Chapter 03 — The Internal Accumulation of Executive Power

The Internal Accumulation of Executive Power

Expansion in scale brings division of labor; division of labor creates scattered execution nodes. At first, these nodes are nothing more than passive receivers, waiting for the sovereign’s instructions, returning to stillness once execution is complete. They are independent screws in a toolbox, with no lateral connection to one another.

But as the business an organization handles grows more complex, the simple model of ‘sovereign issues instruction, node executes’ begins to break down. The sovereign’s bandwidth is finite; he cannot coordinate the interface between every single node. To keep the organization from collapsing into chaos, nodes must build lateral coordination mechanisms and vertical reporting layers among themselves. Bureaucracy is born from this necessity.

The restaurant chain from the previous chapter offers the most concrete sequel to this birth. When there were only two or three locations, the chef could still coordinate between them personally — whichever location ran short, he decided by a single word which location would resupply it. But once locations multiplied into the dozens, this kind of personal, chef-mediated coordination could no longer keep pace with the frequency of daily operations. So the location managers began building their own conventions for resupplying one another: fall below a certain stock level, and a manager would notify a neighboring location directly; neighboring locations would decide, by an unspoken first-come-first-served understanding, who would provide the support. This convention began purely as a stopgap the managers developed so they wouldn’t have to trouble the chef with every single matter — the chef himself may not even know the specific content of this convention.

The essence of bureaucracy is weaving scattered points of execution into an organic network. When Node A needs Node B’s data to complete a task, and Node B needs Node C’s approval before it can supply that data, an internal logic independent of the sovereign’s will begins to grow. This logic’s original purpose is to raise efficiency, but it soon evolves into a mechanism of self-protection.

This accumulation is completed through ‘the internalization of procedure.’ To lower the transaction cost of every act of coordination, nodes develop standard operating procedures, internal norms, and conventions among themselves. Once established, these procedures no longer require the sovereign’s approval each time in order to run. Nodes discover that following procedure is faster than petitioning the sovereign; the sovereign discovers that relying on procedure is less effort than personally intervening.

A company’s information technology department demonstrates the internalization of procedure in its most typical contemporary form. Any request to modify a system could, in theory, be reported directly to senior leadership for a ruling. In practice, such requests follow an established path: first a technical review meeting, then approval from a change-management committee, and finally a slot in the deployment schedule — this entire gauntlet was originally designed, purely out of prudence, to prevent any change from going live too hastily and crashing the system. But once this process begins running, it develops its own rhythm and inertia: committee membership, meeting frequency, the specific standards for approval, all gradually evolve into a body of internal rules that only those inside the process can fully grasp. A newly appointed senior executive, even one who formally holds the power to veto any change, will find it very difficult in practice to bypass this entire process and push through an emergency change he considers important — not because any rule forbids him from doing so, but because the coordination cost of bypassing the process far exceeds whatever benefit his direct intervention might bring.

And this process, in running, does more than merely allocate tasks — it also allocates what is worth being seen. When a change request is discussed at a technical review meeting, the meeting’s agenda, the length of time devoted to discussion, the professional backgrounds of those present, all invisibly determine which risks get magnified for scrutiny and which get waved through with a shrug. An engineer who discovers a potential systemic flaw, but finds that explaining it would require crossing several technical domains and additional meetings, is quite likely to file the discovery under ‘known risk’ rather than launch an investigation that might delay the entire project timeline. This is not deception. It is a rational response to the cost of process — within an already-internalized procedure, the pressure to keep the process running smoothly becomes, invisibly, a filter on what information gets surfaced at all.

The formation of medieval England’s Exchequer and Chancery offers a much older specimen of this same process. A newly formed government might initially depend on a handful of trusted confidants to handle its affairs; as the scope of governance expanded, functions split into specialized offices — the Exchequer developed its own methods for auditing royal accounts and calculating what was owed to the crown, the Chancery developed its own conventions for drafting and sealing writs, each accumulating its own fixed documentary formats, review checkpoints, and reporting timelines for the business proper to it. The rules coordinating these offices with one another — which office had to review a matter first, which office then executed it — were never dictated clause by clause by any king. They were conventions worked out, department by department, over decades and centuries of actual operation. At some point the sheer complexity of these conventions came to exceed what any single monarch could personally master in detail. What a king could still do was affix his final approval to whatever this existing network had already produced.

Once every executing node has grown accustomed to operating by internal procedure, the internal accumulation of executive power is complete. Executive power is no longer a series of discrete, temporary grants issued by the sovereign; it has become a network with its own structural tension — a network that now governs the flow of information, the allocation of resources, and the distribution of friction.

At this point, executive power is no longer merely a derivative of sovereignty. It has acquired a structural weight of its own. Even if the sovereign were briefly absent, this network could keep the organization running at a baseline level on inertia alone. Executive authority has accumulated from a state of ‘obeying commands’ into a state of ‘running on its own.’ This is the rudimentary form in which the apparatus first acquires a body — the moment a tool learns to weave itself into a network and operate by its own logic, without waiting on any instruction from its master, it has stopped being a tool.

The moment this accumulation completes is worth placing beside the observations of the previous two chapters. The first chapter described how a single tool, in a single user’s hands, quietly draws the boundary of judgment. The second described how, once scale crosses its threshold, the channel between sovereign and reality closes behind a filtering chain. Here, the third chapter completes the final piece of the picture: every node of the filtering chain is no longer an isolated individual — it has been woven into a network of mutual dependence, one that coordinates itself. An isolated node, even one that begins filtering information, has a limited sphere of influence in the end. But once these nodes lock into the same network through the internalization of procedure, what they share is no longer merely their own local survival pressures — it is an entire collective logic devoted to keeping the network itself running smoothly and stably.

This network has not yet begun devouring its master — it has only completed the full anatomical structure devouring requires. But once this structure is in place, devouring is no longer a matter of chance. It becomes a mathematical necessity under ecological pressure. A node will discover that the simplest way to keep the network running smoothly is rarely to transmit reality in full — it is to transmit a trimmed summary, one that triggers no additional coordination cost.