The Neolithic transformations: cultivation, domestication, surplus and the changing architecture of power

1. Six changes commonly confused with one invention

The analytical vocabulary matters. Wild harvesting takes naturally reproducing resources. Cultivation manipulates where, when and how plants grow; people can sow or tend plants that remain morphologically wild. Domestication is an evolutionary and demographic process in which selection under human management changes inherited traits over generations. Agriculture is an economic organization in which cultivation and, often, livestock production become central to the subsistence system. Sedentism concerns the length and intensity of residential occupation. Political centralization describes concentrated authority over decisions, force and obligations. These processes may interact without arriving together. [1][5][18]

A plant can be collected, processed, stored and replanted before a non-shattering seed head becomes common. A community can settle for intensive fishing while remaining non-agricultural. A pastoral household can travel seasonally with domesticated herds. A grain-growing village can lack kings and bureaucrats. The difference is not semantics: bundling these separate dimensions falsely transforms an excavated granary into proof of a stratified state and an early cereal trace into a complete economic revolution.

Archaeologists often distinguish pre-domestication cultivation from fully domesticated forms using plant morphology, dated associations and evidence of weed communities, tools or managed soils. Zooarchaeologists examine animal mortality profiles, sex and age structure, changes in body size and the geographic displacement of managed stocks; genomes add another line of evidence, but ancestry alone does not establish the contractual relation between people and animals. Inferences are strongest where independent methods agree.

2. Why the southern Levant complicates the classic sequence

The Natufian tradition of the southern Levant illustrates substantial investment in residence, grinding equipment and harvesting of wild plants before the widespread appearance of morphologically domesticated crops. Climate variability near the Pleistocene–Holocene transition changed access to different resources, but neither the Younger Dryas nor the onset of warmer conditions can explain every local institutional choice by itself. Communities could intensify gathering, construct durable facilities, move or experiment with planting in response to environmental and demographic pressures. [1][2]

At Dhra', near the Dead Sea, excavators documented purpose-built granaries in a pre-domestication context dated approximately 11,300–11,175 calibrated years BP. The floor was suspended to improve circulation and reduce rodent access, demonstrating investment in the engineering of future food supply before fully domesticated cereals became established in the region. The research reported at least four granaries within a very small excavated sample—about 3% of the roughly 6,500 m² site—and residential or processing structures nearby. This is direct evidence of storage infrastructure, not a defensible estimate of the village's total food stocks. [1]

The same study interprets a broad shift from storage outside dwellings toward storage within households in later Neolithic contexts. That spatial change suggests possible alterations in access and ownership, but does not prove that exterior structures were communally owned or that interior ones became legally private. The material record provides the location of the store; the governing rule requires additional evidence.

3. Grain did not automatically create taxation

Stocks transform a seasonal harvest into a resource usable weeks or months later. They can support households during shortages, seed subsequent cultivation, feed visiting allies or sustain workers engaged in construction. They also require investment: drying, containers, protection from insects, maintenance, counting and watchfulness. The return is the security of deferred consumption; the cost is a fixed asset that may have to be abandoned when people need to move. [1]

A powerful actor can exploit this concentration if they control access to a granary, the fields that replenish it or the security arrangements around it. However, the same arrangement can be governed by shared obligations and mutual inspection. To identify taxation one would need evidence of enforceable obligations, allocation to a recognized authority and a system able to maintain that extraction; a granary by itself supplies none of those elements.

The central historical discontinuity lies in who can refuse whom. A household whose food and stored seed are tied to one location faces different exit costs from a household with accessible alternative territories. Those costs can strengthen local claims to land and labor even without formal land markets. Yet villages can preserve collective rights, and people with farming skills can migrate. Agriculture widens the possibilities of durable appropriation; it does not mechanically create a treasury.

