1. Who Gets to Question the Numbers?

By the end of 2018, India’s National Sample Survey Office had completed its 2017–18 Periodic Labour Force Survey (PLFS) – the government’s own instrument for measuring employment. The report was not released. Weeks later, a version leaked to the press showed unemployment at 6.1%, its highest level in forty-five years. Two members of the National Statistical Commission, which had approved the report, resigned over its withholding.¹ The immediate dispute was not over a mathematical error in the figure, but over whether the public would be allowed to see it.

The PLFS controversy showed how much political weight a number can carry. Numbers are central to public debate: debt ratios, growth rates, unemployment rates, poverty lines, price indices, and export values help decide whether an economy is ‘recovering’, whether workers are ‘employed’, and whether public spending is treated as ‘affordable’. But withholding a number is only one way to control its political meaning. Numbers can also obscure reality through the definitions, classifications, comparisons, and formats that make them appear neutral.

Unemployment rates are a good example of what numbers can reveal and obscure at the same time. An unemployment rate is never a simple head count of the jobless. It is a ratio between those classified as unemployed and those counted as part of the active labour force. That ratio depends on definitional boundaries: who qualifies as employed, who is actively seeking work, who has stopped searching, who is immediately available to work, and which forms of informal or unpaid labour fall outside the metric.

The rate can fall because employment has genuinely expanded, but it can also fall because discouraged jobseekers stopped searching, because precarious informal work is counted as employment, or because underemployment has been swept into a different statistical bucket. On its own, the number cannot tell us which underlying social process it records.

These examples point to a deeply political problem. Accuracy is only part of it. The deeper question is who has the power to interrogate a number. Who can probe its definitions? Who has the means to audit the denominator, distinguish a count from a rate or an index, or tell a nominal value from a real one? Who can trace a polished indicator back to the realities of labour and class, gender and region, debt and the world market?

These questions have real-world implications that affect the lives of billions. International Monetary Fund (IMF) surveillance and its formally voluntary data-dissemination standards shape the formats through which states report economic data; World Bank lending has financed and shaped the development of national statistical systems; the United Nations’ System of National Accounts defines the international standards for measuring gross domestic product (GDP); and sovereign credit ratings from Moody’s, S&P, and Fitch influence borrowing costs through methodologies whose final judgements are not fully open to public scrutiny.²

This concentration of statistical authority leaves researchers in the Global South in a bind. They may ask how sovereign debt constrains public spending, or how commodity prices bind national production to an unequal world market. But the data through which they must answer those questions are often organised elsewhere, sometimes by the very institutions and standards they seek to interrogate. The pattern is not new. Raúl Prebisch’s account of core–periphery terms of trade³ and Walter Rodney’s account of colonial economies organised for extraction⁴ point to an older problem: economic domination shapes not only the movement of commodities, labour, and capital, but also the terms through which those movements are counted, compared, and understood. The question today is how that older problem reappears in datasets, indicators, metadata, and institutional standards. Nor does the Global South confront this problem from a single standpoint. A finance ministry technocrat, a business lobby, and a union research department do not approach IMF numbers with the same interests. The question is therefore not only where statistical authority is located, but whose interests its categories and measures serve.

For researchers, movements, and public institutions that have long had to rely on metrics pre-packaged by dominant institutions, the challenge is not simply a lack of data but unequal power to turn raw numbers into evidence. The emergence of AI tools changes this terrain by lowering the cost of that work. But AI’s political significance can only be understood by first looking at the labour that comes before analysis: the choices, checks, and transformations that decide what a number can be made to show.

2. Numbers Are Not Evidence Until Someone Makes Them Evidence

Quantitative research is often imagined as a linear process: the data exist, then they are cleaned, analysed, visualised, and interpreted. This sequence feels intuitive, but it starts too late. Much of empirical inquiry happens before formal analysis begins, in the laborious effort to determine whether a given figure can answer the question in front of it.

A cell value in a table is not yet evidence. It becomes evidence only when connected to a specific research question, a rigorous definition, a unit of observation, a standardised measure, and a basis for comparison. A number may be valid inside its own statistical system and still fail to answer the political question being asked. Recognising when a national total hides deep distributional disparities, or when a trend in money terms hides a change in what that money buys, is part of the conceptual work of quantitative inquiry. So is seeing when a formal head count of workers leaves out labour that is socially necessary.

This connective labour – the work that bridges raw figures and substantive claims – is measurement work. It is the labour of transforming raw administrative or statistical outputs into evidence. Its inputs might be a PDF table, a brief metadata note, or an isolated time series. Its outputs are explicit variables, verified transformations, documented sources, and defensible comparisons.

