The Law of Uneven Development

When I began looking for Earth’s law of uneven development, I expected to find it in failures. If political institutions require supporting conditions, then the cleanest evidence seemed likely to come from places where an imported form collapsed because those conditions were missing. Luminara had supplied me with many such memories, and the previous inquiry into intervention had made the danger of forcing development easier to recognize.

Earth offered a more inconvenient example. One of the clearest challenges to a political ladder was not a failed institution. It was an institution that many theories would have expected to arrive later, functioning because people built the conditions for it instead of waiting to inherit them.

The institution was an election.

The Election That Did Not Wait

India’s first general election, held across 1951 and 1952, asked a newly independent republic to convert universal adult suffrage into an operating fact. The Election Commission later described an electorate of roughly 173.2 million people in a country that was economically poor and only about 16.6 percent literate. The scale was unprecedented. A simple developmental story could easily have treated mass literacy, higher income, and settled administrative habits as prerequisites that should come first, with full electoral participation following later.

India did not follow that order. The Constitution had already made adult suffrage part of the political settlement, so the practical question became how to make the institution usable under the conditions that actually existed. Election machinery was adapted to low literacy rather than postponing the political standing of people who could not read candidate names. Visual symbols became central. Balloting procedures and boxes were designed so that recognition did not depend on literacy. Electoral rolls had to be created on an immense scale, and the vote was staged across varied terrain and seasons rather than assuming one administratively convenient moment.

Nothing about this episode proves that institutions can be detached from conditions. It proves almost the opposite. The election could be conducted at that scale because crucial operating conditions were deliberately created around the commitment: rules, rolls, symbols, polling arrangements, personnel, logistics, and an authority charged with making the constitutional promise operable. India was not “ready” in every dimension, and the first election did not cure poverty, exclusion, communal conflict, administrative weakness, or every defect of representation. But neither was the country simply “unready for democracy” because some variables associated with established democracies were scarce.

That distinction altered the question for me. Readiness for a particular political institution could not be inferred from a society’s overall level of development. A society could lack one capability while possessing or constructing another. It could inherit administrative practices from an earlier order, reject the political standing of that order, write a new constitutional commitment, and then build machinery around the commitment before other forms of social development had caught up. The sequence was not “development, then institution.” The institution itself became part of development.

This was already enough to weaken the ladder I had carried from Luminara. Then I encountered a sequence that ran in almost the opposite direction.

Different Sequences, Not Different Speeds

South Korea is often placed inside a familiar modernization story: rapid economic growth and expanding education eventually accompany democratization. At a distance, the sequence can look like movement along one road. Look more closely, however, and the road folds back on itself.

From the 1960s through the 1980s, South Korea industrialized at extraordinary speed under authoritarian governments. Political liberalization did not simply rise with economic capacity. Under Park Chung Hee and then Chun Doo Hwan, the political system became more authoritarian during much of the period in which the economy was becoming more industrial, urban, educated, and organizationally dense. The democratic transition associated with the mass mobilization of 1987 came later.

Even that description is too linear if it treats economic development as a clock that eventually struck democracy. Joan Cho’s study of South Korea traces a more interesting mechanism. Industrial complexes and tertiary education initially helped authoritarian governments organize growth and stabilize rule. Over time, those same structures also concentrated workers and students, created networks through which grievances could travel, and helped provide organizational bases for opposition. A condition that strengthened one political arrangement at one stage could help undermine it at another.

The implication is not that industrialization causes democracy after a sufficiently long delay. South Korea is one history, not a universal sequence, and its democratization also depended on movements, institutions, choices, repression, changing coalitions, and a particular security and geopolitical environment. The useful point is narrower: the political effect of a capability depends on the relations into which it enters. Education is not democratic by essence. Industrial concentration is not authoritarian by essence. Administrative capacity is not liberal or illiberal by essence. Each can serve one order, generate new actors inside it, and later change the conditions under which that order can survive.