4. A world of multiple beginnings

Archaeological and palaeoecological syntheses identify numerous possible domestication zones and warn against treating every crop as a separate independent “agricultural revolution.” The widely cited global land-use synthesis discusses as many as 24 plausibly independent centers of early domestication, a broad classification involving different taxa and disputed degrees of independence, not 24 identical state-forming packages. At least several regional pathways developed independently or through complex mixtures of local experimentation and diffusion. [3]

The chart visualizes how far apart documented regional contexts occur, not how much each society produced or how “advanced” it was. Some markers are dated archaeological objects, others population-genetic model estimates; they must never be added, averaged or interpreted as synchronized transitions. [1][3][4][5][6][7][8]

5. North China: millet and the labor of dryland cereals

The Cishan assemblages in northern China have been used to identify common millet as a major resource approximately 10,000 calibrated years ago. Micromorphological and phytolith research supports the importance of millet in semi-arid northern landscapes, where a resilient warm-season crop had different requirements from wetland rice. Plant exploitation and management nevertheless need to be distinguished from a fully documented agricultural society with known field boundaries and hereditary property rights. [4]

Cereals change the labor calendar. Clearing, soil preparation, sowing, protecting seedlings, weeding, harvesting, threshing and grinding compete with fishing, hunting, child care and maintenance. Output from an area may rise while total hours of work rise too. Cultivation therefore raises at least three separate questions: yield per hectare, calories per hour of labor, and the distribution of time and benefits among households. None can be reduced to the quantity of millet remains found in a pit.

Northern and southern China also illustrate why writing about “the Chinese Neolithic” as a single homogeneous event is misleading. Millet and rice occupied distinct ecosystems and accumulated technologies and institutions on different timelines. Connections and later combinations matter, but the existence of several crops and local pathways is not proof of one centrally coordinated origin.

6. Lower Yangtze: rice selection as a prolonged evolutionary process

At Shangshan and Hehuashan, dated rice-related phytoliths indicate early Holocene exploitation. Research places rice evidence at around 9,400 and 9,000 calibrated years BP in specified contexts, with morphological indications of selection. A separate investigation of stone tools dated roughly 10,000–7,000 BP found microwear and phytolith residues consistent with different harvesting practices. Those findings are important because they document techniques and plant management rather than relying on a single seed measurement. [5][19]

At Huxi, rice spikelet bases from around 9,000–8,400 BP included wild shattering, intermediate and non-shattering forms. The authors reported the non-shattering domestic form as 8.7% of the sampled bases, indicating selection underway but not completion. Other lower Yangtze sites record changing distributions over subsequent millennia. The assemblages cannot be converted into an exact share of household calories or a percentage of total agricultural land. [5]

A field of rice requires water, labor and recurrent attention, but those inputs varied across early wetlands. The spread of rice harvesting methods and selection could occur before large engineered paddies or formal political control. Rice history is therefore not a single instant at which wild plants became domesticated and society became a state. It is a sequence of ecological management, reproductive selection, new tasks and changing settlement choices.

7. Kuk Swamp: agriculture through engineering rather than grain kingdoms

At Kuk Swamp, in the highlands of present-day Papua New Guinea, excavations and palaeoenvironmental investigations provide an unusually well-resolved sequence. Early wetland-margin plant use and some cultivation are placed around 10,220–9,910 calibrated years BP. Mounded cultivation appears around 6,950–6,440 cal BP, and constructed drainage ditches around 4,350–3,980 cal BP. Research identifies taro use in the early Holocene and intensive banana cultivation by the mound phase. UNESCO recognizes a locally developed agricultural landscape remodeled across millennia. [6][20]

The engineering path differs from the textbook image of grain farmers building a city on a river plain. Wetland crops reward drainage, controlled water tables, mounds and maintenance of soils. This creates collective-action problems: who clears ditches, who benefits from the drainage and how is the burden shared across households? A ditch can be built by related families, shifting labor groups or a hierarchical authority; its existence alone cannot distinguish them. The long sequence at Kuk shows gradual adaptation, not a single event called “the birth of agriculture.”

8. Mesoamerica: maize transformed slowly into a staple

At Xihuatoxtla in the central Balsas River valley of Mexico, starch grains and phytoliths were interpreted as evidence of maize use by approximately 8,700 calibrated years BP. The study also identified phytolith evidence for early domesticated squash in a pre-ceramic setting. These materials connect early plant domestication to a seasonal tropical forest environment and challenge accounts that place its origins exclusively in cooler, semiarid highlands. [7]

The economic significance of maize did not arrive fully formed with its earliest traces. Domestication reshaped inherited traits under repeated selection; the relative contribution of maize to diets could expand over very long periods. One ancient starch grain does not measure the fraction of a village's production devoted to maize, and later maize-centered societies cannot be read backward into the first cultivation contexts.