The term is specific to data, but the kind of translation it refers to has a longer history. Bowker and Star showed that classification systems are built through political and institutional labour.⁵ Marx showed how capitalism translates different kinds of concrete labour into abstract, exchangeable value.⁶ Measurement work names a related translation: the process by which raw figures are made into evidence that can circulate as public truth.

Naming this labour matters because it reframes the primary question from ‘do we have the numbers?’ to ‘what has been done to make these numbers usable as evidence?’

GDP, for instance, can be expressed as an absolute level, a year-on-year growth rate, a per capita value, a nominal series, or a deflated real index. A poverty line can make material deprivation visible, but where it is drawn also decides how many people count as poor at all. A consumer price index captures inflation, yet its political implications diverge dramatically depending on whether the focus is on food security, fuel costs, real wages, or basic household survival. These are not minor technical adjustments. They decide which parts of social life a number can show, and which it leaves out.

Take the World Bank’s ‘Pink Sheet’ on commodity prices.⁷ The downloaded annual workbook is a complex file containing multiple sheets: nominal prices, real prices, price indices, and weight structures. A researcher studying agricultural dependency has to choose which version of the price to use. Should they track the price as it appeared at the time, or adjust it to show its purchasing power across years? Which commodity and unit of measure fit the question? This selection is measurement work. It can make the same decades look like recovery – or like a long decline in what those prices could actually buy.

Similarly, a basic GDP growth series can support very different, even contradictory, arguments depending on how it is handled. While a raw annual rate highlights short-term volatility, a three-year moving average smooths out short-term fluctuations to expose cyclical trends, and a five-year average makes longer trends easier to see (Figure 1). The underlying data remain the same, but the choice to aggregate or smooth them is a conceptual intervention. Without the technical capacity to construct these alternative measures, a research team is forced to inherit the version used in official press releases to frame national performance.

Figure 1. Real GDP growth, read as the annual rate and three- and five-year moving averages (illustrative constructed series, per cent per year; not any specific country).

None of this makes statistical metrics meaningless. It means their authority rests on the chain that connects a raw value to a public claim.

Yet this chain is deeply unequal. Well-funded institutions can spread the pre-analytical labour across dedicated data teams, advanced software pipelines, and robust documentation systems. Researchers, social movements, and public agencies in the Global South often have to carry the sequence themselves: locating raw files, auditing sources, deconstructing tables, writing code, validating totals, and defending their findings in public debates.

This division of labour determines which questions can be asked in the first place. If researchers cannot verify the construct behind a number, an inquiry may never progress to formal analysis. If a spreadsheet cannot be reshaped into structured evidence, the patterns inside it may remain locked away from public debate. When only a handful of well-funded organisations can perform this translation at scale, they retain the power to define economic and social reality through numbers.

Measurement work is a cornerstone of quantitative research capacity. It is not a preliminary chore, but one of the conditions that make empirical inquiry possible.

Yet this foundational labour is routinely erased. The polished indicators in global databases appear as finished facts, with the decisions behind them pushed out of view. This erasure creates two barriers. It discourages researchers from interrogating official statistics, and it leaves them without practice in handling raw administrative data. Often, the numbers exist, but they arrive in formats that are hard to question. A table may be easy to read and still be difficult to check, compare, or reuse.

3. A Readable Table Is Not Yet Data

The gulf between a presentation table and an analysis-ready dataset is visible in a typical worksheet on labour-force indicators (Figure 2). The sheet looks orderly: the two sexes run across the top, sectors down the left, and three related blocks – labour-force participation, work, and unemployment – are stacked one under another. A researcher can read it and a spreadsheet program can display it. But tools such as R, Python, or SQL cannot safely treat it as data, because the meaning of each value is carried by position, formula, and block order rather than by explicit variables.

Figure 2. Labour-force indicators by sector and sex, India PLFS 2017–18 (presentation worksheet; usual status – principal plus subsidiary activity; persons aged 15 and above; rates in per cent).⁸

In the example above, the labour-force participation rate, the worker–population ratio, and the unemployment rate appear in matching positions across separate blocks. To understand how they relate to one another, the reader has to follow the layout and formulas by hand. That may be enough for one careful reading, but it is not enough for analysis that another researcher can repeat and verify.

An analysis-ready dataset has to be organised differently. The meaning of each number must be built into the dataset itself: the sector, sex, measure, unit, and value should appear as explicit fields, not as clues scattered across the layout. Definitions and formulas should travel with the measure as metadata, rather than be inferred from where a number sits in the grid. The question shifts from ‘can I see the right number near the right label?’ to ‘can another researcher check, group, compare, and reproduce the record without guessing?’ Reshaping the sheet into that form (Figure 3) turns a human-readable layout into machine-checkable evidence.