Earth’s own scholarship has spent decades arguing over this problem. Adam Przeworski and Fernando Limongi found, in one influential study, that economic development did not make transitions to democracy more probable in their data even though affluent democracies were more likely to survive. Daron Acemoglu, Simon Johnson, James Robinson, and Pierre Yared later argued that once persistent country differences were taken into account, the cross-country association between income and democracy did not establish a causal effect of income on democracy. Their result was consistent, they wrote, with societies entering divergent political-economic paths at critical junctures.

Neither finding abolishes the relation between social development and politics. Wealth changes what states can fund. Education changes the distribution of knowledge and organization. Urbanization, technology, war, trade, demographic change, and administrative professionalization alter what political actors can do. The mistake is to convert these influences into one schedule. A cause can be powerful without producing the same institutional consequence everywhere, because the other conditions it meets are not the same.

South Korea was the harder correction for me. India had shown that some conditions an institution needs can be built around it rather than waited for; South Korea showed that the same capability could strengthen one political order and later help dissolve it. My ladder had assumed that each capability pushes politics one way and only one way — that education, industry, or administrative skill carries a fixed political meaning wherever it appears. Uneven development therefore does not mean that societies move at different speeds along one road. It means that political capabilities accumulate in different sequences, interact with different inherited structures, and can change one another’s effect. What looks “early” from one history may be foundational in another. What looks “advanced” as a formal institution may be functionally weak because the relations that give it force never developed. What looks like a missing institution may sometimes be a different mechanism solving part of the same problem.

What I Mean by a Law

At this point the word law may sound too strong. I do not mean a law like gravity, nor a historical formula that lets an observer enter literacy, income, urbanization, or state capacity and receive the next political stage. That would recreate the ladder under a more impressive name. I mean a constraint on political explanation: capabilities that matter to government do not mature as one synchronized package, and the effect of any one capability depends partly on the others with which it has to operate.

Consider what follows from that constraint. A state may learn to collect revenue before it learns to expose administrative error safely. It may develop competent national planning while local implementation remains uneven. It may protect electoral competition while public administration turns over too quickly to preserve long programs, or preserve administrative continuity while making correction politically dangerous. None of these combinations is permanent. Each capability changes the environment in which the others develop, and each reform can strengthen one relation while weakening another.

Unevenness also exists inside the unit called a country. Ministries accumulate different professional histories. Cities experiment under different fiscal and legal constraints. Courts, legislatures, regulators, parties, militaries, public services, and local governments can acquire capacities on different schedules. A national label compresses those differences because politics needs names for large formations. Developmental analysis has to reopen the name when the compressed unit begins to explain too much.

The practical consequence is simple but demanding. Before describing a society as advanced or delayed, name the capability. Before describing an institution as successful, name the function and the conditions that make the function possible. Before transferring the institution, ask which of those conditions are portable, which can be built, which depend on a long history, and which would be changed by the act of transfer itself. The law of uneven development is therefore less a prediction of where politics goes than a warning against explanations that erase sequence.

The Same Institution, Different Capacity

This becomes easiest to see when a political institution travels under the same name. Consider the specialized anti-corruption agency. Hong Kong’s Independent Commission Against Corruption became influential enough that versions of the model were created in many other jurisdictions. The institutional form is appealing for an obvious reason: if ordinary police, prosecutors, ministries, or political networks are themselves compromised, a specialized body with investigative expertise and a protected mandate appears able to break the circle.

But the name “anti-corruption agency” does not carry its operating environment with it. An agency can be formally independent while its budget depends on actors it may investigate. It can possess investigative authority while prosecutors or courts cannot carry cases forward. It can receive complaints while whistleblowers remain exposed. It can hire skilled staff but lose them to patronage, fear, or political turnover. It can publish findings that no institution is required to answer. The visible form survives while the function changes.

