Precomputed result
In the near-maximal network of the IIT 4.0 paper, a four-unit lattice excludes the integrated six-unit whole
| Substrate | Six binary units A–F with states in {−1, 1}, each a logistic of its weighted inputs (Eq 60) with k = 4, all-to-all connected. The strongest connections are A→B, B→F, and E→A (0.8), C→D (0.6), F→E (0.42), and D→E and F→C (0.4); the body gives the full weight matrix. PyPhi ships the substrate, with the weights supplied by the authors, as pyphi.examples.iit4_2023_fig6e_substrate. |
| State | (1, −1, −1, −1, −1, −1) — Abcdef, unit A ON |
| φs | 0.2473031161096712 |
| Minimum partition | For the first complex Abef, the directed partition severing B→A, B→E, F→A, and F→E. For the second complex cd, the cut of D→C. The same under both formulations. |
| Formalism | iit-4.0-2026 |
| PyPhi | 2.0.0rc4 |
| Script | albantakis-2023-fig6e-near-maximal.py |
The weights, from the unit in each row to the unit in each column:
| from \ to | A | B | C | D | E | F |
|---|---|---|---|---|---|---|
| A | 0.06 | 0.8 | 0.02 | 0.02 | 0.02 | 0.08 |
| B | 0.02 | 0.02 | 0.08 | 0.06 | 0.02 | 0.8 |
| C | 0.02 | 0.06 | 0.22 | 0.6 | 0.08 | 0.02 |
| D | 0.08 | 0.02 | 0.26 | 0.22 | 0.4 | 0.02 |
| E | 0.8 | 0.08 | 0.02 | 0.02 | 0.06 | 0.02 |
| F | 0.02 | 0.02 | 0.4 | 0.08 | 0.42 | 0.06 |
Albantakis et al. (2023) use this network to show exclusion at work on an integrated whole. From the extrinsic perspective the six units behave as one integrated system. From the intrinsic perspective only the four-unit subset Abef satisfies maximal irreducibility, and the remaining units form a second complex that serves as background conditions for the first.
| candidate | φs, Fig 6E | 2023 | 2026 |
|---|---|---|---|
| Abef (first complex) | 0.27 | 0.2702 | 0.2473 |
| cd (second complex) | 0.08 | 0.0800 | 0.0614 |
| Abcdef (excluded) | 0.15 | 0.1520 | 0.1520 |
The intrinsic information requirement of Mayner et al. (2026) lowers φs of both complexes but leaves the full system unchanged. Abef still exceeds the full system, so the exclusion and the two complexes are the same under both formulations.
| quantity | Abef, Fig 6E | Abef, computed | cd, Fig 6E | cd, computed |
|---|---|---|---|---|
| distinctions | 8 of 15 | 8 | 3 of 3 | 3 |
| Σφd | 2.4763 | 2.4764 | 0.5176 | 0.5176 |
| relations | 67 | 67 | 7 | 7 |
| Σφr | 0.7643 | 0.7643 | 0.8017 | 0.8017 |
| Φ | 3.24 | 3.2407 | 1.32 | 1.3194 |
The Φ-structures are the same under both formulations. The distinctions of Abef are the four first-order distinctions A, b, e, f and the higher-order distinctions be, bf, ef, and bef; the script lists their purviews and φd.
The script uses the default configuration of each formulation. To reproduce, run the script with uv, which installs PyPhi 2.0.0rc4 for the run:
uv run --with "pyphi==2.0.0rc4" python canon/albantakis-2023-fig6e-near-maximal.py
Sources
- Albantakis et al. (2023). Integrated information theory (IIT) 4.0: Formulating the properties of phenomenal existence in physical terms. PLOS Computational Biology. https://doi.org/10.1371/journal.pcbi.1011465 — Fig 6E and its caption; Results and discussion, "Near-maximal substrates, extrinsic entities, and exclusion" (iit:ref/albantakis-2023b)