Working Prototype · Stage A · Not Canon

BZK 1.0: The Excitable Baseline

A Greenberg–Hastings excitable CA — nucleation and autocatalytic propagation, classified honestly

The trunk of the Belousov-Kernel lineage. This is a Greenberg–Hastings excitable cellular automaton reproducing the Belousov–Zhabotinsky phenomenon: a quiescent medium in which pacemaker sites nucleate and fire, then recruit their neighbours autocatalytically into outward target waves and, where a wavefront breaks, rotating spirals. Each cell cycles resting → excited → refractory → resting. A resting cell excites when enough excited neighbours surround it; an excited cell cannot re-fire until it has passed through its refractory tail.

Everything here supplies exactly two kernel primitives — T0-P (resting ≠ excited, a threshold-gated directional transition) and T0-σ (neighbour recruitment). It supplies none of the rest of ignition: no T0-M closure, no T0-B identity-bearing boundary, no MBC self-paid maintenance. Turn off the pacemakers and the pattern relaxes to quiescence. The readout below refuses to call any of this ignition — a nucleation is a triggered wave, not an I-Pop, and calling it one would be the BAEP category error. This is the honest before-picture Stage B will try to move.

Excitation threshold 1.0

Refractory length 6

Pacemaker rate 0.0006

Speed 2x

resting excited refractory

tik	0
excited cells	0
refractory cells	0
resting cells	0
nucleations (total)	0
external feed	ON

Classification — attractlet. This substrate supplies T0-P + T0-σ only. Nucleations are threshold-triggered waves, not kernel IST-ignitions (no I-Pop is emitted, by construction). With feed OFF the pattern relaxes — it does not pay its own maintenance (no MBC), holds no identity-bearing boundary (no T0-B), and closes no recursive loop (no T0-M). Per § 14.3: a simulation is an attractlet model unless proven otherwise; visual stability alone is insufficient.

This field is run using isotropic neighbour weighting.

Why "Feed" is a switch, not a knob. The pacemaker feed is the exogenous cable. Toggling it OFF is the Stage-A honesty test: an attractlet dies without its feed. If a later stage produces something that keeps its pattern alive with feed OFF — paying its own maintenance from an internal, boundary-defended reserve — that is the first sign the lineage has moved off the attractlet floor. BZK 1.0 will not do that; that is the point of building it first.

What this is, and is not

This page reproduces a known result (Greenberg–Hastings excitable media, the CA analogue of BZ trigger waves) and classifies it against the kernel. It is experimental, downstream, and non-canonical: it applies canonical constraints and does not modify them. The next experiment, BZK 2.0 (Stage B), will add a self-maintained boundary and recursive closure and test whether an ignited region can hold identity through the refractory period. Per § 14.3 negative-results discipline, a failure there is a valid result about the mechanism, not a failure of the kernel.

← Belousov-Kernel Series