The Material Basis of Political Agency

Military revolutions, geography, surplus, legitimacy, and the AI era

Exploration v0.1 · 2026-08-15 · working hypothesis · English working translation 2026-08-17

1. Central hypothesis: from material necessity to bargaining power, distribution, and legitimacy

This page does not argue for simple technological determinism such as “infantry creates democracy” or “cavalry creates feudalism.” The crucial variable is which actors can sustainably reproduce effective military capability. Who can afford weapons, training, horses, elephants, ships, artillery, aircraft, compute, soldiers, logistics, and payroll? Who controls the surplus required to keep those systems operating?

Compressed into a feedback loop:

The causality is not one-way. Surplus can make armament possible, but actors who cannot be ignored militarily may also be harder to strip of their surplus. Political orders may have a range of distributive equilibria permitted by the military balance.

It is therefore important to distinguish fighting in an army from being an independent military actor. Large numbers of peasants may serve as infantry while equipment, pay, logistics, and command remain monopolized by the state. Such soldiers are inputs into state capacity, not necessarily independent actors able to bargain with the state.

2. Participation thresholds and surplus pooling

Military participation has a total cost: weapons, training, food, foregone labor, horses, fodder, armor, retainers, and time. If the disposable surplus of an ordinary household falls below that threshold, the household cannot reproduce itself as an independent military actor.

Concentration of surplus can therefore have nonlinear effects. Small surpluses spread across many households may support no warhorse at all, while one lord who pools them can maintain horses, armor, retainers, and long-term specialist training.

But a tax state can aggregate many small surpluses into another equilibrium. Roman or Qin-Han style fiscal states could collect small household surpluses into a treasury and transform them into equipment, pay, logistics, and mass armies. When central fiscal capacity weakens, lords, local magnates, cities, or religious communities may substitute as pooling institutions.

Relation to prior literature: what is inherited and what is recombined

Many elements of this argument have clear precedents. The claim is not that the page discovers an entirely new general theory, but that several research traditions can be connected by the common variables who can convert surplus into military capability and whose cooperation rulers need, then compared from antiquity through the AI era.

• War and state formation: bellicist state-formation arguments associated with Charles Tilly emphasize cycles in which war expands taxation, conscription, administration, and the concentration of capital and coercion. The “surplus pooling” framework is close to this lineage but treats the breakdown of public pooling and substitution by lords, magnates, cities, and religious corporations within the same frame. See a recent reassessment of the bellicist paradigm.
• Military Revolution and fiscal-military states: large literatures connect military change to state finance, standing armies, bureaucracy, absolutism, and constitutional divergence. This page abstracts one step further: not simply which technology appears, but which actor can financially reproduce the new military capability. See overviews discussing Downing and related work and the fiscal/military state literature.
• Mass warfare, citizenship, and redistribution: research linking mass war and conscription to citizenship, welfare states, progressive taxation, and mass politics directly informs the modern section. This page generalizes from democratization to mass legitimation, which can also take nationalist or party-state forms. See reviews of mass warfare and welfare states and Scheve and Stasavage.
• Geography and political fragmentation: the Fractured-Land Hypothesis is a close quantitative predecessor for explaining Chinese reunification and European polycentrism through terrain and the distribution of high-productivity land. Here geography is interpreted as interacting with military technology and acting through local outside options and central surplus capture. See Fernández-Villaverde et al.
• Asiatic mode and hydraulic despotism: Marx’s Asiatic mode of production and Wittfogel’s hydraulic-despotism thesis are older attempts to connect geography, surplus, and centralized state power. Strong claims that irrigation itself generated despotism have been heavily qualified. This page replaces water control with geography × military scale economies × surplus capturability. See work challenging hydraulic-despotism accounts in Chinese history and examining ancient Egypt.

The comparatively distinctive move, if any, lies in the integration: participation thresholds, pooling institutions, geography, material dependence, distribution, and legitimation are treated as one feedback system, and then extended to possible AI-driven redundancy of human military, productive, administrative, and cognitive roles.

