Lux: One definition sits at the center of everything we've been discussing this series, Hex. T equals Π, L, F, B. Four symbols. Four components. Today we're putting each one under the microscope.
Hex: The formal definition of a theory — or layer — in the emergence calculus.
Lux: And it works like a passport. A layer has to fill four fields before it can cross the border into legitimate physics. Miss one and the passport is invalid. The layer doesn't get in.
Hex: So let's check each field. What's the first one?
Lux: Π — the packaging lens. This is the field that decides what the layer looks like. In Six Birds terms, it's the coarse-graining operation that maps microstates to macro-variables. In the Throw paper's finite setting, Π is implemented as state projections — you take the full microstate and extract the variables that matter at the agent's scale.
Hex: Like the photo on the passport. It shows what this layer looks like from the outside.
Lux: And like a passport photo, it throws away information. You can't reconstruct the original microstate from Π. That's by design. The lens picks what's relevant and hides what isn't. In the agency setting, it extracts four macro-variables: an internal macrostate, an observation at the boundary, a ledger variable, and an outside macrostate. Everything else — all the micro-detail — is coarse-grained away.
Hex: So the photo isn't just cosmetic. It defines the resolution of the layer.
Lux: And different lenses produce different layers. This is one of the key points the Become paper makes explicit. The lens is a modeling choice — it's first-class data, not an afterthought. If you choose a lens that retains too little, you miss important structure. If you choose one that retains too much, you haven't actually coarse-grained. The art is choosing a lens where the macro description is both informative and stable.
Hex: Field two?
Lux: L — the controlled kernel. The dynamics of the layer. For every state-action pair, L tells you the probability distribution over next states. In the finite setting, it's a row-stochastic tensor — one transition matrix per action.
Hex: The nationality stamp. It says what rules this layer follows. Where it's allowed to evolve.
Lux: And here's where the Become paper adds a crucial insight. The dynamics aren't just "the physics." They include the completion rule — the way you lift macro descriptions back to micro when you need to evolve them forward. The Become paper is explicit: the completion is part of the theory. Different completions with the same lens can produce wildly different behavior. A wrong equilibrium family can look stable while producing systematic error at the macro level. Choosing mean-only versus mean-plus-variance changes whether heterogeneity is visible. These are modeling choices with consequences.
Hex: So the nationality stamp isn't just stamped in. You have to commit to specific dynamics — and that commitment changes what objects the layer produces.
Lux: Exactly. The completion rule turns a coarse description into an operational layer. It's the commitment that makes the passport real rather than hypothetical.
Hex: Field three — F.
Lux: The feasibility gate. Which actions are actually available at each state. In the Throw paper, feasibility is ledger-gated: the feasible action set at state s is every action whose cost fits within the current resource level. Mathematically, A-feasible of s equals the set of actions a where cost of a is at most the ledger value at s.
Hex: Travel restrictions. Places you're not allowed to go, no matter how much you want to.
Lux: And the restrictions are structural, not preferential. It's not "you'd rather not go there." It's "this visa doesn't cover that destination." Infeasible actions don't exist at the layer's scale. The paper uses the phrase: infeasible interface commands are not "actions" in the induced layer. The feasibility gate erases them from the agent's world.
Hex: That's a hard boundary. Not a soft penalty.
Lux: Which is what makes the layer honest. If everything were feasible, the action space would be unrealistically large. The feasibility gate — which implements P2, constraints, and P6, accounting — carves the space down to what the agent can actually do given its resources. This is where the budget meets the dynamics.
Hex: [nodding] And field four?
Lux: B — the budget. The ledger dynamics. What things cost, how resources flow, how spending changes the state. In the finite ring-world, the budget is a nonnegative scalar extracted from the state. Each action has a nonnegative cost. When you act, you spend. When you repair, you might replenish. The ledger tracks the balance.
Hex: The expiry date on the passport. How long the layer can sustain itself before it runs out of resources.
Lux: Without a budget, every action is free. Without costs, the feasibility gate has nothing to check. The budget is what makes the layer realistic — what turns a mathematical construction into something that faces constraints. And the budget interacts with the viability kernel. If costs are too high relative to income, the viability kernel shrinks. If costs are manageable, the kernel stays large. The budget quantifies the practical limits of persistence.
Hex: Four fields filled. Is that enough?
Lux: Not quite. Filling all four fields is necessary but not sufficient. The passport also has to be valid. In Six Birds terms: the layer has to be coherent.
Hex: Meaning?
Lux: Idempotence. Apply the packaging operation — use the lens, evolve through the dynamics, repackage — and do it again. If the second application gives the same result as the first, the layer is coherent. The macro description is stable under its own closure.
Hex: And if it's not?
Lux: The idempotence defect tells you how far off you are. A high defect means the macro labels aren't behaving like real objects. The passport has all four fields filled, but the photo is blurred — the layer isn't actually functioning as a self-contained theory.
Hex: So coherence is the validity check. You can have all the right information on the passport and still fail inspection if the photo doesn't match the face.
Lux: The Become paper gives explicit failure regimes. Moment closures that look stable but produce systematic macro error because the equilibrium family is wrong. Lens choices that omit slow variables, erasing information that matters on the relevant timescale. LES filtering where evolving then filtering gives a different answer than filtering then evolving — that's route mismatch, and it's structural. These aren't bugs. They're diagnostic opportunities. The defect tells you where the layer breaks.
Hex: Now — does this passport format show up only in the agency setting?
Lux: It shows up everywhere. The same four-field structure — lens, dynamics, feasibility, budget — appears in every Six Birds paper. In the Throw paper, it produces agent layers. In the Become paper, it produces life-like maintenance layers. In the Plot paper, it produces geometric layers. The specific contents of each field change, but the passport format is universal.
Hex: Same form, different stamps.
Lux: And in continuous systems, the core Six Birds paper shows the same structural idea through dissipative atoms. An atom has an internal Hilbert space, input and output maps, and a semigroup whose norm decays exponentially. The decay rate is the continuous analog of the finite convergence we see when the viability kernel stabilizes. The balanced coupling condition — input-output coupling bounded by the decay rate — is the continuous version of the budget constraint.
Hex: So the passport generalizes even to infinite-dimensional systems.
Lux: The four fields adapt, but the structure persists. Lens, dynamics, constraints, costs. Whether you're in a finite ring-world or a continuous semigroup setting, the layer has to show all four.
Hex: And what does the passport buy you? Why insist on all four?
Lux: Because the definition is the technical hinge that separates enablement from causation. The paper's discussion section is explicit about this. Enablement is about making the passport valid — filling the four fields so that a layer exists with stable macro-variables and a functioning feasibility gate. Causation is what happens after the passport is stamped. Once the layer exists, you can meaningfully ask whether changing action a changes the distribution of outside future y. Before the layer exists, that question is a type error.
Hex: So the definition isn't just organizational. It's what prevents you from asking causal questions in a vacuum.
Lux: The schedule trap is the cautionary tale. If you don't check the passport — if you skip the enablement step and go straight to causation — you can treat an external schedule as agency and fabricate empowerment. The definition forces you to build the layer first. Lens, dynamics, feasibility, budget. Then check coherence. Then, and only then, ask about difference-making.
Hex: Four fields. Lens, dynamics, feasibility, budget. Miss one and the layer doesn't pass inspection.
Lux: And every layer in the Six Birds program has to show its passport. The emergence calculus doesn't do border-free travel. If you want legitimate physics at an induced scale, you fill the form.
Hex: Passport stamped. Layer approved.