Figure 3. From workbook layout to an analysis-ready long table, India PLFS 2017–18 (usual status – principal plus subsidiary activity; persons aged 15 and above; rates in per cent; data and definitions as in Figure 2).

Once the measure is explicit, the evidence chain can be verified and inspected. Code can test whether totals add up, whether categories repeat consistently, whether rates match their inputs, and whether cases are missing. The record that travels with the dataset – source, layout, definitions, formulas, exclusions, checks – lets another researcher trace each step.

Without that structure and record, a chart may look polished while resting on a number whose meaning was guessed from its place in the table. The evidence cannot be independently checked or reliably reused beyond the person who cleaned it. The same opacity recurs across government reports, international databases, survey summaries, commodity-price sheets, and statistical PDFs – wherever the conditions that make numbers usable as evidence are hidden in layout, metadata, definitions, and institutional practice.

4. AI Lowers the Cost of the Work before Analysis

In this context, AI matters because it lowers the cost of pre-analytical labour. It is most useful for researchers who have a clear question, understand the statistical constructs, and can judge the political stakes, but lack years of formal training in data engineering.

For resource-constrained research teams, a major bottleneck is the difficult transition from a published table to a dataset that can be searched, checked, and analysed. Researchers have to identify the implicit variables, write the code that reshapes the table, verify the formulas, and record where each value came from. Many promising inquiries are abandoned at this stage not because the question is weak, but because the path from source format to evidence is too costly.

AI can help at several points in this transition, so long as each step remains open to human inspection.

Reading the Layout. AI can reduce the friction by helping researchers map the structure of the source data. A researcher can describe a table in plain language – quarters across columns, age cohorts down rows, sex categories repeated under each period, and measures stacked in blocks – and ask the system to propose a structure: period, age group, sex, measure code, unit, and value. The point is not to outsource the design, but to get a first draft that can be inspected, corrected, and finalised faster than starting from a blank script.

Drafting the Code. AI can help turn the researcher’s plan into a first working script. A researcher can explain what needs to happen to the table: move repeated headings into their own columns, combine related tables or lookup lists, remove footnote markers and other recurring text patterns, make labels consistent, turn numbers stored as text into usable values, or record where the original figures came from. The script still has to be checked against the source. But AI can lower the barrier between knowing what needs to be done and being able to run the code that does it.

Running the Checks. AI can also help with verification. A researcher can ask it to generate validation checks: Do sex categories add up to the listed totals? Do regional sub-aggregates reconcile with national aggregates? Does a rate match its numerator and denominator? Were any rows dropped during the reshaping process? These checks are the primary way in which a small team can prevent an empirical claim from drifting away from its material grounding.

Documenting the Work. AI can help turn a messy, interactive data-cleaning session into a structured record of the work: which sheet and cell ranges were read, which rows were excluded, how variables were renamed, what assumptions were made, and which validation tests succeeded. That record can be written in plain Markdown, JSON, or any other format the team can review. Without it, the analysis remains an isolated craft. Researchers are forced to reconstruct the same steps again, while institutions with stronger data-management systems retain the authority to define the mainstream narrative.

Diagnostic visualisations extend the same logic. Line charts, bar charts, and histograms can reveal structural breaks, duplicated periods, missing categories, or unusual outliers while evidence is still being made, before a claim has been written around a mistaken pattern.

One political caution belongs here: cleaning sensitive administrative data with a foreign AI tool can make those data newly legible to the company running the model, depending on its retention practices. The same act can build research capacity and open a new point of extraction.

Taken together, these steps place AI inside the measurement workflow rather than above it (Figure 4). AI can lower the cost of moving from a source to a structure, a set of checks, and a record. But the work that gives numbers public meaning – confirming the source, checking definitions, and interpreting the result – stays in human hands.

Figure 4. Where AI enters the measurement workflow (schematic; no data).

AI can automate parts of the quantitative workflow, but it cannot replace quantitative research as a whole. Nor does it make programming obsolete. The political opening is that researchers can spend less time fighting syntax and more time on statistical definitions, conceptual design, and political-economic interpretation.

In practice, this means that local research teams, trade unions, and regional public institutions in the Global South can now transform dense, poorly formatted administrative reports into auditable databases that others can inspect, reuse, and contest. This extends the Bandung Circuits’ AI for Social Science (AI4SS) framework into the terrain of measurement, applying its concern with research sovereignty to the work of turning numbers into evidence.⁹

5. AI Cannot Replace Judgement

Making data preparation easier is not the same as replacing intellectual judgement.

An AI system can reshape a file without comprehending its substance. It can infer a tabular layout without knowing whether a given value is a numerator, denominator, total, rate, share, or index. It can produce a clean dataset from a flawed premise. It can accelerate research, but it can also accelerate structured error.

