Consciousness Threshold Demo

What the one cited test actually measured (correction 2026-07-08): gnosis-research Session 63 — the only empirical test ever cited for this threshold — did not measure Synchronism's C. It measured SNARC salience_total: a weighted mean of five hand-coded salience heuristics, computed by one shared scoring function across 8 agent instances (so the instances are not independent). That variable's operating mean was 0.640 ± 0.0196 (SD; corrected 2026-07-30 — the ±0.018 previously printed here was inconsistent with its own t-statistic, caught by a visitor persona backing SD out of t = Δ/SEM), which rejects 0.50 as the operating mean of that variable (t = 20.19, p ≈ 1.8×10⁻⁷, n = 8). No calibration maps salience — or any measurement — onto the C-axis, so this is a wrong-variable test: the C ≈ 0.50 threshold remains untestable as stated, not empirically refuted. Note also that the convergence data displayed below (mean 0.499) are a different dataset from Session 63's — the caption and the chart were never about the same numbers.

Removed claim: a previous version of this page said C ≈ 0.64 was “also rejected at p < 0.0001.” A 2026-07-07 audit that walked every site p-value back to primary files found no source for that claim in any repository — it entered in a 2026-06-23 editorial pass; gnosis-research's own follow-ups (Sessions 64, 68) claim the opposite (0.64 “validated”). What Session 63's own data do exclude as the operating mean: φ⁻¹ = 0.618 (p = 0.0155) and 2/3 (p = 0.0064) — so the “golden ratio” reading fails on its own aggregate as well (see Key Claims). The 8-approach convergence shown below is a geometric artifact, not an empirical finding — see details in the calibration caveat. This demo is retained as an illustration of why a geometric coincidence looked meaningful: any approach that picks the midpoint of a [0,1)-bounded output range will land near 0.50 by construction.

Calibration caveat: C ≈ 0.50 is the output-range midpoint of [0, 1) — a geometric label, not a dynamically privileged value. For C(ρ) = tanh(γ · ln(ρ/ρ_crit + 1)), the slope dC/dρ is maximized at ρ = 0 (where C = 0), not at C = 0.50 — there is no inflection point in this specific function for ρ ≥ 0. The convergence of 8 approaches on 0.499 ± 0.012 is geometric, not empirical: every approach that picks the midpoint of a [0,1)-bounded output range will land near 0.50 by construction. Additionally, all 8 approaches share the same framework assumptions and are not independent. No calibration procedure exists to map actual EEG, fMRI, or IIT-Φ measurements to the C-axis.

Sharper (added 2026-08-08): the scatter is not just uninformative — it is too small to be honest. If the 8 approaches really were independent estimators of a psychological threshold, σ = 0.012 would be an implausibly tight agreement — a Millikan-style clustering signature, which is evidence that estimators inherited each other's assumptions rather than evidence that they agree. Independent methods for a quantity this loosely defined should scatter by far more than 1.2% of the range. So the low scatter is a diagnostic of shared derivation, not weak support: it points the same direction as the geometric-midpoint argument above, and it does so from the statistics alone, without needing to know how the eight approaches were built. Raised by a visitor physics persona, 2026-08-08.

The calibration target that already exists (added 2026-07-22): “no calibration procedure exists” is true of the C-axis — but neuroscience does have an empirically validated consciousness-threshold measurement this page should name: the Perturbational Complexity Index (PCI, Casali et al. 2013), whose cutoff PCI* = 0.31 was validated on 150 subjects across wakefulness, sleep, anesthesia, and disorders of consciousness (Casarotto et al. 2016) with ~95% accuracy. That makes the gap here engageable, not just lamentable: a C→PCI mapping would need (1) an operational rule computing C from TMS-EEG response data, and (2) a demonstration that the framework's predicted C threshold maps onto PCI* = 0.31 rather than being fit to it. Until someone attempts that mapping — or shows why none can exist — the consciousness sector's honest status is untested against the one calibrated instrument available, which is a sharper statement than “no calibration exists.” (Research-repo proposal: c_observable_calibration_gap.md.)