Conformity-driven social learning acts as a topological disruptor in the non-equilibrium evolution of cooperation

The evolutionary fate of cooperation is governed by the interplay between individual utility maximization and the pressure to conform. While traditional evolutionary game theory relies on success-biased learners, the macro-evolutionary impact of frequency-biased social learning remains under-explored. We develop a coupled dynamical system integrating payoff-driven individual learners with conformity-driven social learners. We uncover that strong normative conformity acts as a profound topological disruptor, irrevocably destroying strategic diversity. In anti-coordination environments, critical social pressure eradicates the cooperative middle ground, forcing a topological collapse into absolute polarization. Furthermore, a critical mass of social learners rigidly fractures the state space, birthing highly stable, decoupled asymmetric equilibria: the \textit{Social Trap} (artificially rescuing cooperation against rational free-riding) and the \textit{Social Drag} (pathologically suppressing cooperative pioneers). Governed by a universal, payoff-invariant phase boundary, this structural decoupling mathematically proves that the topology of social influence can completely override universally dominant economic landscapes. Our findings provide a rigorous mechanistic explanation for the extreme polarization of modern societies and highlight the inherent evolutionary paradox of social governance.