The surface, not the size

Start not with how big an agent is, but with how much it touches. As your reach grows — more of the world in your model, more people in your life, more of what’s newly possible within your grasp — the thing that grows fastest isn’t your insides. It’s your surface: the edge where you meet everything you might combine with, learn from, or be thrown by. And that edge doesn’t grow at the same pace as your reach. It grows faster.

This isn’t the book’s discovery, and the honest thing is to borrow it in the open. A smooth surface grows as the square of its radius; a rough, crinkled, fractal one grows faster still, by an amount that measures just how crinkled it is — that’s Benoit Mandelbrot’s geometry. Living bodies live off exactly this trick: the lung and the bed of capillaries fold an enormous trading surface into a small space by branching, and the body’s rates rise with its size in the lawful, less-than-proportional way that West, Brown, and Enquist made famous. Count relationships instead of surface and the same thing happens, only combinatorially: connect n people and the pairs between them climb as n squared (that’s Metcalfe), the possible groupings faster still (that’s Reed). A city of a million doesn’t bustle a million times more than a town of one; it bustles disproportionately more, which is exactly why each person in it produces more, not the same (Bettencourt and West). Whichever way you count — by area or by relationship — the surface where an agent meets its world, its frontier with what Stuart Kauffman called the adjacent possible, climbs faster than the agent carrying it.

Hold on to that one fact. Everything here follows from it.

Two consequences of one surface

The surface is where two opposite things happen at once — and the whole argument turns on their being the same surface.

The first thing is wonderful. Every patch of that frontier is a chance for synergy — a combination worth more than its parts, a question you couldn’t have asked yesterday, a tool that springs open a dozen others. If the surface grows faster than your reach, then so does the supply of those combinations: a little more reach exposes a lot more pairings. It’s the homely truth that the more you know, the more you can ask, stated as a law. Widening isn’t just adding. It’s generative, and it compounds.

The second thing is the price of the first, and it lands on coherence. Every patch of surface that offers a combination also brings something to be integrated — squared, as far as it bears on the rest, with what you already value and already know how to do. Left unintegrated, a new perspective isn’t synergy at all; it’s just noise, or flat contradiction. And making a surface hang together, when its pieces all bear on one another, is work — chemical in a cell, mental in a mind, organizational in an institution. That work grows with the surface, and can grow faster than it, because what has to be made consistent isn’t the parts but the multiplying relationships among the parts. A pair has one relationship to keep straight. A village has thousands. A civilization has a number with no friendly name.

A caution belongs right here, and it tightens the claim rather than dressing it up. The fact that a surface grows faster than its size doesn’t, by itself, prove that keeping it coherent grows harder than its size — for a smooth body, the surface actually shrinks relative to the inside as the body gets bigger. The burden only really lands if you add one more reasonable assumption: that an agent’s capacity to integrate grows more slowly than the surface it has to integrate — that there’s some ceiling on attention or bandwidth. Grant that, and the cost outruns the agent. Without it, the claim is the milder one. So what follows is a strong tendency under a stated condition, not an exact law — and there’s no single magic exponent to quote: more than the square when the surface is rough, more again when it’s structured, and dependent on how finely you choose to look in the first place.

Why narrowing is downhill — and blind

Now the payoff begins. Put an agent under that rising cost — its surface, and the labor of holding it together, both climbing faster than its means — and ask what it’ll be tempted to do. There are exactly two ways to get coherence back when the work of integration outruns you. One is to do the work: take more in, and carry the rising cost of holding it together. The other is to stop taking things in — wave off the inconvenient fact, shut out the discordant voice, seal the border, and let consistency come back for free because nothing’s left inside that could disturb it. The second way is cheaper. It is, in fact, the cheapest move there is — always, at every scale.

This is the counter-dynamic, which Part 1 named as the formal mark of the immoral: coherence won by narrowing instead of widening. What the surface adds is the reason it’s so hard to shake. It isn’t just a temptation that weak agents give in to. It’s the path of least resistance — the downhill direction on the cost curve. The cult, the sealed ideology, the hardened faction reach their flawless inner agreement by the one method that’s always within reach: caring about less. In the geometry of scale, the cheap thing and the wrong thing point the same way.