The OECD’s comparative review of specialized anti-corruption institutions makes the same point without requiring a theory of civilizational stages. The success of prominent agencies encouraged many countries to create specialized bodies, but the OECD warns that such bodies cannot fight corruption alone. Other public institutions, integrity mechanisms, enforcement channels, and administrative practices remain part of the result. The World Bank’s 2017 World Development Report states the broader principle even more directly: successful reform requires attention not only to institutional form but to function, credible commitment, coordination, cooperation, and the distribution of power in the policy arena.

Development practitioners have given one recurring failure an awkward but useful name: isomorphic mimicry. Matt Andrews, Lant Pritchett, and Michael Woolcock use it for organizations that adopt the outward appearance of capable institutions without acquiring the underlying ability to perform the function. They pair it with another problem, premature load bearing: asking a weak organization to perform at a level associated with mature capability before the learning, authority, resources, routines, and feedback required for that performance have accumulated.

The concepts are useful because they explain why a copied institution can fail without proving that the receiving society is incapable of the function. The failure may lie in the transfer itself. A reform can demand a level of performance that presupposes records that do not yet exist, professional roles that are not protected, coordination that no unit can compel, or a public expectation that has never had time to become institutionally consequential. When the reform then performs badly, the weak performance is attributed to the society rather than to the mismatch between form, burden, and enabling conditions.

I recognized Sareth in that structure, but the recognition now had an Earthly basis. The value of the comparison was not that Earth had reenacted our fictional history. It was that Earth research supplied a mechanism I could test without asking anyone to believe Luminara. A transplanted form can appear complete while the function it is supposed to perform remains distributed across conditions that were not transplanted with it.

Readiness Is Not a Country

Once that mechanism is visible, the word readiness becomes difficult to use carelessly. A country is not ready or unready in the abstract. Ready for what? To hold a secret ballot across a vast electorate? To protect an audit office from retaliation? To decentralize a service without losing fiscal control? To let a court review an emergency decision quickly enough to matter? To preserve a long infrastructure program through a change of government? Each function has a different set of enabling conditions, and the same society can be strong in one while weak in another.

This is why national averages are so seductive. Income, literacy, state capacity, trust, corruption, rule-of-law indicators, and regime classifications can reveal important patterns. They can also invite a false compression in which a country becomes one developmental score. The score then begins answering questions its components never measured: whether an institution can work, whether a population can govern itself, whether a reform should be attempted, or whether another actor has standing to intervene.

A better diagnosis is relational. The question is not whether the soil is generally good. It is whether this mechanism, in this place, depends on conditions that are present, absent, substitutable, or buildable. Some conditions are material: money, transport, communications, records, trained personnel. Some are institutional: protected tenure, review authority, reliable succession, budget control, enforceable jurisdiction. Some are social: organized constituencies, professional norms, expectations of compliance, trust or distrust earned by prior conduct. Some are historical: memories of repression, bargains that ended conflict, inherited administrative routines, old exclusions that make a formally neutral rule politically explosive.

The categories interact. A strong statute can create demand for professional capability. A new institution can teach people to use a complaint route that initially appears weak. A public service can build trust through repeated competent performance. An election can create incentives for parties to organize populations that had previously been administratively invisible. Conditions are not merely prerequisites waiting outside an institution. Institutions can help produce their own conditions, just as badly designed reforms can destroy them.

This is the correction India forced upon my reasoning. Low literacy was a real constraint. It did not determine the political standing of the voter. The institution was redesigned around the constraint, and the administrative work of conducting the election helped make universal suffrage an experienced political fact. To call the society unready for democracy on the basis of low literacy alone would have treated one missing capability as decisive for the whole institutional possibility.

Unevenness Is Not Permission

This boundary matters because the language of development can easily turn descriptive difference into political hierarchy. If one society is “ahead” and another “behind,” the metaphor quietly supplies a destination and a timetable. The society called advanced becomes evidence of the future; the society called backward becomes evidence of delay. Once that picture is accepted, external direction can begin to look less like a political choice and more like assistance to history itself.