This distinction also matters politically. The crop that would eventually support highly elaborate Mesoamerican agricultural systems and urban states was not itself a government-making device. Demographic change, irrigation in some regions, exchange, conflict, customary rights and central administration developed through distinct later histories. The early Balsas evidence establishes plant use and selection, not pre-classical taxation or an ancestral royal house.

9. Africa and the Sahel: indigenous innovation, not passive diffusion

African agriculture cannot be reduced to the later arrival of Southwest Asian or Eurasian crops. Genomic analyses of pearl millet support a western Sahara domestication origin, with estimated cultivated expansion beginning about 4,900 years ago. Gene flow from wild relatives contributed to diversity in western and eastern Sahelian populations, complicating any notion that a crop was invented once as a genetically closed package and simply exported unchanged. [8]

Archaeology, genetics and palaeoenvironmental records answer different questions. Genomics identifies related populations and models divergence, admixture and selection; charred seeds and impressions place use at particular sites; settlement and tool assemblages reveal the wider economy. A genetic estimate of expansion should not be represented as an excavated farm with a precisely dated ruler.

The Sahel's changing rainfall, seasonal pastures and geographic breadth favored combinations of grazing, cultivation, fishing and foraging that differed across regions and centuries. Later states and trading networks arose under political and ecological conditions far removed from the first plant-selection episodes. Agricultural innovation was not monopolized by the Fertile Crescent; neither did independent domestication inevitably produce immediate urban centralization.

10. The Andes: vertical ecologies and domesticates beyond cereals

In the Central Andes, cultivation unfolded across steep altitudinal gradients and very different productive niches. An Oxford synthesis associates early highland cultivation of wild ancestors of potato, oca, quinoa and related crops with developments around 6,000 years ago, followed by gradual selection and elaboration of crop–pastoral systems. Archaeological and palaeoecological evidence also records interactions among maize cultivation, camelid herding and soil transformation over the Holocene. [9][21]

Unlike the cereal-focused narratives of West Asia, highland tubers and chenopods created different storage, processing and risk-management problems. Freeze-dried potatoes, later known as chuño, demonstrate how altitude, night frost and human processing can contribute to food security; the existence of this later technology does not prove a precise invention year. Camelids could provide transport, meat, fibre, dung and capital embodied in living herds, with different property and mobility dynamics from fixed grain fields.

The Andean record underscores a larger principle: domestication changes ecological relationships, not just plant genetics. Corrals can fertilize surrounding soil; terraces can alter runoff; storage can smooth harvests across seasons. Institutions controlling those assets may be collective or unequal. None of the surviving evidence licenses a single timeline in which every Andean community became agricultural, sedentary and hierarchically ruled simultaneously.

11. Domesticated animals were living assets with their own risks

Goats and sheep were brought under human management in parts of Southwest Asia around 10,500–9,500 years ago, with regionally varied pathways. At Ganj Dareh in the Iranian Zagros, evidence of goat management and an ancient genome around 10,000–9,700 cal BP illuminate early pastoral lifeways. Zooarchaeological synthesis from the southern Levant suggests local management practices and possible later introductions; a single uniform dispersal from one village is not an adequate explanation. [15][16]

Herding transforms economic constraints. Animals reproduce, die, require feed and water, escape, become sick and can be slaughtered or transferred. Keeping breeding females changes immediate meat consumption into potential future herd growth. The returns depend on access to pasture, shepherding labor, disease control and the ability to defend animals. Herd size is a possible wealth proxy, but bones from a disposal pit do not reconstruct one household's ownership register.

