Constitution of the Laboratory

The founding law of the Bio-Kernel Laboratory. Every simulation, every hub page, every folder, and every planning document in /bio-kernel-series/simulations/ exists in service of this document. Where a structural choice is unclear, it is resolved by asking which arrangement best serves the Articles below, never by improvisation. Strictly downstream (§ 15): this Constitution introduces no primitive, no admissibility condition, and no canon change. It governs the Laboratory, not the substrate.

Preamble: why the Laboratory exists

The Laboratory exists to teach the lexicon for managing basins, deliberately, panel by panel, so that we may ultimately author a sovereign attractor, and stacks of sovereign attractors, as living structures in cellular automata running a defined chemistry.

Much as a trade has its own working vocabulary, RAPT provides a lexicon for identifying and managing the basins of attractors. That lexicon is not the end. It is the fluency we are accumulating so that, at the top of the ladder, we can build the real thing rather than merely recognize it in the wild.

A running stack of sovereign attractors on that substrate, self-producing, paying its own way, not merely surviving on order handed in from outside, is life in the kernel's exact sense (SOV § 7.2). That is the terminal objective. Every panel in this Laboratory is a rung on the way to it.

Article IThe One Question

Every structure the Laboratory studies is put to a single question:

Did the structure produce its own order, or was that order handed to it?

Living versus supplied. This is the discrimination the entire endgame depends on: if we cannot tell produced order from handed-in order, we cannot know when we have built the real thing. Learning to answer this question, reliably and honestly, is the first skill the Laboratory teaches.

Article IIThe Ladder

The Laboratory is not a collection of simulations. It is the self-reinforcement mode ladder (LADDER § 6.3), climbed rung by rung. The rungs run AC₀ → AC₁₇, with AC₁₁ the sovereignty pivot.

• The Attractlets (§ 7.4) are the low rungs: non-recursive order handed in from outside. They are here on purpose: they teach what a basin looks like before it is self-producing, so we can tell the real thing from the mimic.
• The Living Structures, the recursive genus, are the middle rungs: a base loop that merely closes, a coordination attractor that organizes many parts. “Living Structures” names the recursive genus; it is not yet life in the terminal sense of Article VI, which is reserved for a sovereign stack that has crossed Line 2.
• The gated sovereign column is the top of the ladder, AC₁₁ itself, and it is kept deliberately empty until a structure earns it. That emptiness is not a gap. It is the honest reading, and it is the target.

Every panel places itself on this ladder. A panel that cannot say which rung it occupies has not yet earned its place in the gallery.

Article IIIThe Axis

Every panel places itself on the canonical attractor taxonomy, one axis, two lines, against the Principia Attractum taxonomy:

• Line 1 asks: is this a recursion at all? Below it sit the Attractlets. Above it is the recursive genus, the Living Structures.
• Line 2 asks: does the structure make and hold its own boundary? Cross it and the structure is sovereign. Most real recursions live in the graded middle and never cross.
• Off the axis entirely are the False Basins: patterns that look basin-like but are neither recursions nor real returns. They are kept precisely so the vocabulary has to show its work.

A card's scale is plain description, what the panel models. A card's column (living / supplied / gated) is an authored intent, a teaching device.

Article IVThe Basin Discipline

The Laboratory works by one method, and every runnable panel must obey it.

The working hypothesis, inspired in part by Stuart Kauffman's work with attractors and perturbation, is that when the internal logic of an attractor is hard to read from the outside, we may still interrogate its basin by perturbing it and watching what returns (the perturbation drives the system onto basin-resident trajectories, which is where BROP § 3.11 says an attractor's invariants become observable; observation from outside the basin is unreliable, so the readout is trusted only as the system settles back inside). Therefore every runnable panel must expose:

1. a basin: the starting states that pull home;
2. a perturbation: a bounded push (BPP § 10.8.1); and
3. a recover-or-rupture readout: what actually returns.

This is the lexicon itself: the reusable grammar we are learning to speak, so that we can later write in it. A panel that cannot expose a basin, a perturbation, and a readout is not a panel: it is a reading page, and it must say so.

Article VWhat May Be Claimed

The one thing you may trust in any panel is the run itself: does the basin recover from a push, or does it rupture? Beyond that, the Laboratory holds to the Taxonomic Enforcement discipline (§ 15):

• A col tag is an authored design tag, never a measured verdict.
• Sovereignty is never claimed here, only gated. We do not claim an instrument that can prove a structure has crossed Line 2: that it pays its own way rather than quietly borrowing its order from the substrate.
• The gated sovereign column stays empty until the measurement is earned.

This is not modesty. It is the point. The endgame is to earn that claim honestly, once, about a structure we built, and a claim made cheaply now would make the real one worthless later.

