Dispatch #44

Ideal organization as ideal pathology

Human organization is a set of genres: we recognize things like businesses, religions, and clubs. This continues a line of inquiry into how organizations maintain epistemic sovereignty in adversarial environments: how they know what they know, and how they survive attempts to make them know differently. The genre of an organization gives us a good first approximation on what to expect from it even if we are not familiar with the particular instance we are encountering within that genre. (We have a hard time parsing crossovers like Scientology—such suffer from bad market positioning, in a sense, much as a book that doesn′t cleanly fit into an existing genre is hard to talk about.)

Genre tells us what an organization claims to be, but our actual relationship to it reveals what it is. Our relationship to an organization ranges from purely transactional (or even exploitative) to full interdependency. The nature of this relationship is governed by belief. The sociology of belief suggests that we expect more interdependency the more belief there is. You′re going to commit more effort to your child′s sports team (even though it′s somewhat transactional) than you are to the DMV. And you will do much for your religion for free forever. (Indeed, you could probably run the equation backwards: what you do for free is in fact the index into your religion.)

But a lot of belief is in fact bonding, on a spectrum from friendly to hysterical (to traumatic). Alief, or belieflike attitudes that guide action even when they contradict explicit belief, matters here: people know an organization to be pathological but alieve it to be necessary, sustaining scripts that move the organization toward its genre’s behavioral mainline.

Organizations don’t suffer pathologies; they are intrinsically pathological constructs. Idealized organizations are not perfect. They are perfectly pathological. (Venkatesh Rao, “The Gervais Principle”)

(As something of an aside, this suggests an A. A. Milne-style typology of firm pathologies:

• Piglet organizations: Anxiety-driven, constantly seeking reassurance, risk-averse to paralysis

• Eeyore organizations: Depressive realism, ‘nothing matters anyway,’ learned helplessness

• Tigger organizations: Manic optimism, ‘reality distortion field,’ ignoring warnings

• Rabbit organizations: Obsessive process worship, form over function

• Owl organizations: Credentialism masking incompetence

Identification with particular real organizations is left as an entertaining exercise to the reader.)

For an organization, a coordination of two or more humans, the map becomes the territory. The processes of maintaining the organization immediately begin to assume prominence, distorting perceptions around the collaboration. The org chart becomes in some sense more real than the task itself. The quarterly review cycle dictates what problems get solved (those solvable in 90 days) and which get ignored (everything else). The metrics dashboard determines what’s ‘real’—if it’s not measured, it doesn’t exist, the corollary of “what gets measured gets managed”. This isn’t a corruption of an ideal form: it’s the natural state of organization.

Pournelle’s Iron Law kicks in: those devoted to the organization’s perpetuation outcompete those devoted to its mission. The process of institutional survival becomes the mission; not as sabotage but as thermodynamics. Maintaining the organization is immediate and measurable; fulfilling the mandate is distant and uncertain. (The fewer the number of people involved, the easier it is to resist this tendency.)

At the limit, the ideal organization may become a different kind of pathology: deceased. Most organizations fight death—they find new missions after completing the original one (mission creep), they perpetuate themselves through process (bureaucratization), they refuse to admit the work is done. But some should accept mortality as health.

The Benedictine Rule hasn’t been updated in 1,500 years, not from rigidity but from recognition that its work is done. The 26-letter alphabet is stable because elaboration would be corruption. Further updates would be attack vectors, not improvements—whatever change you could propose to the monastic rule, it would certainly make things worse.

This suggests a counterintuitive thesis: the greatest possible trust reposes in permanent stasis. Every upgrade is a supply chain attack vector. Every update path is a potential capture point. The only truly sovereign system is one that no longer accepts updates. But this requires accepting that the work is finished—and organizations are pathologically incapable of believing their work is ever finished, because completion means death. The rare organization that can accept completion and freeze its protocols permanently is either genuinely sovereign or genuinely dead—and from the outside these are perhaps indistinguishable.

If every update is an attack vector, then institutional sovereignty requires either: accepting permanent stasis (the Benedictine solution), or building update mechanisms with cryptographic verification that prevents capture through the supply chain (the zero-knowledge solution). The former admits the work is done. The latter admits it never will be.