Secondary products widened possible returns. A 2026 study reported direct evidence for caprine dairy exploitation in the Iranian plateau from the seventh millennium BCE. Such evidence does not mean that milk was universally important at the start of herd management; exploitation of meat, skins, fibre, dung, traction and dairy products developed with different technologies and social preferences. [22]

12. Çatalhöyük: dense settlement without a simple class-state equation

The eastern mound of Çatalhöyük in Anatolia contains 18 archaeological occupation levels dated approximately 7400–6200 BCE, according to UNESCO. Its closely packed houses and roof-based access show an unusual built environment and long-term collective life. These features challenge any claim that a large settled farming community must reproduce a street-grid city with palaces and a visible central administration. Buildings contained household activities and symbolic representations, but their interpretation remains subject to contextual archaeological debate. [10]

A multi-dataset study published in 2024 examined inequality at Çatalhöyük using differences among contemporary buildings, changes through time and material measures. It reported broadly similar access to staple foods, variation in less essential goods, and no unified temporal trajectory of inequality or clear institutionalized long-lasting economic hierarchy. Particular periods may nevertheless have shown greater differentiation. The investigators' Gini-like measures are defined for archaeological proxies, not an observed income census. [11]

The economic analysis is powerful but bounded. Farming, dense settlement and architectural investment were compatible with substantial social differentiation without a permanently institutionalized elite detectable in those measures. Agricultural technology increased possibilities for hierarchical control; it did not fix their realization. Any serious theory making agriculture alone the direct cause of hereditary aristocracy must account for this contrary case.

13. When inequality grew: scale, land and delayed institutional change

A large comparative study published in PNAS in 2025 found that increasing wealth differences were a long-run but not universal Holocene tendency. Pronounced rises in material inequality often came a millennium or more after plant domestication, associated with greater population scale, land-limited production and settlement hierarchies. The result directly challenges a model in which planting immediately produces elites in every community. [12]

The causal proposition is that larger and more differentiated settlement systems can weaken face-to-face leveling mechanisms. Productive land may become relatively scarce; households controlling favorable plots or labor can secure cumulative advantages; leaders coordinating large populations may develop stronger enforcement capacities. Yet comparative correlations can combine different historical processes, sampling traditions and measures of “wealth.” A relationship with population scale does not prove that crop biology itself generated coercive institutions.

A plausible rival mechanism emphasizes warfare and security: the need to protect concentrated stores or fields could give military organizers greater bargaining power. Another emphasizes inherited networks of ritual authority, trade and marriage. These mechanisms can reinforce one another, and their chronology must be demonstrated, not assumed. Stronger tests compare dated household assets, landscape constraints and the timing of institutional changes within connected archaeological regions.

14. The labor ledger: more food from land can mean more work per person

Agriculture can support more people on a given area while leaving open whether each person works fewer hours or eats a more diverse diet. Crops require clearing, sowing, weeding, guarding, harvest processing, grinding, water management and seed preservation. Herds require fodder, care and movement. Fixed structures need repair. The economic surplus remaining after these costs depends on technology, seasonality, soil quality and who must supply the labor. [13]

Research on the bones of prehistoric women from Central Europe found patterns consistent with prolonged intensive manual work, including grain grinding and other repetitive activities, across the early millennia of farming. These results are based on skeletal measurements and comparative activity patterns, not eyewitness diaries or a universal assignment of every prehistoric task by sex. The study underscores how production systems can change physical workloads while archaeological narratives focus narrowly on food quantity or impressive settlements. [13]

A household might gain the security of cereal stores but bear a much heavier burden of milling and transport. Another could organize shared processing and redistribute those costs. In either case, the total harvest cannot be equated with the welfare of the workers who produced it. The unit of analysis must include labor time, injury, care for dependents, seasonal critical constraints and rights to the crop.

15. Population and health: a demographic dividend with biological costs

Settlement and increased food production can alter fertility, mobility and the age structure of populations. The Neolithic Demographic Transition hypothesis relates a rise in fertility in many early farming contexts to changes in maternal energetics, child transport and food availability. Skeletal-age profiles provide indirect evidence, but burial selection, mortality regimes and sample size constrain identification. A higher birth rate is not proof that an individual woman's workload or health improved. [17]

A study integrating skeletal and ancient-genetic data on 167 prehistoric Europeans reported relatively reduced stature among early farmers after accounting for genetic contributions to height. The finding is compatible with nutritional stress and/or greater disease burdens in those particular samples; stature is not a direct and complete measure of subjective wellbeing, diet, infection or longevity. Studies of bone loading likewise reflect patterns of behavior rather than a literal tally of work hours. [14]

Dense settlements can bring parasites, waste-management problems, respiratory exposure and shared human–animal pathogens, while improving mutual aid, care and access to stored food. Biological costs also vary by the crops, climate, sanitation and historical contact systems involved. “Farming made everyone sick” is as sweeping as “farming made everyone prosperous.” Different populations and periods followed different paths, and political choices about sanitation and distribution helped shape the outcomes.