The core principle remains: AI can transform the file, but the analyst must verify what the file actually represents.

Verification is where statistical definition, economic interpretation, and political responsibility converge. Researchers must still decide what activities qualify as labour, which denominator belongs to a rate, whether historical categories remain comparable over time, whether a nominal series must be adjusted for inflation, and whether a chosen price index genuinely addresses the research question. They must also ask whether an official indicator obscures informal work, unpaid care, ecological degradation, or class distribution.

The same limit appears in research on informal work. An AI system can isolate categories such as self-employment, own-account work, or contributing family labour. But it cannot determine whether a category is best understood as an independent livelihood, disguised wage labour, a household survival strategy, or a position of acute subordination within a broader chain of extraction. Translating a statistical category into a concrete social relation requires political-economic judgement. No machine can assume responsibility for how researchers interpret class, value, dependency, or exploitation.

AI can recommend validation checks, but it cannot discern which discrepancy matters politically. A small mismatch in a total may be a rounding issue; a denominator that quietly excludes migrant workers may change the social meaning of the entire rate. AI can render a chart, but it cannot take responsibility for the public claim attached to it.

Using AI responsibly does not mean outsourcing judgement to an algorithm. It means embedding the machine within a disciplined research workflow – checking sources, reconciling totals, documenting each step – and interpreting the result politically. The more powerful these computational tools become, the more clearly researchers must understand their limits.

6. From Data Access to Research Capacity

By this point, the problem is no longer only whether data can be accessed. It is whether researchers have the capacity to turn those data into evidence that can be tested and defended. A dataset may be public and still be politically unusable if researchers lack the time, tools, concepts, and institutional support to work with it. What matters is who can sustain the whole chain: formulating the question, structuring the data, validating transformations, interpreting results, and carrying findings into public debate.

If knowledge production has its own ‘means of production’, then quantitative inquiry is no exception. It requires data and software, statistical concepts and economic theory, infrastructure and documentation – and, above all, institutional time. We call this broader ensemble of conditions the means of quantitative inquiry. It is not a single software package, but the collective capacity to interrogate numbers. Measurement work is a foundational, yet often hidden, layer of this capacity, the point where the structural conditions of quantitative inquiry first become labour.

The central change is not in the questions researchers ask, but in the resources required to answer them with evidence. By lowering that resource threshold, AI can change who builds and defends an evidence base. The first shift is in wider access to materials. A research team that once stopped at a scanned PDF table can now build an auditable database. The second is in deeper access to statistical constructs. A civil society researcher who once had to accept a pre-packaged external indicator can now dismantle and examine its numerator, denominator, and measurement logic. The third is the capacity to build an evidence base of one’s own. A local public institution that once passively consumed international classifications can now test whether those frameworks align with its mandate and build rigorous counter-evidence when they do not. The United Nations Economic Commission for Latin America and the Caribbean shows what that third capacity looks like when sustained at an institutional scale. Prebisch directed it from 1950 to 1963, and it has produced alternative development statistics for the region for nearly eight decades.¹⁰

This is, fundamentally, a question of methodological sovereignty. It does not mean that every institution should invent its own statistics or ignore comparability. It means having the capacity to examine inherited measures, test whether they answer local and regional questions, document alternative choices, and make those choices accountable to public argument. For a graduate student, this may mean reproducing an official chart instead of simply citing it. For a union, research institute, or public agency, it may mean maintaining an alternative evidence base that can survive public scrutiny. Kerala’s People’s Plan Campaign, launched in 1996, is a working example of that balance. Local bodies gathered their own data and set their own development priorities, within a common planning framework that kept plans comparable across the state – and the decentralised planning it built continues under the Kerala State Planning Board.¹¹

Methodological sovereignty matters most where statistics decide a country’s future. In Pakistan – as Tricontinental: Institute for Social Research documents in dossier no. 69, How the International Monetary Fund Is Squeezing Pakistan – the IMF has made its lending conditional on austerity: removing electricity, gas, and fuel subsidies, raising interest rates, and letting the market set the exchange rate.¹² The Fund justifies these conditions through an aggregate assessment of debt sustainability. Its judgement of what is ‘sustainable’ reaches the household as the price of cooking gas and electricity. To show how that burden is distributed, local researchers and trade unions need the capacity to work with household survey microdata, reweight price indices according to local consumption patterns, and estimate the effects of the proposed cuts on different groups, particularly working-class households. Owning the means of quantitative inquiry allows researchers to do more than react to the IMF’s claims: they can challenge its arithmetic with evidence that withstands technical scrutiny and equip unions, movements, and governments prepared to contest its prescriptions.

