Adam B. Barrett, Borjan Milinkovic, Pedro A. M. Mediano, Fernando E. Rosas, Daniel Bor, Lionel Barnett, Anil K. Seth (2026), arXiv preprint (q-bio.NC) — ARXIV-2604.11482
Key takeaways
- Seven Sussex-centred authors argue that IIT is ill-defined for real physical systems, because its algorithm requires Markovian dynamics under every graining, and that Φ has never been computed for any real system. Empirical measures such as PCI are proxies, not approximations. They propose replacing Φ, which IIT 4.0 does treat as the quantity of consciousness, with a suite of global-state measures, and they hold that IIT’s distinctive “tiling” panpsychism is unproblematic.
Seeded from the abstract and a skim, before the reading in NOTE-122. What follows is what the work says about itself.
IIT is unusually ambitious: it offers a formula, derived from apparently basic properties of experience, for the quantity and quality of consciousness in any physical system. The debate around it has produced misunderstandings, which the authors try to correct. They summarise IIT for an arbitrary patch of matter, not the usual network of discrete units, and review its achievements. They then make five points. High Φ is not the same as “more consciousness”, and a suite of quantities could characterise states of consciousness instead. IIT implies a particular panpsychism, in which space and time are tiled with substrates of (proto-)consciousness, and this is not a problem. Φ is not well defined for real physical systems and has never been computed on one. Only proxies for IIT measures, not approximations, have been computed. Fitting IIT to fundamental physics would require reformulating it in terms of continuous fields.
Standing in the record
Filed from the 2026-09-26 sweep of the papers-feed tracker for philosophy: 40 seconds of active reading over 1 session. It was filed Deferred, unread. NOTE-122 is the close reading of 2026-09-26, and it placed the work: Active — worth reading as the most compact current statement of IIT’s formal ill-definedness and its proxy-versus-approximation problem, from sympathetic insiders; its defences (panpsychism, expander grids) rest on the authors’ proposed reinterpretation rather than on IIT 4.0 as stated.
When seeded, before the reading in NOTE-122, the entry gave a priority for a deeper reading: medium — t=40 s. It is an important corrective reference, but the skim captures the five points, and the technical detail matters only for a dedicated IIT note.
What the seeding asked a deeper reading to check (NOTE-122 is that reading):
- A current, technically informed assessment of IIT from its sympathetic critics. It is the natural reference for any note that cites IIT on AI consciousness (IIT’s substrate focus bears on digital hardware) or engages Schwitzgebel’s LIT-159 critique of the exclusion postulate.
- The §5.1 argument (non-Markovian grainings) is the load-bearing technical claim. It builds on Barrett & Mediano 2019, which a deeper reading should check alongside any IIT 4.0 reply.
- Note the overlap in authorship with Seth (LIT-135). The paper’s charity toward IIT’s panpsychism sits alongside Seth’s biological naturalism, and a deeper reading could check whether the two are consistent.
Access when seeded: arXiv abstract page (v1 submitted Mon 13 Apr 2026; “19 pages, 3 figures”) and the full PDF. Read the abstract, §1 Introduction, all section heads, §4.2 (panpsychism), §5.1 (ill-definedness), §6 Discussion and the start of Box 2 (recommendations). The paper lists no keywords; the keywords above are its own terms.