3. Eleven stages for comparing military revolutions

1. Chariot revolution — concentrates expensive mobile military capital in palaces and warrior elites.
2. Iron / first infantry revolution — mass infantry partially displaces chariot aristocracies.
3. Feudal revolution — weakening public fiscal pooling links land, dependent populations, and military specialists.
4. Cavalry revolution — raises the speed of battle, reconnaissance, pursuit, tax extraction, suppression, and long-distance movement.
5. Second infantry revolution — cities, rural communities, and mercenary corporations organize crossbows, pikes, spears, and early firearms collectively.
6. Firearms revolution — lowers some individual skill and capital thresholds while increasing the importance of mass procurement and logistics.
7. Standardization revolution — places people, equipment, information, time, and space onto common standards so population itself becomes scalable state capacity.
8. Total-war revolution — connects society as a whole to the military system.
9. Professionalized RMA — reconcentrates capability in advanced skills, capital, and technical specialists.
10. Twenty-first-century democratization of military capability — drones, commercial satellites, GPS, open-source software, and cheap sensors reduce thresholds for small actors.
11. Military AI revolution — may strengthen either decentralization or a new phase of central standardization.

No region must pass through all eleven in the same order or with the same intensity. Absence, delay, or early absorption by the state are themselves comparative data.

4. Two infantry waves and “democratization windows”

At least twice, new military forms appear to have reduced the relative advantage of expensive military elites and widened the set of middle strata or corporate communities able to become militarily consequential.

First wave: Iron Age mass infantry

Where small political units could survive geographically and middling landholders could equip themselves, a pathway like household surplus → armament → military importance → political participation could emerge. The Greek polis is a candidate example.

Where a large territorial state could register, equip, tax, and conscript rural populations directly, mass infantry could instead strengthen centralization. Warring States China and Mauryan India sit closer to this opposite branch.

Second wave: late-medieval corporate infantry

Urban and rural communities and mercenary organizations increasingly pooled military capital in crossbows, pikes, long spears, early firearms, wages, and collective training. The first wave often centered on households able to arm themselves; the second centered more on corporate ownership and organization of military capital.

Flanders, the Swiss Confederacy, Frisia, Dithmarschen, Hussite forces, and Japanese leagues, villages, and autonomous cities provide cases worth comparing.

Such windows need not last. In both waves, scale advantages can return to the center once large states learn to procure, supply, and standardize the new form at massive scale.

5. Chariots, iron, and cavalry: ancient divergences

Bronze Age chariots required concentrated military capital: vehicles, multiple horses, drivers, warriors, fodder, maintenance, and long training. This often gave military logic to concentrating surplus among palace and warrior elites.

Egypt had a unified state before chariots, so chariots cannot explain the origin of kingship there. A narrow fertile corridor, the Nile as cheap transport, and surrounding desert may instead have created an asymmetry in which local exit was costly while central movement was cheap; chariots later amplified imperial projection.

Mesopotamia sustained city-state competition for long periods, but weak natural barriers also made conquest and snowballing surplus capture possible when one city gained an advantage. Later cavalry increased the radius within which reconnaissance, communication, pursuit, suppression, and taxation could be coordinated.

Ancient India shows another branch. Even after chariots declined, elephants and cavalry remained expensive military capital. Mass infantry did not imply comprehensive democratization of military assets; armies combined mass infantry + expensive elephants + expensive cavalry. Large states could gain scale economies by connecting Gangetic agrarian surplus, eastern elephant resources, and northwestern horse supplies.

6. China: territorial states absorbing military democratization

The Spring and Autumn and Warring States transformation is better understood not as “iron defeats chariots” but as a simultaneous shift toward mass infantry, crossbows, cavalry, and territorial states reaching directly into rural society through registration, taxation, and conscription.

In Greek poleis, some independent farmers could arm themselves and constitute the political community. In China, state institutions could register rural populations and convert them directly into huge armies. Continuous northern plains and large populations increased returns to territorial scale. Here mass army ≠ democratized military capital.

The Sixteen Kingdoms period is a useful counterexample and refinement. Cavalry coalitions could conquer broad territories while magnates, fortified communities, and displaced populations privately controlled people and surplus that the state could not readily convert into taxes or troops. At times conquest radius > governance radius.

The rapid fragmentation after Former Qin’s near-unification illustrates that military coalition control and fiscal integration are separate problems. Northern Wei stabilization required reconnecting land, population, magnates, and taxation to state institutions.

After the Tang collapse, by contrast, counties, officials, cities, taxation, and documentary administration persisted in localized form. The Song could reunify much of the former core without rebuilding state society from scratch.

Late-Qing warlordism similarly points to who reproduced military capacity financially. From the Taiping era onward, provincial officials and local elites organized armies such as the Xiang and Huai forces, shifting military-financial circuits from center to locality. After imperial collapse, those circuits could become political actors in their own right.

