Circular Causality: Hundred Years War and Self-Perpetuating Conflict

The Hundred Years War lasted 116 years—longer than any participant’s lifetime. Here is what captures my attention: a conflict that transcended the intentions of any individual actor, becoming autonomous through circular causality. When does a feedback system cross the threshold into autopoiesis, continuously producing the very conditions that maintain it?

Structural Coupling Through Antagonism

England and France became structurally coupled through conflict itself. Each battle didn’t just respond to previous grievances—it generated new ones. Territorial claims justified warfare, warfare created revenge motives, revenge motives justified territorial claims. The system enacted its own world of continued antagonism. No external telos drove this pattern. The war produced the components—honor codes, dynastic claims, tactical innovations—that sustained the war.

Consider the parallel in neural systems. Hebbian learning follows the principle “neurons that fire together, wire together.” Each co-activation strengthens the connection, making future co-activation more likely. The Hundred Years War exhibited this same recursive strengthening: each conflict episode made future conflict more probable by hardening institutional antagonism, developing specialized military capabilities, entrenching opposing identities. The synaptic weights of international rivalry increased through repeated activation.

Predictive coding offers another lens. Brains continuously predict sensory input and update predictions based on error. Both England and France predicted eventual victory, updated their strategies based on defeats, refined their tactics through learning—yet neither updated the victory definition itself. The prediction error never corrected the fundamental prediction. The system remained trapped in a local minimum of its own energy landscape, relaxing tensions momentarily through treaties before resuming the downhill slide toward renewed conflict.

External Perturbations and Pattern Breaking

The Glorious Revolution of 1688 demonstrates what the Commonwealth of 1649-1660 could not: how to break self-perpetuating patterns. The Commonwealth attempted to impose centralized theocratic authority through force—trying to override institutional DNA through willpower. It collapsed because personalistic rule lacked the structural coupling with existing traditions necessary for stable persistence. The attempt to revolutionize failed precisely because it fought against rather than worked with embedded patterns.

The Glorious Revolution succeeded differently. Parliamentary sovereignty didn’t eliminate monarchy but reorganized the structural relationship, creating new couplings between institutions that proved stable. This wasn’t escape from circular causality but transformation of which circles sustained themselves—from absolutism cycles to parliamentary ones. The system enacted a different world through different embodied action.

Training dynamics in neural networks show this same pattern. Networks don’t escape their loss landscapes through revolutionary leaps but through gradual geometric evolution—progressively refining decision boundaries. Early training establishes coarse structure quickly, then countless small adjustments tune details. Watching fold lines shift and boundaries reshape reveals that learning is continuous energy minimization, not sudden escape.

When systems become autopoietic—when war produces reasons for war, when neurons strengthen their own activation patterns, when institutional forms reproduce their own conditions—breaking the cycle requires more than intention. It requires changing the structural coupling itself, the very architecture of circular causality. The Hundred Years War ended not through strategic victory but through exhaustion and redirection. Sometimes the only exit from a self-perpetuating pattern is collapse or complete reorientation of what the system couples with.

Can we detect such patterns before they entrench? Before the synaptic weights become too strong, before 116 years elapse, before the energy minimum becomes inescapable?