What AI changes depends on who deploys it. In the hands of dominant institutions, it can produce the same pre-packaged indicator narratives at a faster rate. In the hands of scholars, trade unions, social movements, and public agencies in the Global South, it can widen who is able to build and defend an evidence base. That wider access does not mean lower standards. It requires clearer definitions, more traceable data lineages, more reproducible transformations, more explicit verification protocols, and more responsible interpretation. The objective is not to bend numbers to a predetermined narrative, but to let more researchers check what the numbers can and cannot say.

7. Who Sets the Terms of Counting?

Rigour matters because numbers carry power in public life. But not everything of value should be turned into a metric. Some relations of solidarity and subsistence resist counting, and should. The narrower claim is that whoever sets the terms of counting holds real power, and that power should not sit only in the North. That is the test to hold AI to as well: not AI thinking for researchers, but more researchers gaining the conditions to think with numbers.

Who gets to interrogate the numbers? Who gets to turn numbers into evidence? Who gets to use that evidence to challenge frameworks arranged by others? In the Global South, methodological sovereignty begins there: in the capacity to question inherited numbers, defend the evidence built from them, and mobilise it for political and social emancipation.

Notes

Collage by Tricontinental: Institute for Social Research drawing from a Dymaxion map projection and Islamic, Mesoamerican, Ethiopian, and Chinese maps made between the twelfth and seventeenth centuries.

Images sourced from: Muhammad al-Idrisi, Tabula Rogeriana, 1154, 1929 reconstruction and transliteration by Konrad Miller via Wikimedia Commons; Matteo Ricci and Li Zhizao, Kunyu Wanguo Quantu (坤輿萬國全圖), 1602, Library of Congress Geography and Map Division via Wikimedia Commons; Códice Xólotl, Lámina 1, Tetzcoco, Mexico, c. 1541 via Wikimedia Commons; Tigrayan circular map from the Māshafā Aksum (Book of Aksum), pre-eighteenth century, Bibliothèque nationale de France, Collection Antoine d’Abbadie, via Wikimedia Commons; and Justinkunimune, Blank World Map, Dymaxion Projection, 2023, Wikimedia Commons, CC0 1.0.

The maps in this dossier trace how the hegemonic world order has organised and represented the world, from colonialism, the post-World War II international order, and Cold War bipolarity to internationalised capital, the North–South divide, and US unipolarity. More than representations, they distil concepts imperial and colonial powers have used to exercise power. Every dot of ink is a choice: what sits at the centre, which countries appear at the top, and which distortions are considered acceptable. The cover gathers cartographic traditions with different centres, orientations, and cosmologies, asking how, beyond poles and blocs, we might understand the unequal social relations that constitute the world order itself.

The Mercator projection was developed in 1569 for European navigation during the colonial period. During imperial expansion, it became the standard map projection, still used today.
This map, which shows the spheres of influence of the ‘great powers’ and their ‘possessions’, was created by H.E.C. Robinson in 1908. Sourced from the National Library of Australia via Wikimedia Commons.

The language of ‘unipolarity’ and ‘multipolarity’ – whether global power is concentrated in one dominant state or distributed among several major states – has become almost unavoidable in political debates. US and allied policymakers have often described the post-Cold War international system as a ‘liberal international order’ increasingly threatened by ‘revisionist powers’ such as China and Russia. Many governments and movements in the Global South, by contrast, welcome what they see as the emergence of a multipolar world in which Washington can no longer dictate the terms of international life. This aspiration is understandable. Any weakening of the United States’ capacity to impose blockades and unilateral sanctions, organise coups, and wage wars can open political space for sovereign development and for class struggles to advance.

‘Multipolarity’ may function as a slogan of resistance to US primacy and offer a rough description of how certain forms of state power are distributed. Its political appeal, however, should not be confused with explanatory power. The number of powerful states tells us little about the social relations that produce their power, the class interests organised within them, or the imperialist hierarchies that connect them. Multipolarity does not, by itself, provide a Marxist account of the world; when treated as such, it can mislead us into believing that a wider distribution of state power is inherently progressive.¹

Venezuela’s former President Hugo Chávez, who had every reason to welcome a world of many centres, was dissatisfied enough with the term ‘multipolarity’ to coin ‘pluripolarity’ against it – insisting that what mattered was not the number of centres but whether popular subjects were the agents of transformation.² The need to specify this political content was itself revealing: the number of poles alone tells us nothing about the class forces or popular subjects shaping the resulting order.

A Marxist analysis must therefore replace the arithmetic of poles with the analysis of a chain: an unevenly developed system of social formations, each internally constituted by class struggle and externally bound to the others by relations of exploitation and dependence. The question is not how many poles there are but how the chain is held together and where it is most susceptible to being broken.

