Scale of the self

In 2019 Levin defined the cognitive boundary of an individual as the most distant set of events in time and space that the system can measure and try to regulate.¹ It marks what the system can care about, and in advanced agents it is also the edge of the self-model.¹ Levin notes that the formalism sets no upper bound on its size, and he asks what a being with a far larger boundary would be like.² Ciaunica, Shmeleva and Levin (2023) tie the idea to self-organization. They argue that living systems must form boundaries that define an internal space while staying coupled to their surroundings.⁵

Levin (2022) treats gap junctions as a case study in how the borders of the Self expand or contract.³ He uses it to frame the many-into-one problem, which asks how competent smaller selves bind into a higher Self.³ In his account the Self scales by bringing subunits into one goal-directed loop, not by micromanagement. Short-sighted cells then yield long-term collective rationality.⁴ Levin (2023) makes a similar claim for neural networks. Homeostatic competencies of single neurons are integrated into a higher Self with memories and preferences that no component has alone.⁷

Levin (2023) describes the scaling from single cells to tissues as an enlargement of the cognitive light cone. Collectives gain longer memories, better prediction from past experience and wider spatial measurement.⁶ He lists as an open problem the link between this view and the biophysics of a tissue's electrical circuit.⁶ McMillen and Levin (2024) generalize the point. They hold that all intelligences are made of parts, and that evolution seems especially good at scaling the cognitive light cone of cells.⁸ The held passages come from Levin and co-authors. They offer a framework and proposals, not a direct measurement of a boundary.

Levin (2025) reads cancer as a rolling back to the smaller cognitive light cones of the deep past.⁹ He says this fits the atavistic theory of cancer, which has been confirmed for transcriptional profiles, and the criticisms of the mutation theory.⁹ His own claim is that cancer cells are not more selfish but have smaller selves.⁹ They run action loops with smaller local goals and treat neighboring cells as external environment.⁹ The 2022 paper describes the same pattern as myopic cell action that harms the host.⁴