ARGUMENT FLOW
Dynasties Don't Escape the Cycle Because of Bad Men
A widely upvoted Zhihu answer attributes dynastic collapse to institutional sclerosis rather than corruption. This piece maps its argument chain, its historical anchors, and the three gaps it leaves open.
Source: Zhihu · Teresiya · Text marked with Source is excerpted from the original answer; everything else is this site's independent analysis
I · Chapter One
Four Stages of a Dynasty
Temporal evolutionStage One
Founding: eliminate uncertainty
Founders assume people are unreliable, so they write laws, build institutions and fix conventions — patches applied to the machinery of rule.
Patching
Stage Two
Early period: the patches work
Power is pulled out of the instability of personal rule and turned into a predictable production line. This is the early golden age of every dynasty.
Running smoothly
Stage Three
Middle period: patches grow hosts
Each institution breeds a group that lives off it — canal administration and the boatmen's guild, the examination system and the gentry, compliance review and the consulting industry.
Distributional coalitions
Stage Four
Late period: the patches bite back
Two centuries of accretion form an airtight net. When reformers try to turn the ship, every layer of protection has become a constraint.
Cocoon of one's own making
What the four stages share: Not one of them requires a villain. Nearly all of the answer's effort goes into establishing exactly this — collapse is what rational people produce together by following rules that were once correct.
* * *
From stages to deadlocks
II · Chapter Two
Four Structural Deadlocks
Mechanisms of resistance2.1
The legitimacy paradox: reform is not defeated, it is ruled illegitimate
Trigger
Fiscal exhaustion or invasion convinces the ruler that the ship must turn.
Fiscal exhaustion or invasion convinces the ruler that the ship must turn.
↓
Form of resistance
Vested interests never say «no». They invoke a higher-ranking source of legitimacy instead.
Vested interests never say «no». They invoke a higher-ranking source of legitimacy instead.
↓
Deadlock
As a matter of law they are correct, and the reformer has no ground to answer from. Chongzhen and Louis XVI met the same structure; only the texts cited differed.
As a matter of law they are correct, and the reformer has no ground to answer from. Chongzhen and Louis XVI met the same structure; only the texts cited differed.
Result → resistance wears legal dress, and reform dissolves inside the procedure
2.2
Anti-corruption fails: it cannot touch lawful rigidity
Intuitive fix
Purge corrupt officials in the late period and the state can be saved.
Purge corrupt officials in the late period and the state can be saved.
↓
Distinction
Corruption is unlawful greed — cheap to identify, with well-defined remedies.
Corruption is unlawful greed — cheap to identify, with well-defined remedies.
↓
Deadlock
What actually drains the state are bureaucrats who stall projects by following procedure, experts paid to demonstrate that reform is unworkable, and elites who harden family interest into tradition — all of it entirely lawful.
What actually drains the state are bureaucrats who stall projects by following procedure, experts paid to demonstrate that reform is unworkable, and elites who harden family interest into tradition — all of it entirely lawful.
Result → corruption metrics improve while governing capacity does not recover
2.3
Transaction costs consume the surplus: the same task keeps getting more expensive
Symptom
The later the dynasty, the more gates any action must pass through.
The later the dynasty, the more gates any action must pass through.
↓
Contrast
In the early Tang a single edict moved armies; in the late Ming, moving one unit required wrangling through the ministries of war and revenue, the Grand Secretariat and the palace eunuch bureau.
In the early Tang a single edict moved armies; in the late Ming, moving one unit required wrangling through the ministries of war and revenue, the Grand Secretariat and the palace eunuch bureau.
↓
Deadlock
Social resources stop flowing toward production and innovation, and are consumed by internal transaction costs instead.Source
Social resources stop flowing toward production and innovation, and are consumed by internal transaction costs instead.Source
Result → extreme refinement and extreme incapacity appear together
2.4
The cure mutates: the department for reducing entropy generates its own
Remedy one
If the disease is entropy rather than corruption, create an agency whose sole job is repealing law.
If the disease is entropy rather than corruption, create an agency whose sole job is repealing law.
↓
Reality
That agency grows its own headcount, procedures and interests, and ends up part of the thing it was meant to cut.
That agency grows its own headcount, procedures and interests, and ends up part of the thing it was meant to cut.