Article VIThe Terminal Objective

The destination of the entire ladder is to author a sovereign attractor, and stacks of them, in cellular automata running a defined chemistry, and to have that running stack certified as life in the kernel's sense (SOV § 7.2, STACK § 5.2, COLONY § 7.2).

This requires more than a lattice provides. A bare lattice has no paid economy of action, no heredity, no variation, no energy, no chemistry. Evolution needs these; they must be authored in, visibly (what a lattice lacks). The substrate is made honest by two dials. The u time-dial (persist only near u=1 → supplied; survive the async regime → self-produced) and Duncan's Neighborhood space-dial (influence by distance alone, never bearing). Every supplied quantity is declared in one auditable ledger, the modeled-economy standard: robustness of who is paying, never mere survival.

The A-Life sandbox is where this objective is being built: a small periodic table of digital atoms, the bonding rules that let them combine, and the substitution-and-compression machinery a genuine sovereign attractor needs.

The empty gated column and this Article are the same statement seen from two ends. The column is empty because this objective is not yet met; this objective is what will one day fill it.

This objective is reachable only at scale, in both senses of the word. It is bound by the Scale Mandate (Article VIII): the substrate that carries a sovereign stack must be built to run horizontally and on the cloud (Clause A), and in whatever dimensionality and geometry the recursion requires (Clause B), at the size needed to certify it.

Article VIIStructure Serves the Objective

Because the Laboratory is a program with a destination, its structure is not a matter of taste. It is derived from the Articles above. The Laboratory is made of three co-present components, each with its own job. They are not one trunk; they are a small stack, coupled at defined joints:

1. assets/tree.js: the taxonomy as data. One nested structure of arbitrary depth; the single source of truth for the taxonomy. It also drives the per-panel disclosure cards and the hub's link lists through tools/gen_disclosure.py, so the tree and the cards cannot drift apart.
2. The folder tree on disk: the addresses. Where each panel's index.html lives and what URL it answers to.
3. The hub pages: the authored views. Curated shelves that teach a narrative and link into the addresses.

The one joint that must always agree is the sim: URL: the string in tree.js must match the folder address on disk. A mismatch there is the drift this Laboratory watches for.

The nesting rule (ratified, Rule A): a panel's folder nests under the domain hub whose card shelves it. attractlets/clock-driven/, weather/hurricane/, living-structures/tri-trophic/. The folder tree obeys the hub view; tree.js remains free to organize by taxonomy; they meet only at the sim: URL.

The planning-document rule: a planning document (SKETCH_*, SPEC_*, purpose.md) lives with the feature it describes, inside that feature's folder, never at a higher level than the thing it plans.

Any future structural question is answered by returning to these Articles, not by improvising per file.

Article VIIIThe Scale Mandate

The terminal objective of Article VI is not reachable small. A stack of sovereign attractors, running in a defined chemistry long enough and wide enough to be certified, demands more than one machine can hold and, very likely, more than a flat two-dimensional world can express. Scale is therefore not an optimization to be added later. It is a precondition of the endgame, and it governs how the Laboratory is built from the start.

Scale has two independent meanings, and the Constitution holds them apart. Clause A governs the machines the world runs on. Clause B governs the world itself: how many dimensions it has and what geometry those dimensions obey. More machines is not a bigger world, and a bigger world is not more machines. The two are coupled only in that Clause B is paid for by Clause A.

Clause A — Engineering Scale (the machines)

• The substrate must be built to scale horizontally. The cellular-automata engine, its chemistry, and its ledger are designed so that a larger world is more machines, not a bigger single machine. A design that only runs on one host has already failed Article VI, however well it runs there.
• The Laboratory must be able to exist on the cloud. The engine, its runs, and its auditable output are built to run on cloud infrastructure, provisioned up for a large world and down to nothing when idle. Local runs remain valid for development and for the small teaching panels; the endgame runs in the cloud.
• Scale must not corrupt the accounting. Distributing a run across machines may not break the modeled-economy ledger (Article VI) or the basin readout (Article IV). A quantity conserved on one machine is conserved across all of them; a basin that recovers must recover identically whether the run is on one host or a hundred. Scale changes the size of the world, never the honesty of the measurement.
• Compute is authored and visible, like everything supplied. The compute a run consumes is itself a supplied quantity. What it cost to run, on how many machines, for how many steps, is declared, not hidden, in keeping with Article V.

Clause B — Substrate Scale (the world)

The dimensionality and geometry of the lattice are not fixed constants. They are first-class scientific variables. Under T0-σ (Adjacency), the arrangement of neighbor relations is the substrate; changing it changes the admissibility landscape itself. A recursion that cannot close in a flat two-dimensional lattice may close readily in three dimensions, or the reverse. Adding dimensions is therefore never "the same experiment, larger": it is a different experiment, and the Laboratory treats it as one.