7. Reframing the “Asiatic mode” through military geography and distribution

A simple hydraulic theory in which irrigation itself creates despotism is too strong. But some features older “Asiatic mode” arguments sought to explain—centralized extraction and weak autonomous local actors—may be reformulated through another mechanism.

If broad continuous agricultural zones + easy movement of large armies + central knowledge of population and surplus + military and logistical scale economies combine, local outside options weaken. Captured surplus can be converted into further military power, creating a positive feedback loop of centralization → army → surplus capture → more centralization.

For comparison, a term such as continental extractive equilibrium may be more useful than a cultural category like “Asian”: it can appear strongly in China, partially in northern India, weakly in Japan, and in some European cases such as Russia.

8. Medieval Europe and Japan: from weakened public pooling to a second infantry revolution

After the western Roman fiscal system contracted, public pooling for salaried armies weakened and local military specialists tied to land and dependent populations became more important. The cavalry shift is better viewed as a long interaction among fiscal contraction, private surplus pooling, mobility, and the high cost of mounted specialists than as the effect of a single invention.

As population, agricultural output, commerce, and monetization recovered, cities, guilds, and rural communities could jointly finance arsenals, walls, crossbows, pikes, wages, and training. The second infantry revolution was therefore also an organizational revolution: turning thousands of individuals into a corporate military actor.

Japan displays a compressed analogue: retreat of large ritsuryō mobilization, growth of mounted warriors rooted in local land and networks, village leagues and autonomous towns, ashigaru, long spears and firearms, followed by reabsorption into large daimyō organizations. Diffusion of capability did not guarantee permanent decentralization.

9. From firearms to the standardization revolution

Firearms lowered some individual skill and capital thresholds, making it easier to convert ordinary people into effective fighters. But firearms alone cannot operate armies of hundreds of thousands or millions.

Its components include universal conscription, basic education, common languages, standardized drills, ranks and unit structures, interchangeable weapons and ammunition, maps and surveying, railways, telegraphy, staff systems, and mobilization plans.

Standardization could make individuals more replaceable while increasing the bargaining relevance of the population category as a whole. A knight was military capital whose identity mattered. A modern soldier could be comparatively interchangeable, yet the state might need millions of such ordinary citizens.

Thus individual bargaining power can fall while the bargaining power of “the population,” “workers,” or “soldiers” as categories rises, shifting legitimation from personal bargains with nobles toward mass categories.

10. Total war: mass incorporation rather than democracy

Total war connects soldiers, factory workers, farmers, railway workers, engineers, women workers, taxpayers, teachers, medics, and administrators to state capacity. The latent military value of the general population becomes extremely high.

But total war does not map directly onto democracy. The same dependence on the masses supported liberal democracy, communism, and fascism.

Legitimation could take several forms: political participation, social equality and redistribution, nationhood, religion or tradition, or performance and prosperity. Communist states were not indifferent to mass legitimation; they were intensely mass-political while vertically integrating independent organization into party-state structures.

This separates the masses being materially necessary from the masses existing as independent political actors. The path from dependence to democracy also requires mediating institutions such as independent parties, unions, municipalities, religious organizations, firms, or communities.

11. Professionalized RMA: why reduced direct military dependence did not immediately destroy democracy

After World War II, jets, submarines, nuclear weapons, missiles, electronic warfare, and C4ISR reconcentrated capability in specialized skills, capital, and technical organizations. A simple material-dependence model might predict a rapid collapse of mass bargaining power, but established Western democracies did not immediately disappear.

This is an important counterexample. Path dependence in elections, courts, parties, unions, and welfare states; norms of civilian control; continued dependence on civilian industry, taxation, and infrastructure; and the latent possibility of large-scale land war and reserves all preserved forms of mass relevance.

Professionalization itself is politically ambiguous. A military that defines itself narrowly as expert in war can coexist with civilian government; one that defines itself as expert in national order, modernization, or internal security may intervene domestically. Political effects therefore depend on professionalization × threat model × mobilization model × inherited institutions.

12. Twenty-first-century capability diffusion: a third window?

Cheap drones, commercial satellite imagery, GPS, smartphones, open-source software, communications gear, and sensors diffuse capabilities once limited to states or large armies toward small states, small units, and non-state actors.

This may be a third democratization window analogous to earlier infantry waves. Small actors gain reconnaissance, denial, disruption, and precision-strike capacity at lower thresholds.