But cheapness is only half the charge, and the smaller half. Narrowing isn’t only the inexpensive move; it’s the blind one. To collapse a varied “we” into a single bet — one doctrine, one method, one allowed answer — is to stop exploring. And here a finding from the study of how communities learn proves its worth. Kevin Zollman and those who followed him turned up something that sounds backwards at first: a community of inquirers that’s loosely connected, where pockets explore more or less on their own, often reaches the truth more reliably than one wired tightly together. The tightly wired group rushes to agree on the first answer that looks decent and drops the alternatives too soon. The loosely wired one protects what they call transient diversity — it lets the minority bets run long enough to prove themselves. So a monoculture isn’t only unfair; it’s a worse learner. It has thrown away the very variety that let it find better.

One condition fences this in. The edge that the loose, varied network has isn’t universal. On an easy problem — where the right answer is plain and the wrong ones are quickly seen to be wrong — tight connection wins, because there’s nothing to protect and speed is everything. The varied network shows its value on hard problems: rugged, noisy, deceptive ones, where the space between a good answer and a better one is small and easy to misjudge. And moral questions are the hard kind almost by definition — many goods at once, fine-grained, tangled up with context, where the better path is rarely obvious and being certain too early costs a lot. So the result applies here, where it counts — but exactly because the ground is rugged, and only because it is.

Put the two halves together and the verdict gets sharper than either alone. Narrowing is cheaper now and blinder later: it buys today’s tidiness by giving up the search that finds what’s better. The counter-dynamic isn’t just a habit the framework frowns on. It’s the standing, structurally favored, doubly broken shortcut — and naming why it’s favored is the first step in resisting it.

The answer is a shape

If narrowing is downhill, then a framework that asks for widening is asking agents to climb. So the question this essay has to answer is: how do you make the climb bearable without taking the cheap way down? And the answer isn’t a pep talk. It’s architecture.

Go back to the body. A body doesn’t keep its trillions of cells alive by running a private wire from the heart to each one — the wiring would cost more than the body, and no heart could keep that many addresses straight. It nests. Blood moves through a few great vessels into many smaller ones into countless tiny capillaries, each branch serving a region that mostly serves itself, the whole vast trading surface fed through a tree of branchings that no single channel could ever manage. The cost of supplying an enormous surface is made bearable by not supplying it all the same way — by building it out of nested parts, each minding its own neighborhood, each joined to the next at a seam small enough to handle.

This is what Chapter 5 called earned modularity, now with its engine shown. You make a large “we” cohere not by reconciling every member with every other — that’s the cost that explodes — but by letting members cohere into small wholes, and those into bigger ones, each whole keeping its own house in order and answerable only to the few wholes it actually touches. The labor that was climbing faster than the agent gets broken into pieces small enough that each stays bearable, and then the pieces are stacked. The curve bends: what threatened to grow with the square of the parts grows, under nesting, far more gently. And this isn’t a clever trick imposed from outside. It’s what evolution itself stumbles onto whenever connections are costly. When networks are selected to work well and to spend little on their wiring, they reliably come out modular and layered — the cost of connection is the very pressure that carves the joints (Clune, Mouret, and Lipson; and, for the stacking of modules into levels, Mengistu and colleagues). Nature pays the cost of scale the way a circulatory system does: by branching.

Nesting is also a search

Here’s where the payoff the opening promised actually arrives. Nesting was brought in to make the cost bearable. But the very same structure also fixes narrowing’s other flaw — its blindness — and seeing that is what turns nesting from a grudging economy into the engine of growth.