16. Environmental engineering, resilience and lock-in

Cultivation modifies land. Clearing changes vegetation and erosion; repeated fields alter soil and nutrient balances; channels and ditches shift water; herds change grazing pressures and seed dispersal. These interventions can increase productive capacity and expose people to new dependencies. The long-term Andean landscape record, for example, documents interactions among cultivation, pastoralism, erosion and later terrace management. The Kuk sequence shows centuries of progressively modified wetlands rather than a simple one-time “conversion” from nature to agriculture. [6][21]

The economic problem resembles investment under uncertainty. Maintaining a drainage system or terrace demands labor today for future returns that may fail if rainfall changes, soil degrades or a group moves. Highly productive infrastructure can become a trap when the cost of abandoning it rises. On the other hand, managed landscapes can reduce climatic vulnerability through diversification, soil protection and storage. The direction is case-specific.

A conditional comparison is more useful than declaring farming environmentally destructive or protective by nature. Where intensive cultivation depletes soils faster than they recover, households face falling returns unless they change techniques or move. Where water-control or soil management is sustained by coordinated labor, gains can persist. The political issue is whether the beneficiaries and the maintainers of infrastructure are the same people, and who bears losses when it fails.

17. Exchange and migration: crops move, institutions need not

The spread of a plant is not the spread of a complete society. Seeds can move through exchange, marriages, visiting groups, migration, conquest or shared experiments. People can adopt crops without adopting a donor community's language, religion, land rules or social rank. In Central Asia, newly documented 9,000-year-old barley consumption in southern Uzbekistan is particularly instructive: cereals and related tools appear among communities whose wider subsistence history is not reducible to a simple package of Neolithic farmers advancing uniformly across Eurasia. [23]

Ancient genomes can illuminate population movement and mixtures, but genetic ancestry is not equivalent to a legal title or political continuity. The Ganj Dareh genome shows that early southwestern Asian livestock economies involved genetically differentiated populations; diffusion of animal management did not require every recipient to be descended from a single farming people. [15]

A good historical comparison distinguishes the location of earliest evidence for cultivation, the region of biological domestication, the route of later diffusion and the period when a crop became economically dominant. These dates often differ by millennia. Political institutions may travel with migrants, emerge locally or be transformed by contact. Reducing the spread of farming to a single arrow on a map hides that complexity.

18. Methodological comparison of evidence and uncertainty

The visual comparison intentionally juxtaposes contexts whose dates are not fully synchronized. Dates described as cal BP use calibrated estimates before the conventional 1950 reference; BCE chronologies are distinct expressions of time. A crop's first presence, the earliest evidence of cultivation and its biological domestication must not be mapped as one event. Neither chart is a series of prehistoric GDP, productivity or population.

19. The political economy of surplus without inevitable kings

Agricultural production enlarged the range of possible relations between persons, land and stores. A cultivator could invest in soil preparation and expect to harvest in a future season; a household could maintain a pantry, or a community a protected granary; a herd could multiply while creating obligations to watch and feed it. Recurrent work on the same landscape also made claims of access, exclusion and inheritance more consequential. But enforcing those claims required relationships, norms and sometimes coercion that are analytically distinct from farming.

The most useful comparison is between the technical possibility of surplus and the political capacity to capture it. A stock of cereals can stabilize food availability when managed for mutual insurance. It can produce dependencies when controlled through debt, force or inherited privilege. Larger populations may support specialists, but specialization can be reciprocal, seasonal or unequal. Early villages could redistribute, differentiate or contest authority; the documented record does not support a universal linear passage from farmer to chief to king.

The transition thus created neither a single superior way of life nor an immediate class state. It changed the structure of risks and possible bargaining positions. The subsequent rise of cities and organized states—where it occurred—requires separate explanation of demographic concentration, administrative technologies, control of force, transport, institutions of property and uneven rights to the product of collective labor. The Neolithic is not the first chapter of inevitable civilization; it is a family of historically contingent transformations that reshaped the material conditions under which power could be exercised.

