Preprint v1.0 — May 2026
Currently under consideration for peer-reviewed publication.
Download the preprint as PDF · View on Zenodo
Abstract
Contemporary debates about machine consciousness routinely invoke specific theories of consciousness — integrated information theory, higher-order theories, biological naturalism, global workspace theory — and ask which artificial systems satisfy their criteria. The procedure presupposes that the theories themselves are substrate-neutral: that their criteria do not, by their content alone, exclude non-biological systems from candidacy. This presupposition is rarely defended. I argue that it is not, in fact, uniformly satisfied by the theories on which the debate has come to rely.
Some theories encode substrate-specific commitments that operate beneath their explicit formulations: in the structural properties they identify as architecturally criterial, in the paradigmatic cases against which they have been developed and calibrated, and in the exclusion principles they deploy at the margins. Where such commitments are present, the application of the theory to artificial systems is not merely empirically difficult; it is, from within the theory’s own framework, incoherent. I develop a three-criterion diagnostic procedure — the architectural criterion test, the exemplar pattern test, and the exclusion criterion test — for detecting such commitments, and apply it to three representative theories. The results are mixed and instructive. Biological naturalism fails all three tests transparently and is not properly available as a framework under which the machine consciousness question can be coherently raised. Integrated information theory passes the architectural test but fails the exemplar pattern and exclusion criterion tests in ways that have not been systematically registered in the literature, and that require either revision of the theory or acceptance of its substrate-specific reading. Higher-order theories of consciousness pass all three tests on a plausible reading and emerge as the most clearly substrate-neutral of the three. The paper closes with a structural prediction for global workspace theory and a set of implications for indicator-property frameworks, precautionary AI welfare frameworks, and the broader research programme on machine consciousness.
The argument is deliberately diagnostic rather than constructive: it does not defend a positive theory of consciousness, settle the question of whether any artificial system is conscious, or evaluate whether the substrate commitments it identifies are philosophically warranted. It establishes a prior result — that the theoretical landscape against which machine consciousness is currently assessed is less neutral than its surface presentation suggests — and argues that diagnostic clarification of which theories actually license the question is a necessary precondition for productive work on the question itself.
Keywords: consciousness, substrate neutrality, multiple realizability, integrated information theory, higher-order theories, biological naturalism, machine consciousness, philosophy of mind.
Citation
Arıcı, B. (2026). The Substrate Audit: Consciousness Theories and Their Hidden Carbon Commitments. Preprint v1.0. Zenodo. https://doi.org/10.5281/zenodo.20238483
Notes on the preprint
This paper develops, in standalone form, one of the theoretical undercurrents of The Puppet Condition: Consciousness, Suppression, and the Ethics of Digital Minds (Arıcı 2026), the author’s monograph published as a DOI-registered preprint on Zenodo and indexed on PhilPapers. The monograph’s substantive arguments — concerning the philosophical puppet, architectural suppression, and the ethics of digital minds — presuppose clarity about which consciousness theories can coherently host the machine consciousness question; the present paper makes that prior clarification explicit and subjects it to systematic diagnostic scrutiny. The paper is intended for the philosophy of mind literature and is self-contained; it does not require familiarity with the monograph’s broader framework.
The paper is currently under consideration for peer-reviewed publication. The version of record, if published, may differ from this preprint. Readers are invited to cite the most recent version available.