↓
Remedy two, same fate
Sunset clauses: Jefferson argued constitutions, laws and public debt should all carry a mandatory expiry of at most nineteen years. Madison's rejection was that coalitions attached to the clause would turn it into indefinite renewal.Source
Sunset clauses: Jefferson argued constitutions, laws and public debt should all carry a mandatory expiry of at most nineteen years. Madison's rejection was that coalitions attached to the clause would turn it into indefinite renewal.Source
Result → both cures are eaten by the very mechanism they were meant to treat
Where the four deadlocks meet
The property that makes an institution effective and the property that makes it unrepealable are the same property.
Purging corrupt officials solves unlawful greed. It does nothing about lawful rigidity.
Teresiya · original answer
* * *
From deadlocks to the loop
III · Chapter Three
A Self-Reinforcing Loop
Feedback cycleThe loop needs no external enemy — every step is executed by rational actors inside the system.
* * *
From the loop to the numbers
IV · Chapter Four
Nineteen Years Against Two Hundred
The only numeric contrast in the pieceAn institution's actual lifespan versus its designed expiry
The answer barely uses numbers. Its one comparable pair falls in the section on remedies: how long institutions actually accrete, set against the mandatory expiry Jefferson proposed.
Actual accretionSpan described in the answer
Jefferson's proposed capExpiry for constitutions, laws and debt
04080120160200
!
More than a tenfold gap.
The gap is not a technical problem. Madison's rejection names the difficulty precisely: even with an expiry in place, the coalition attached to the clause will push for indefinite renewal — so the expiry itself needs an expiry. This is the fourth deadlock reappearing at the level of institutional design. This site has not independently verified the primary Jefferson–Madison correspondence; it is presented as relayed by the original answer.
* * *
From the numbers to the verdict
V · Chapter Five
Four Transferable Rules
Verdict and transferThe value of this argument chain is not its conclusion. That institutional sclerosis drags down regimes is not a new claim; social science has mature formulations of it. What is worth having is the decomposition — it breaks «why do intelligent people collectively fail to decide correctly» into four separately testable links, and the decomposition transfers to other domains far more readily than the conclusion does.
1
Measure the two decays separately
Unlawful greed and lawful rigidity run on different mechanisms. Falling corruption metrics do not imply rising governing capacity, and using the former to vouch for the latter is a common misreading.
2
Count veto points, not rules
The more reliable proxy for institutional burden is not the total volume of law but the number of independent veto points one task must clear. The first is easy to dress up; the second is hard to hide.
3
Ask who the host will be
Any rule that lands breeds a group that lives on it. If repeal conditions are not written at the moment of creation, no one will later have an incentive to execute repeal.
4
A cure must say who ends the cure
When creating a streamlining agency or a sunset clause, the first question is who terminates the mechanism itself. The answer does not address it, and neither did Jefferson's proposal.
* * *
This site's verdict
A complete argument chain on solid historical anchors, missing one falsification test and one attribution.
Strength
The Chongzhen–Louis XVI pairing turns «institutional resistance» from metaphor into testable structure: in both cases opponents concede sovereign authority and invoke a higher-ranking legitimacy instead.
Gap
No counterexample is addressed. Polities with equally long institutional accretion have preserved mobility through incremental constitutional amendment or federal division of power.
Attribution
The core model overlaps heavily with Olson's distributional-coalition theory, uncited here; the originality lies in the historical instantiation and the prose rather than the theory.
* * *
Editor's Note
How to Read This Article
Why it matters
It moves dynastic collapse out of moral narrative and into mechanism, and supplies one operational distinction: unlawful greed and lawful rigidity run on different mechanisms and therefore cannot be treated with the same instrument. The Chongzhen–Louis XVI pairing is not decorative — the form of resistance is identical in both, with vested interests conceding sovereign authority while invoking a higher-ranking source of legitimacy. That is a level above the «the last emperor was incompetent» account this topic usually receives.
Caveats
The argument runs in one direction only. No counterexample is engaged: polities with comparable two-century accretion have preserved mobility through incremental amendment or federal division of power. Without that test, «no dynasty escapes» slides from conclusion back to selective illustration, which is the main reason the rating stops at 3 rather than 4. The modern parallels — Western infrastructure permitting, the government efficiency department — are asserted without data or case evidence.
Transferable insight
The operative mechanism is that a rule's beneficiaries and its repeal authority coincide — whoever eats from the rule holds the veto over removing it. The same structure appears in software engineering: every compatibility layer grows callers that depend on it, the cost of deleting an old interface rises monotonically, and whether it may be deleted is usually decided by the team that maintains that layer. Same conclusion: if repeal conditions are not written at creation, no one with an incentive to repeal will exist later.
Source excerpt