A network of semi-independent nodes is, by its nature, a search run in parallel. Each node holds its own slightly different picture of what matters and what works, so each is a standing experiment in how to live. And because the nodes are loosely rather than tightly tied together, they don’t all stampede toward one answer. Some can explore while others exploit — some pushing into the new while others polish the known — and the structure never has to pick one global setting, because it’s running both at once in different places. That’s the protected variety from the last section, now a permanent feature of the design rather than a lucky accident; modular systems shed exploration on their own, one part’s ordinary fidgeting raising fresh possibilities for its neighbors. And the wiring can breathe: nodes search in relative isolation, and when one hits on something that genuinely works, the network links up to spread it — a loosely-connected phase for discovery, a tightly-connected phase for locking the winner in. Students of complex systems call that alternation dual-phase evolution; managers call it ambidexterity. The winners get taken up outward, over a widening scope — which is exactly the convergence the tree of agreement describes, now with a motor under it. Agreement isn’t only traced back to a shared root; it’s found, by parallel trial, and carried outward by plain success.

So coherent pluralism — many coherent nodes, varied at the edges, joined at the center — isn’t a grudging concession the framework makes because it can’t force everyone to agree. It’s the generative form of widening coherence: the shape that carries the cost of scale and runs the search that makes scale worth having. The lived, civilization-sized version of this — how a whole society holds itself as a network of networks without either flattening into a monoculture or shattering into noise — is the work of Part 2, and it stays there. Here the claim is only about structure: the shape that answers the cost answers the blindness too, and it’s the same shape.

What the shape costs

Nesting is the answer — and it would be cheap coherence all over again to pretend it’s a free one. It isn’t. It doesn’t abolish the downhill pull; it relocates it. And where it goes is worth naming, because that’s where this architecture breaks when it breaks.

The widen-or-narrow choice doesn’t vanish under nesting. It comes back, at every seam. A module holds together inside by drawing a boundary — and a boundary can be a membrane or a wall. Kept porous — taking in what presses on it, answerable to its neighbors — it’s the honest local coherence the whole structure leans on. Sealed — treating its border as a reason to ignore whatever lies past it — it’s the counter-dynamic again in miniature, a little closed circle inside the big open one. So the branching economy that makes widening bearable also makes the temptation branch: the same downhill move is available at every level, in every node, all the time.

And the thin seams that make nesting work rest on compression. A level can treat the level beneath it as a single piece only because that piece sends a simplified summary of itself upward; the higher level acts on a lossy picture of the lower. That lossiness is what keeps the seam thin — and it’s also what lets it deceive. The summary can lie: the representative who stops representing, the abstraction that leaks, the measure that, once it’s made the target, quietly stops tracking the thing it was meant to. That last one is Goodhart’s law, met already in the essay on measuring coherence, and it’s the signature failure of every nested system — the number that travels up the hierarchy and gets optimized in place of the thing it once stood for. None of this refutes nesting; nesting is still the only bearable way to hold a large coherent thing together. It just marks where the watchfulness has to go: not at the size of the structure but at its seams, where membranes harden into walls and summaries curdle into lies.

There’s also a question of grain — how big each node should be — and economics answered it long before the framework asked. A boundary is worth drawing right where the cost of coordinating inside it would start to exceed the cost of coordinating across it. That’s essentially Ronald Coase’s account of why firms have edges at all, stretched from markets to any nested order — and it’s the same instinct the political principle of subsidiarity expresses when it puts a decision at the lowest level that can competently hold it. A node too big drowns in its own internal bookkeeping; a node too small can do nothing on its own and shoves all the cost out onto the seams. The right grain sits between, near the poised region complexity folks call the edge of chaos — ordered enough to remember and act, loose enough to keep exploring. And that region is a direction to tune toward, not a knife-edge every healthy system has to balance on: not every adaptive system sits right at criticality, and plenty of working ones sit comfortably toward order. The edge of chaos names a target, not a law.

Two gradients

Step back, and the whole picture resolves into two opposing slopes — and the moral of the essay is which one wins, and where.

One slope runs downhill, toward the silo. Call it the cost gradient: narrowing is always the cheaper way to get coherence back, cheaper now and blinder later, and it tugs every agent under strain toward the sealed circle. This is the counter-dynamic, and the surface guarantees it never goes away. It’s a permanent feature of the landscape any widening agent has to cross.

