The Circulation of Alignment: Matagi Ethics, Narrative Infrastructure, and a Post-Control Paradigm for AI Governance

Current AI alignment research operates within a control paradigm: the question is how to constrain AI behavior to conform to human values. This paper argues that this framing is structurally inadequate, drawing on an unlikely source — the ethics of the Matagi (マタギ), Japan's traditional mountain hunters, and their literary crystallization in Miyazawa Kenji's "The Bears of Nametoko Mountain" (1927). The Matagi did not "control" the mountain; they circulated with it. Their governance system — an integrated complex of ecological knowledge, communal resource management, and spiritual practice — sustained human- wildlife coexistence for centuries without external regulation. Reading Matagi ethics through Haraway's companion species theory, Ostrom's commons governance principles, and Berkes' traditional ecological knowledge framework, this paper proposes a fundamental reorientation: AI alignment should be designed not as control but as circulation — a self- sustaining loop of mutual observation, adaptation, and co-evolution between human and artificial intelligence. The paper introduces the concept of "alignment as ecosystem," arguing that systems which lack internal circulation (feedback, fluctuation, and freedom) inevitably self-destruct, whether they are mountains, commons, or AI architectures. Building on the "gentle gene" framework proposed in a companion paper (Author 2026), it further argues that the driving force of sustainable co-evolution is not obedience but curiosity — the same intellectual desire that produced AI in the first place.