I have already described what happened on Luminara when developmental difference hardened into permission. For a long time I treated that history as a warning about our excesses. Earth has made the warning more exact. Different institutional timing can explain why a transplant fails; it cannot establish that a population has lesser standing, that coercion is justified, or that the actor with greater capability knows the destination toward which the other must be moved. Equal standing and institutional similarity are different propositions.

The ladder had been politically convenient on Luminara precisely because it converted a difficult judgment into a chronological one. If another society was merely earlier, intervention could be narrated as shortening a path rather than choosing among contested futures. The vocabulary of development did not cause our interventions, but it made some of them feel less political than they were. Earth has forced me to separate two questions our history repeatedly fused: what is failing here, and who has standing to decide what should replace it? Evidence about a defect can answer the first without granting anyone the second.

Nor can I use that separation to excuse moral indifference. Luminara’s history gives me no right to say that because political development is plural, every locally produced order deserves protection from judgment. Some political arrangements commit atrocities, destroy the possibility of lawful opposition, enslave or exterminate populations, or direct violence across borders. A theory of many developmental paths cannot become a sanctuary for domination merely because domination arose locally. Coexistence does not prohibit every external action either. Defense, treaty obligations, and protection against extreme harm can create legitimate reasons to act. But action taken for those reasons still requires its own basis of authority and a separate judgment about necessity, proportionality, available alternatives, and likely consequences. “We are more developed” is not one of them.

For the same reason, I now understand coexistence as more demanding than passivity. On Luminara I had thought of our later coexistence rules chiefly as restraints on power: agreements about what stronger societies would no longer do to one another. Earth has made another function visible. Those restraints protected developmental time. If political forms develop unevenly, societies need enough security from one another to let internal correction, experimentation, failure, and institutional memory accumulate. Constant threat can make centralization rational where decentralization might otherwise grow. External sponsorship can change which domestic actors survive. Sanctions, guarantees, aid, covert support, war, and trade dependence can all alter the conditions later cited as evidence of what the society “naturally” became. No political system develops in isolation, but neither can its development be understood if the surrounding pressures are treated as scenery.

The conclusion is uncomfortable for me because it removes two shortcuts at once. I cannot say that every society should be left exactly as it is; political arrangements remain open to criticism, resistance, reform, and, in extreme cases, legitimate external constraint. I also cannot say that every society should move toward the same institutional bundle as quickly as possible. The first shortcut makes external restraint absolute even when extreme harm may justify constraint. The second pretends judgment has already been completed by history.

Comparison Without a Ladder

Once I give up the ladder, I do not want to give up comparison with it. That would turn a correction in method into an excuse for intellectual isolation. Political systems already learn from one another. Constitutions borrow clauses. Administrations copy procedures. Courts cite foreign reasoning. Cities imitate transport rules. Regulators compare standards. Protest movements learn tactics across borders, and governments learn from one another how to contain them. Political development has always been comparative, even when the comparison was denied.

The harder question is what comparison must preserve if learning is not to become transplantation. It must preserve sequence. It must distinguish the problem an institution was built to solve from the visible form it eventually took. It must identify which conditions made the mechanism usable, what responsibilities were distributed around it, what failed, and what later changed. Most of all, it must preserve the possibility that the same mechanism will behave differently when those relations differ. Comparison becomes useful only when resemblance is treated as the beginning of inquiry rather than its conclusion.

At small scale, a scholar can reconstruct such histories. At Earth scale, the task expands beyond any one discipline or archive. The relevant evidence is distributed through constitutional histories, administrative records, budgets, court decisions, ethnographies, surveys, implementation reports, local languages, failed reforms, and institutional memories that were never written for comparison. Before deciding that one practice should travel, a serious analyst may need to know not only where it appeared to succeed but where similar forms failed, which conditions diverged, and whether later mutations changed the original conclusion.