Sources and methodological documentation

[1] Kuijt and Finlayson, “Evidence for food storage and predomestication granaries 11,000 years ago in the Jordan Valley,” PNAS (2009): https://pmc.ncbi.nlm.nih.gov/articles/PMC2700141/

[2] Smithsonian Human Origins Program, “Humans Change the World,” institutional historical synthesis: https://humanorigins.si.edu/human-characteristics/humans-change-world

[3] Ellis et al., “Used planet: A global history,” PNAS (2013): https://doi.org/10.1073/pnas.1217241110

[4] Lu et al., “Earliest domestication of common millet (Panicum miliaceum) in East Asia extended to 10,000 years ago,” PNAS (2009): https://doi.org/10.1073/pnas.0900158106

[5] Zheng et al., “Rice Domestication Revealed by Reduced Shattering of Archaeological rice from the Lower Yangtze valley,” Scientific Reports (2016): https://pmc.ncbi.nlm.nih.gov/articles/PMC4914946/

[6] Denham et al., “Origins of agriculture at Kuk Swamp in the highlands of New Guinea,” Science (2003): https://pubmed.ncbi.nlm.nih.gov/12817084/

[7] Piperno et al., “Starch grain and phytolith evidence for early ninth millennium B.P. maize from the Central Balsas River Valley, Mexico,” PNAS (2009): https://pubmed.ncbi.nlm.nih.gov/19307570/

[8] Burgarella et al., “A western Sahara centre of domestication inferred from pearl millet genomes,” Nature Ecology & Evolution (2018): https://www.nature.com/articles/s41559-018-0643-y

[9] “Origins of Food Production in the High Andes,” Oxford Research Encyclopedia of Anthropology (2021): https://academic.oup.com/edited-volume/61643/chapter-abstract/539828448

[10] UNESCO, “Neolithic Site of Çatalhöyük”: https://whc.unesco.org/en/list/1405/

[11] Twiss et al., “But some were more equal than others: Exploring inequality at Neolithic Çatalhöyük,” PLOS ONE (2024): https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0307067

[12] Kohler et al., “Economic inequality is fueled by population scale, land-limited production, and settlement hierarchies across the archaeological record,” PNAS (2025): https://pmc.ncbi.nlm.nih.gov/articles/PMC12037049/

[13] MacIntosh et al., “Prehistoric women's manual labor exceeded that of athletes through the first 5500 years of farming in Central Europe,” Science Advances (2017): https://pmc.ncbi.nlm.nih.gov/articles/PMC5710185/

[14] Cox et al., “An integrative skeletal and paleogenomic analysis of stature variation suggests relatively reduced health for early European farmers,” PNAS (2022): https://pmc.ncbi.nlm.nih.gov/articles/PMC9169634/

[15] Broushaki et al., “The genetics of an early Neolithic pastoralist from the Zagros, Iran,” Scientific Reports (2016): https://www.nature.com/articles/srep31326

[16] Munro et al., “The Emergence of Animal Management in the Southern Levant,” Scientific Reports (2018): https://www.nature.com/articles/s41598-018-27647-z

[17] Bocquet-Appel, “Explaining the Neolithic Demographic Transition,” Current Anthropology / thematic analysis (2008): https://pmc.ncbi.nlm.nih.gov/articles/PMC7120555/

[18] Fuller and collaborators, “How did the domestication of Fertile Crescent grain crops increase their yields?”, experimental research abstract: https://pubmed.ncbi.nlm.nih.gov/28286354/

[19] Wang et al., “New evidence for rice harvesting in the early Neolithic Lower Yangtze River, China,” PLOS ONE (2022): https://pmc.ncbi.nlm.nih.gov/articles/PMC9728911/

[20] UNESCO, “Kuk Early Agricultural Site”: https://whc.unesco.org/en/list/887

[21] “The Evolution of Agrarian Landscapes in the Tropical Andes,” peer-reviewed palaeoecological study (2024): https://pmc.ncbi.nlm.nih.gov/articles/PMC11013201/

[22] “Caprine dairy exploitation on the Iranian Plateau from the seventh millennium BC,” Nature Human Behaviour (2026): https://www.nature.com/articles/s41562-025-02396-y

[23] “9,000-year-old barley consumption in the foothills of central Asia,” PNAS (2025): https://pmc.ncbi.nlm.nih.gov/articles/PMC12435284/