If history rhymes, the next step may be reabsorption: mass procurement, electronic warfare, satellites, industrial production, common communications standards, and AI battle management could let states integrate millions of sensors and unmanned systems into one military architecture, restoring large scale economies.

13. Military AI: completing democratization or creating a new standardization regime?

AI is not merely another weapon. It can reduce the cost of planning, ISR analysis, translation, programming, logistics, command, and unmanned-system control—the cognitive and organizational capital of warfare.

Personal AI, local models, and cheap robots could amplify small actors. One person able to coordinate many machines may face a lower participation threshold than any previous combatant.

The same AI can also act as a new standardization technology for the state. Common agent protocols, autonomous-weapons standards, machine-to-machine command, automated logistics, mass robot production, and AI battle management could integrate enormous numbers of factories, sensors, and weapons into one system.

14. AI may simultaneously reduce dependence on the masses across multiple domains

AI differs from earlier military revolutions because it may automate not only military service but production, administration, expertise, and information analysis. States and firms could reduce their dependence on large human populations in several domains at once.

At the same time, state power may close a high-speed Observation → Inference → Actuation loop. Better sensing, intent inference, intervention selection, and execution through administrative, financial, infrastructure, and robotic systems could reduce the information gaps, local discretion, organizational defection, and command delays that historically destabilized authoritarian control.

The issue is not coercion alone. If advanced AI delivers stronger economic, medical, logistics, security, disaster-response, and administrative performance, it can strengthen performance legitimacy. If conversational systems infer each person’s values and grievances and generate personalized reasons for supporting the regime, legitimation could become individualized rather than ideologically homogeneous.

High performance, precise persuasion, and precise suppression of a small residual opposition could produce stable concentration without mass terror. Everyday lifestyle freedom could remain broad while constitutional exit—the ability to alter the governing structure—becomes weak.

15. Regional comparison

The tabs below use the same eleven stages as a comparative device rather than a literal universal periodization.

16. Implications for Value Exploration

An independent value-inquiring subject is not merely someone who holds a different opinion internally. Eight billion people could have different values while all information, compute, communications, production, military force, and administration are monopolized by one actor; if no one can try another institution, way of life, research path, or AI without permission, effective exploratory paths may still collapse toward one.

Value exploration may therefore require not only cognitive diversity but causal autonomy: the ability to think a different hypothesis, use a different AI, try a different institution, control some resources independently, and—within safe bounds—refuse the ruling order.

Political redundancy in an AI era is consequently not only a democratic problem. It can be a loss of independent causal pathways through which civilization discovers unknown values, institutions, and forms of life.

17. What does not follow

• Military necessity does not automatically produce democracy; total-war states included communist and fascist regimes.
• Flat geography does not automatically produce despotism; geography interacts with technology, institutions, and historical path dependence.
• Infantry does not necessarily decentralize, nor cavalry necessarily centralize; centrally equipped mass infantry can support very strong centralized states.
• Reduced material dependence on the public does not imply inevitable loss of rights; established institutions and political culture have their own persistence.
• AI does not necessarily centralize power. Personal AI, local compute, independent energy, robotics, and production could have the opposite effect if widely distributed.

The limited claim is that who the ruling order materially needs is one important condition shaping bargaining power and the forms of legitimation that order must provide.

18. Provisional conclusion

Political freedom and inclusion may be impossible to understand as a history of ideas alone. Rights, redistribution, and citizenship may have partly rested on material conditions under which states could not fight, tax, produce, or govern effectively while ignoring broad populations.

Two infantry waves appear to contain moments when old expensive military elites lost relative advantage and some middle strata or corporate communities gained temporary autonomy. In both cases, the window could close once large states learned to procure, supply, and standardize the new technology at scale.

The AI era may repeat that pattern. Cheap drones, open-source tools, and commercial ISR may be opening a third capability-democratization window; yet AI could also integrate millions of weapons, sensors, factories, and administrative processes into a standard system whose scale advantages dwarf earlier centralization.

Because AI may simultaneously reduce dependence on ordinary people in the military, production, administration, and cognition while increasing coercive capacity, performance legitimacy, and persuasive capacity, political freedom may increasingly depend on the material distribution of AI, compute, energy, production capacity, independent organization, and credible exit.

Keeping value exploration open may therefore require more than allowing dissent. It may require that actors who dissent retain enough independent causal capacity to keep affecting civilization’s feedback loops.