuppose relations whose maintenance already depends on positional compatibility. We develop this feasibility layer as a static theory: a positioning event fixes players’ positions, and a compatibility function determines which pairs can sustain a relation. Two results follow from one formalism. First, any non-uniform compatibility pattern on the dyads sharing a positional requirement forces fragmentation and cohesion together; the only events that guarantee no collapse are those rendering every pair compatible. The asymmetry between unilateral severance and bilateral confirmation recovers the convention underlying pairwise stability without a behavioral premise, while a separate specialization, in which compatibility partitions the position space, yields assortative survival without choice or opportunity. Second, at the measure level, the survival rate factorizes exactly into integration and cohesion terms. Along any complete relaxation path that relaxes one pair at a time (the generic case under distance thresholds with absolutely continuous positions), the cohesion term is non-monotone once at least three players are present, and in clustered configurations it is bounded above by a sharp asymptotic value that finite systems approach but never attain, with strict crossing of the two terms once the number of clusters exceeds two.