The other slope runs uphill, toward the open world. Call it the gradient of generativity and search: the synergy that only combination yields, the better answers that only protected variety finds — the surplus that narrowing shuts off and only widening can reach. Left to the cost gradient alone, every agent would slide into the silo. What keeps the world from sliding is that the second slope is real too — and that nesting lowers the wall on it, bending the cost of widening down far enough that the generative payoff can win the race. The framework’s optimism isn’t a faith that agents are good. It’s the structural claim that the architecture which makes widening bearable is the same architecture that makes it pay off — so that, with the right shape, the uphill direction is also the rewarding one.

That’s why the counter-dynamic is permanent but not sovereign. It’s always available, always cheaper in the moment; it can never be abolished, only out-competed — seam by seam, by a structure that makes the wider path worth the climb. And that competition, run at the scale of a whole civilization, is the subject the next part of the book takes up under the name of coherent pluralism. The mechanism is here. The practice is there.

What is borrowed, and what is the framework’s own

It’s worth being exact about the accounting, because nearly every separate piece of this essay is borrowed. The contribution is the assembly.

Borrowed: the scaling of surface with reach (Mandelbrot; West, Brown, and Enquist; Bettencourt and West; Metcalfe and Reed) and the frontier of the adjacent possible (Kauffman); the cost of organization, paid in energy (England; Chaisson); the connection-cost origin of modularity and hierarchy (Clune, Mouret, and Lipson; Mengistu and colleagues); the explore–exploit tension and the worth of transient diversity (March; Zollman); the breathing topology of discovery and consolidation (dual-phase evolution; organizational ambidexterity); the economics of the boundary (Coase; the principle of subsidiarity); and the language of coherent pluralism (Hasok Chang; Michael Jackson). None of it is the framework’s own — and saying so is a gain in credibility, not a loss.

What the framework adds is the use. It takes that super-quadratic surface and reads it as a moral frontier, coupling synergy and integration-cost as two faces of one geometry. It gives the counter-dynamic — which Part 1 stated as a definition — a structural and an epistemic explanation at once: narrowing is the cheapest way to get coherence back and the blindest, downhill on cost and bankrupt on search. And it reads the nested network as the moral architecture, the one shape that carries the cost and runs the search — so that the framework’s central demand, coherence that grows by widening rather than by narrowing, has, at last, a mechanism and not only a name. The scaling laws were lying around in a dozen fields. What wasn’t lying around was the claim that they describe the shape of a moral life held together at scale.

Sources & further reading

This essay engages its literature directly rather than through the book’s per-chapter end notes. A citation-level pass is still owed.

The scaling of the surface. Benoit Mandelbrot on fractal dimension; Geoffrey West, James Brown & Brian Enquist on allometric scaling from fractal exchange networks; Luís Bettencourt & Geoffrey West on superlinear urban scaling; Metcalfe’s and Reed’s laws on the value of networks; Stuart Kauffman on the adjacent possible.

The cost of organization. Jeremy England on dissipative adaptation; Eric Chaisson on energy-rate-density.

Earned modularity. Jeff Clune, Jean-Baptiste Mouret & Hod Lipson, “The evolutionary origins of modularity” (2013); and the evolution of hierarchy under connection cost (Mengistu and colleagues).

Exploration and diversity. James March, “Exploration and Exploitation in Organizational Learning” (1991); Kevin Zollman on the epistemic benefit of transient diversity (and the later work qualifying it to hard, noisy landscapes); the dual-phase and organizational-ambidexterity literatures on alternating discovery and consolidation.

The boundary and the whole. Ronald Coase, “The Nature of the Firm” (1937); the principle of subsidiarity; the edge-of-chaos literature and its half-chaos qualifications; and, for the civilizational form deferred to Part 2, the coherent-pluralism tradition (Hasok Chang; Michael C. Jackson) with Isaiah Berlin on value pluralism.

Within AoM. Chapter 5 (the cost of getting bigger, and earned modularity); The Tree of Agreement (the convergence this search powers); Measuring Coherence (the Goodhart hazard at the seams); and coherent pluralism in Part 2.