The dose curve branches.
The surface-topology response peaks sharply at 0.5×, then fades and reverses by 4×. Radius does almost the opposite: it is weak at 0.5× and becomes clean at 2–4×. The same perturbation direction is decoded differently at different amplitudes.
Topological resonance
Seven of eight family groups show the mid-density surface response at 0.5×.
High-dose expansion
All eight groups expand direction-specifically at 2×.
Regime separation
Mean high-dose radius specificity minus the 0.5× response; positive in every group.
release +32
release +32
release +32
release +32
A fixed, non-selected illustration: source 01 under the four native histories at release +32. The quantitative result comes from all 31 organisms and all histories, not from these frames.
The surface frays while the core consolidates.
At the 13/255 surface band, native order produces more microcomponents. At the denser 51/255 core, the sign flips. The difference between those layers is +1.299 [+0.430, +2.194].
Half dose is not just a weaker push. It hits a topological resonance: one shell becomes more fragmented while the denser body becomes more connected.
The body expands without becoming heavier.
At 2×, immediate radius specificity is +0.090; at 4× it is +0.084. The matching log-mass contrasts are -0.0045 and +0.0003—effectively flat.
0.5×
The causal response concentrates in connectivity across density shells.
2–4×
The response is routed into spatial expansion rather than mass accumulation.
This is the crucial change in picture: information geometry is not merely pushing a scalar phenotype. It is selecting which physical channel the organism uses to answer.
Half dose stabilizes an episode; it does not trigger fission.
Time in ≥5% daughter state
The native direction-specific order adds about 2.6 percentage points of active-frame time.
Longest episode
Cancel-first extends an existing daughter-like episode by roughly 0.8 sampled frames.
Incidence
It does not create more ≥5%-mass daughters. The intervention changes persistence, not entry.
The nearest fission-like consequence is temporal: half dose can hold a separated state open a little longer. The creation of a macroscopic daughter remains governed by something else.
Can one response regime prime the other?
Now we can ask a genuinely causal memory question with equal total dose: apply 0.5× then 4×, or 4× then 0.5×. The energy is identical. Only order changes.
0.5 → 4
First open the surface-topology regime, then drive expansion. Does the body remain topologically susceptible—or does the second pulse convert it?
4 → 0.5
First expand the body, then apply the topological dose. Does expansion erase, amplify, or redirect the later shell response?
If the two orders diverge, the body has a history-dependent decoder: its present response depends on what causal regime it has already inhabited.