The problem, then, lies in the image of the pole itself. It asks us to picture the international system as a set of sovereign states treated as its decisive units, each possessing a measurable quantity of power. Depending on how that power is distributed, the system is described as unipolar, bipolar, or multipolar. This vocabulary tells us how many great powers count as geopolitical poles. It does not tell us what kind of social order each state reproduces, how its power is produced, which class interests shape that power, or through which transnational relations it is exercised. It photographs the chessboard while hiding the social relations that made the board, carved the pieces, and decided who is permitted to play.

Polarity presents itself as a form of measurement. Measured by gross domestic product (GDP) at purchasing power parity (PPP) – which adjusts for differences in price levels between countries – the world appears economically plural: no single economy commands anything resembling the dominance that the word ‘unipolarity’ implies. Measured by military expenditure and the alliance systems organised around it, however, that same world appears far more concentrated: in 2022, the United States alone accounted for about 53.5% of global military spending, while the US-led military bloc accounted for 74.3%. The United States also maintained at least 902 foreign military bases – by far the world’s largest network.³

Measured in a third way – by control over critical markets and infrastructures that function as chokepoints in contemporary production and exchange – the world appears neither plural nor bipolar but close to monopolistic. Yet this concentration does not map neatly onto states. Four firms hold 56% of the world’s commercial seed market and 61% of the global pesticide market. One of them is Bayer, which absorbed Monsanto and its patent portfolio in 2018.⁴ Three firms – Eli Lilly, Novo Nordisk, and Sanofi – control more than 90% of the global insulin market by value and around 95% of the market in low- and middle-income countries.⁵ Palantir, a single US firm, supplies the data platform on which the US Army runs its operations and US immigration authorities track human beings.⁶

The same concentration appears in finance and energy. The dollar accounts for more than four-fifths of global trade finance – a predominance that far exceeds the US share of global output or trade. This is a form of monetary command that a ranking of national economic weight cannot capture.⁷ In energy, ownership of a resource can likewise diverge from control over the circuits through which it reaches the market. This divergence is visible in neighbouring Venezuela and Guyana. Venezuela has owned its oil outright since 1976, yet US aggression – whether military (such as the 3 January bombing), economic (over 1,000 sanctions and other unilateral coercive measures), or otherwise – continues to shape the conditions under which Venezuelan oil can be sold internationally. In neighbouring Guyana, the 2016 production-sharing agreement governing the offshore Stabroek Block restricts the government’s ability to alter its terms without the prior written consent of the ExxonMobil-led consortium.⁸

Three measurements, three different worlds, one reality. The disagreement between those who see unipolarity and those who see multipolarity is therefore not simply a disagreement about the facts. It is also a disagreement about which relations constitute power – and it is precisely this question, which must be settled before any counting can begin, that the vocabulary of poles leaves unanswered.

Polarity also inherits what is often called a Westphalian representation of power: a world composed primarily of sovereign territorial states treated as internally coherent units. This obscures how territorial states, capitalist markets, chartered colonial companies, and cross-border circuits of capital accumulation developed together. The Dutch East India Company, founded in 1602, possessed extensive commercial and quasi-sovereign powers – including the authority to wage war, conclude treaties, and build forts – decades before the 1648 Peace of Westphalia.⁹ The history of the modern capitalist world system is therefore not one in which fully formed sovereign states appeared first and corporations only later crossed their borders. Capitalist power combined territorial, commercial, and imperial forms from the beginning.

The 1941 Atlantic Charter set out an Anglo-American vision for the post-war order, affirming the right of ‘all peoples to choose the form of government under which they will live’ while Britain still ruled a vast colonial empire.
Map created by MacDonald Gill in 1942. Sourced from the PJ Mode Collection of Persuasive Cartography, Cornell University Library via Wikimedia Commons.

The US Genealogy of Polarity

The modern vocabulary of polarity emerged from the institutional and political world of post-1945 US political science. It had older antecedents in European discussions of the balance of power, as we shall see, but it was in the United States, amid the behavioural revolution, systems theory, Cold War strategy, and the rapid expansion of university-based national-security research, that ‘polarity’ came to be treated as a formal property of the international system. The central figure in the consolidation of polarity as a theory of international structure was Kenneth Waltz, the principal architect of neorealism. Rather than beginning with the internal character of states, neorealism sought to explain their behaviour through the constraints of an anarchic international system and the distribution of capabilities among its leading states.

Waltz built on the work of Morton Kaplan, whose System and Process in International Politics helped translate the diplomatic history of balances of power and alliances into abstract, rule-governed models.¹⁰ Kaplan organised international politics into six models – the balance of power, loose bipolar, tight bipolar, universal, hierarchical, and unit-veto systems – giving a systems-theory scaffolding to what had previously been studied largely through diplomatic history. Both Kaplan and Waltz studied at Columbia University under the international relations scholar William T. R. Fox, who coined the term ‘superpower’ in 1944.¹¹ At a time when US political science increasingly sought to model political behaviour in formal, scientific terms, Kaplan and then Waltz drew on general systems theory to analyse international politics. The histories of imperialism, revolution, and class struggle were displaced by abstract interactions among formally comparable units. Class did not have to be explicitly refuted; it disappeared from the object of study itself.

