Flow Lenia · developmental fold onset · 53 fresh organisms

The action world appears before the body.

Immediately before a recognizable form emerges, the system already contains a mature, history-dependent map of possible futures.
We found each organism's first certified visible formation event, then compared score-matched information-directed histories before it, across it, and after it. The body became more folded through formation. The causal response landscape was already there.
The finding

The visible organism is late.

Before the morphology announces an organism, different information-directed histories already send the same-looking state into different causal futures.
0.350pre-formation displacement between matched-history causal futures.
0.348the same displacement while crossing visible formation—essentially unchanged.
0.349after formation. The action-world separation was not created by the visible event.

The strongest reading is not that a finished organism secretly existed in the soup. It is that an organized causal state—a structured map from action to future—preceded the body that later made it visible.

How we looked

Three windows around one birth event.

Beforeq−16 → q−7The last two-pulse window that finishes cleanly before formation.
Crossingq−8 → q+1The same intervention history spans the visible transition.
Afterq+1 → q+9The first matched window beginning after formation.

For every state we built a four-corner square: two final Φ‑R levels, reached through two different two-pulse histories. Within each score level the histories were tightly matched on final Φ‑R, then released and tested against nine later interventions.

Morphological fold

The body keeps deepening through formation.

At release, changing history while holding Φ‑R approximately fixed moves the state about as far as traversing the entire low→high Φ‑R span. That history separation grows before→crossing and continues growing after the visible event. There is no isolated spike; the fold deepens through the transition.

0.120.130.140.150.15History route, matched Φ-RLow→high Φ-R, matched routeBeforeAcross formationAfter
+0.007crossing minus pre; 95% interval +0.003 to +0.011.
+0.003post minus crossing; 95% interval +0.000 to +0.006.
≈1.00×history displacement ÷ low→high Φ‑R displacement at all three states.
Causal future

The action landscape is already mature.

After release we applied the same nine challenges to every corner and followed all 5,724 futures. The route-conditioned future displacement is already about 0.35 in the last clean pre-formation state. It neither peaks nor materially changes when the visible event arrives.

0.310.330.340.360.38Causal-future remappingBeforeAcross formationAfter
visible morphology: still arriving
history-dependent future map: already present

The shift is about one entire local repertoire width before, across, and after formation. The strongest challenge remains the same across matched histories only about 10–14% of the time. The hidden organization is not a small residual—it reorganizes which interventions lead where.

Causal leverage

Before the body locks in, history can move the clock.

The four matched information routes changed when the same morphology event occurred. This leverage is just as large in the last clean pre-window as in the crossing window, then collapses once the event has already happened.

0.000.901.802.703.60Range of formation times across four routesBeforeAcross formationAfter

Pre-formation

Matched histories differed by 1.66 and 1.83 passages on average at the two ending levels.

Across formation

The differences were 1.81 and 2.15 passages. One route did not universally accelerate: the direction depends on ending level and local state.

After formation, the mean four-route range falls by 2.34 passages. That is partly logical—the post-window begins after the measured event—but it cleanly locates the leverage: before and during formation, not afterward.

One organism, untouched

The body barely announces the hidden transition.

This baseline organism's certified event occurs at q31. The images change gradually. Yet the matched causal routes operating just before and across this interval already separate its later future map by roughly one repertoire width.

Last clean pre-window at q16
Last clean pre-windowq16
Crossing window begins at q24
Crossing window beginsq24
Certified visible event at q31
Certified visible eventq31
First post-window at q32
First post-windowq32
One score-matched route pair

Near the event, history can delay the same transition by 15 passages.

For the representative organism above, the two low-ending crossing histories differ by less than 0.9% of the local Φ‑R range at release. One route reaches the certified morphology event at q31; the other reaches it at q46. The images remain deceptively similar, which is exactly the point: the causal divergence is easier to see by intervention than by appearance alone.

Hold history event at q31
Hold historyevent at q31
Loop history event delayed
Loop historyevent delayed
Hold history q46
Hold historyq46
Loop history event at q46
Loop historyevent at q46

This 15-passage example is illustrative, not the cohort average. Across all 53 organisms, route directions vary; what replicates is the existence of timing leverage and history-separated futures.

What this changes

Formation is expression, not origin.

What the visible event does

It marks a reproducible morphological transition. The release-state fold grows through it, and the system becomes visibly more structured.

What it does not do

It does not create the causal repertoire. The history-conditioned action world is already fully separated immediately beforehand.

causal organization → visible formation
not
visible formation → causal organization

That arrow is the bold hypothesis now supported by this experiment's timing: the body appears to be a late readout of an earlier organization in pattern-and-action space.

The next direct swing

Find the first moment the action world exists.

The last pre-formation window was already too late: causal-future remapping was mature. So the next experiment moves backward. At several earlier offsets—while the field still looks like diffuse pre-form material—we will build the same matched history sheets and ask where future remapping first rises above its earliest baseline.

Then we intervene at that onset. If changing the hidden action landscape redirects which body later forms, we will have a direct causal bridge from early information geometry to organism emergence.