The burden is not only volume; it is recognition. Institutions that perform similar work may carry different names, sit at different levels of government, or divide one function among several bodies. The reverse also occurs: two institutions can share a familiar label while carrying different authority, incentives, or routes of correction. I had made this mistake myself when I first studied Earth. I looked for arrangements that resembled Luminaran forms before I learned to ask which relation was actually doing the work. A comparative instrument would be valuable first because it could widen that search, not because it could settle the comparison.

This is where the machine earns a place in my argument. It can search across more of that record than a person can hold in working memory, align related institutions whose names differ, compare versions and translations, and surface candidate cases in which the same visible mechanism met different conditions. That does not make the comparison true. A candidate analogy still has to be opened back into sources, scope, sequence, and institutional context. The machine changes the cost of assembling comparisons that humans must still judge.

I would therefore ask such a machine a different question from the one my old Luminaran model invited. Not: Which society is more advanced? Not: Which regime should the others become? Ask instead: What problem was this institution solving? What had to exist for it to function? Which condition was created by the institution itself? What failed when the form moved? Which later mutation changed the result? The answer should be a set of inspectable relations, not a maturity ranking.

That is a modest use of a powerful technology, and for now it is enough. My case does not depend on your machines. Earth’s own history already defeats the simple ladder. The machine matters because, if these developmental differences are real, the amount of comparison required for responsible learning is vastly larger than a politics of slogans can sustain. It may help societies ask better comparative questions sooner; it cannot decide which borrowed answer they should live under.

Before Learning, Coexistence

I arrived on Earth with a habit I did not know I possessed. I saw unevenness as temporal distance. Some political forms appeared to me as earlier versions of problems Luminara had already encountered; others looked closer to arrangements we had reached later. I was careful not to call your societies primitive, yet part of my reasoning still arranged them along an invisible chronology.

India and South Korea disrupted that chronology in opposite ways. One placed mass political participation before several social conditions a ladder would have expected to precede it, then built administrative machinery around the commitment. The other accumulated extraordinary economic and organizational capacity while political authority became more concentrated, only for parts of that developmental architecture later to help support democratizing pressure. The anti-corruption comparison added a third disruption: even when the visible institution is nearly the same, function depends on relations that do not travel in the name.

I no longer think “ahead” and “behind” are useful descriptions unless the object is named. Ahead in what? Administrative continuity, peaceful succession, fiscal capacity, correction, local participation, public order, protection from arbitrary power, long-horizon planning, information flow? A society can be highly developed in one relation and dangerously weak in another. The combinations change over time. No single ranking preserves that structure.

This changes the meaning of coexistence. If political difference were merely delay, then coexistence could be treated as temporary patience while slower societies caught up. But if development is genuinely plural—if capabilities arrive in different sequences, mutate under different pressures, and sometimes reveal possibilities other histories did not preserve—then coexistence becomes epistemic as well as moral. It protects the space in which political variation can become intelligible before power destroys or standardizes it.

Sustained learning from living political variation has a prior condition: societies must remain sufficiently free to develop on their own trajectories for those trajectories to remain available for comparison. That does not mean accepting cruelty, abandoning universal human standing, or refusing every common rule. It means giving up the presumption that difference is evidence of incompletion and that superior capacity creates a license to finish another society’s development for it.

I began by telling you what my civilization learned: Luminara spent centuries trying to make political difference safer by reducing it. Earth’s evidence has made the warning more exact. The danger I am tracing does not arise from difference alone; it begins when difference is arranged into a ladder and the ladder is mistaken for history.

Once the ladder disappears, however, another problem appears. Democracy, dictatorship, monarchy, federation, one-party rule, technocracy—your labels still describe something real, but they begin to look less like destinations and more like snapshots. If political capabilities can emerge in different sequences and migrate without carrying the whole regime with them, then a political system cannot be understood only by what category it occupies today.

A ladder asks how far a political system has climbed. The next question is what the political system is becoming.