Waltz gave this systems approach its canonical neorealist form. In ‘The Stability of a Bipolar World’ and more systematically in Theory of International Politics, Waltz argued that the international system was anarchic – meaning that no central authority stood above sovereign states. Within this system, states were treated as functionally similar despite differences in their political and social systems and distinguished principally by the distribution of capabilities among them.¹² The domestic orders of these states could differ profoundly, but Waltz’s structural theory abstracted from those differences, and his models transcended ideology. The United States and the Soviet Union could therefore be treated as functionally comparable units occupying two poles, and their conflict could be analysed primarily through the structure of bipolarity rather than through the antagonism between their social systems. Whether capitalism or socialism organised production in each country was treated as an internal difference rather than as something that constituted the international system itself. This abstraction was not innocent. It converted a historically specific Cold War conjuncture into a general theory of international politics. Within this framework, anticolonial revolution, competing social systems, unequal exchange, and US imperialism became analytically subordinate to the number and relative weight of the strongest states. The resulting framework closely reflected the concerns of a US national-security establishment focused above all on strategic stability and the management of rivalry.

During the Cold War, Waltz argued that bipolarity was more stable than multipolarity because the two superpowers depended less upon allies and could assess one another’s capabilities more directly. As the USSR disintegrated, this same conceptual vocabulary was repurposed to describe US primacy. In his 1990 Henry M. Jackson Memorial Lecture, Charles Krauthammer argued that the apparent plurality of world power concealed the predominance of a single superpower: the United States.¹³ The concept soon migrated from political commentary into realist scholarship. William Wohlforth’s ‘The Stability of a Unipolar World’ (1999) argued that the scale of US preponderance made counterbalancing difficult and major-power war less likely.¹⁴ In the same year, Samuel Huntington described the emerging order as ‘uni-multipolar’: one superpower operating alongside several major powers whose cooperation it often required.¹⁵ Although these authors disagreed over the number of poles and the distribution of capabilities among them, they shared the same underlying conception of the international system: states as comparable units whose relative capabilities determined its structure.

As US predominance has eroded in some areas while remaining entrenched in others, the language of polarity has returned to the centre of debate. The categories inherited from Kaplan and Waltz are now used to ask whether China, Russia, India, Brazil, and other states constitute independent poles. The debate commonly turns on indicators such as GDP, military expenditure, population, technological capacity, and diplomatic reach. Yet ‘multipolarity’ is often used less to identify a coherent structure of world power than to describe the relative decline of US power. What often remains unquestioned across these successive descriptions – bipolarity, the unipolar moment, and now multipolarity – is the assumption that territorial states are the basic containers of power and that world order can be understood by ranking them.

The limits of this approach become clearer when we try to interpret the war in Ukraine, the expansion of BRICS+, or the different positions and policies of China, Russia, India, and Brazil. These states cannot be treated as equivalent units merely because they resist, bargain with, or operate outside particular US initiatives. Their ruling blocs, forms of property, relations to international finance, popular pressures, and positions within the imperialist chain differ sharply. Those differences help explain why governments that cooperate within BRICS+ or another multilateral institution may diverge sharply on other questions and why the same state may resist US pressure in foreign policy while reproducing exploitation domestically and relations of dependency internationally.

Judged by its own criteria, however, the predictive record of this framework after the Cold War has been uneven at best. Structural realism was never merely descriptive: its proponents used it to make concrete predictions about how the post-Cold War order would develop. In 1993, Waltz argued that balance-of-power pressures would push Germany and Japan towards full great-power status, including the acquisition of nuclear weapons.¹⁶ Working from similar realist premises, Mearsheimer argued in 1990 that if US and Soviet forces withdrew from Europe and the Cold War alliance system dissolved, Europe would return to a multipolar order more prone to instability.¹⁷ Neither prediction was borne out. Germany and Japan did not acquire nuclear weapons. The United States remained militarily embedded in Europe. NATO expanded.

A Marxist explanation for this sequence of events points instead to relations that structural realism largely sets aside. In Germany and Japan, dominant class fractions were integrated into US-centred circuits of accumulation while their state apparatuses were reshaped within the wider US-led order.¹⁸ This structure did not fail to determine outcomes; the structure was never the one Waltz described.