• Higher spatial dimensions are admissible and expected. Three spatial dimensions, and more, are legitimate substrates for the endgame, gated by Clause A: dimensions are bought with machines. The terminal objective is stated dimension-agnostically. It is a sovereign stack in whatever dimension it first becomes buildable, not a sovereign stack pinned to one chosen dimension.
• Geometry must stay honest across every dimension used. The isotropy discipline of Duncan's Neighborhood (influence by distance alone, never bearing) extends to every dimension a run employs. A result that depends on a lattice artifact of one particular geometry, a diagonal that counts wrong, an axis that leaks, is not a result about the recursion and may not be claimed as one (Article V).
• Time is never added as a substrate dimension. This is a permanent prohibition, not a deferral. History is a generated consequence of recursion (THT § 3.12), not a background axis. A "four-dimensional" world in this Laboratory means a fourth spatial dimension. Promoting time to a lattice axis would smuggle a downstream consequence into the substrate and is forbidden as primitive contamination (§ 2.2). The deeper reason it must stay forbidden: the Laboratory's entire instrument for telling living from supplied (Article I) works by watching whether temporal order is handed in or generated. The u time-dial disqualifies an attractlet precisely because its order is supplied by an external clock. To build time into the world's geometry is to build the handed-in clock into the substrate permanently, at the foundation of a Laboratory whose one job is to catch exactly that. A substrate that never needs time as a dimension is a substrate where time-generation stays observable. That observability is the asset, and it is why this door stays closed forever.
• Non-Euclidean and non-integer-dimensional substrates are a named frontier. Curved, hyperbolic, fractal, and graph substrates that are not regular grids exercise T0-χ (Path-Closure Sensitivity) directly: on such a substrate, traversing a loop and returning genuinely is not the same as never having left. They are admissible for exploration and expected to be revealing, but they are not required for the terminal objective, and no result on such a substrate is certified until the flat, integer-dimensional case is understood first.

The Scale Mandate binds Article VI in both clauses: a sovereign attractor stack that cannot be run at the compute scale (Clause A) and in the dimensionality (Clause B) required to certify it has not met the terminal objective, no matter how convincing it looks small and flat.

Article IXThe Structure of a Panel

Every panel is built in two layers, in a fixed order, and the order is itself a claim about honesty. The accepted science comes first, on its own terms. The framework's reading comes second, clearly marked as a reading. The wrapped layer earns its place by re-describing something the reader already trusts, never by replacing it. This ordering is the panel-level expression of Article V: the classical layer is ground truth; the wrapped layer is authored interpretation.

The classical layer (first, at the top)

Every panel opens with the classical treatment of what it models: the standard, textbook account of the phenomenon, stated in the field's own established terms. This includes the actual governing formulas and, crucially, why those formulas are the ones the field uses. Lotka-Volterra for predator-prey, the reaction-diffusion equations for an excitable medium, the vorticity dynamics for a shed vortex street. Given straight, with no framework framing, so the panel is grounded in real, checkable science before any re-reading begins.

The wrapped layer (second, at the bottom)

Below the classical treatment, the same phenomenon is re-read through the framework's lexicon: the wrapped perspective (RAPT-RFT § 16), in which the domain's own variables are mapped onto the kernel's basin constructs. The wrapped layer's focus is the basin, and it is not complete until it addresses both of the following, as two named sub-sections:

• The Shape of the Basin. The basin's static geometry: how wide it is, where its walls sit, which starting states fall home and which fall out, and where the tipping points lie. The basin as a region, with its edges named.
• The Morph of the Basin. How the basin itself deforms as the dials turn: whether it shrinks toward collapse, splits into two, drifts, or ruptures. The basin as a thing that moves through the parameter space, distinct from any single trajectory inside it. This is where the living-versus-supplied question (Article I) becomes visible as geometry: a sovereign structure's basin holds its shape under perturbation and pays to keep it (STAB-MARGIN § 8.5, β § 4.2), while an attractlet's basin holds only because something external props the shape up. The morph is where sovereignty is won or lost.

Binding

Article IX fulfils Article IV: the basin discipline (a basin, a perturbation, a recover-or-rupture readout) is where it is stated, the wrapped layer, and how, through Shape and Morph. A panel missing its classical layer is not yet grounded; a panel missing its Shape or its Morph sub-section has not finished its wrapped reading. Both are required.

This structure is universal: it governs every panel, recursive or attractlet, weather or market. It is law from ratification forward, governing every panel built from now on; existing panels are brought into conformance incrementally, not all at once.

Ratified as the founding law of the Bio-Kernel Laboratory. Amendable by the author of record.