John Mearsheimer’s work illustrates both the strength and the limits of this tradition. His rejection of the language of ‘democracy promotion’, used to justify NATO’s eastward expansion, can illuminate the strategic pressures surrounding the war in Ukraine more clearly than accounts framed primarily in terms of liberal values. However, by treating great powers principally as functionally similar security-seeking states, realism cannot adequately explain the class forces that shape Russian, Ukrainian, European, and US policy. Nor can it explain the material interests organised through energy, finance, arms production, and property relations or the unequal imperial structures within which strategic choices are made. A critique of liberal ideology is necessary, but without a social analysis of the states and class forces involved it remains incomplete.

Paradoxically, the language that now announces the end of US predominance remains rooted in a conceptual framework developed within the US national security academy during the Cold War. The social and class relations that helped shape the post-Cold War order are precisely those which this framework tends not to register.

This map depicts the two antagonistic blocs of the Cold War and the geopolitical logic of bipolarity.
Robert M. Chapin, Two Worlds 1950, 1950. Source: Time, 2 January 1950, via the PJ Mode Collection of Persuasive Cartography, Cornell University Library via Wikimedia Commons.

The Westphalian Model and the World Market

The theory of polarity rests on what is often called a Westphalian representation of the international order: a world composed of territorially bounded states, each sovereign within its borders and formally equal to the others, interacting in the absence of any higher authority. The term refers to the Peace of Westphalia, the agreements concluded at Münster and Osnabrück in 1648 that ended the Thirty Years’ War – a devastating conflict in Central Europe shaped by religious divisions, dynastic rivalry, and struggles over political authority within the Holy Roman Empire – and were later treated in international relations theory as the birth of the modern sovereign-state system. Yet the 1648 treaties of Münster and Osnabrück did not suddenly invent sovereignty, nor did they produce a clean division between domestic hierarchy and international anarchy. Rather, they settled the war within a layered dynastic and imperial order in which political authority remained overlapping, unequal, and socially heterogeneous. Historians and international relations scholars have therefore challenged the ‘Westphalian myth’ – the idea that the 1648 settlements created the modern system of sovereign states.¹⁹ The importance of Westphalia lies less in what the treaties immediately created than in what they later came to symbolise in international theory: the supposed birth of a world composed of territorially bounded sovereign states.

For Waltz, Westphalia served as a useful historical boundary. In Theory of International Politics he argued that from Westphalia onward no more than eight major states had at any one time competed for mastery over the world system. Waltz used this early European example to demonstrate the structural problem of multipolarity, largely setting aside the personalities of individual rulers and the details of diplomatic manoeuvring.

A second route to thinking about multipolarity emerged from Henry Kissinger’s study of the Congress of Vienna. His 1954 Harvard dissertation, later reworked and published in 1957 as A World Restored, examined how Castlereagh and Metternich used legitimacy, restraint, and equilibrium to stabilise the post-Napoleonic settlement.²⁰ Kissinger asked how diplomacy could manage a multipolar order. Waltz asked how the number and relative capabilities of great powers constrained diplomacy regardless of who governed them. Together, these approaches established the two dominant US conceptions of multipolarity: one centred on diplomatic management and the other on systemic structure.

Within both accounts of interstate order, economic relations remain secondary to relations among sovereign states. What recedes from view is the historical development of mercantile, industrial, and financial capital; the formation of national and international markets; and the legal and political institutions governing trade, property, labour, and investment. Waltz acknowledged that large corporations ‘and other nonstate actors are difficult to regulate and may rival some states in influence’.²¹ Even though the ‘importance’ of nonstate actors is ‘obvious’, he argued, it does not make the ‘state-centric conception of international politics’ obsolete. The issue, then, is not whether Waltz recognised non-state actors but whether they can be treated as secondary forces operating within a system whose basic structure is already defined by states. From a Marxist standpoint, they cannot: the capabilities attributed to states are themselves produced through relations of property, production, finance, empire, and class struggle, which cross territorial boundaries.

For Marxists, this omission matters because modern territorial states did not first develop as sealed political units and only later enter capitalist economic relations. The modern state system and the capitalist world market developed together. The consolidation of European sovereignty was bound up with enslavement, dispossession, and colonial rule abroad.²² The formal equality of sovereign states in international law has long coexisted with deep material hierarchy across the world system: colonies denied sovereignty, formally independent debtor states, occupied territories, protectorates, and states whose policy space is constrained by creditors, trade rules, sanctions, and military threats. Modern territorial sovereignty itself was produced over centuries through war-making, colonial conquest, capitalist property relations, and class struggle.

The pole metaphor converts these relations into attributes of individual states. Financial power appears as something the United States possesses rather than as a set of relationships reproduced through the international role of the dollar, central and commercial banks, debt contracts, payment systems, corporate accounting, and the constraints these arrangements place on other states. Contemporary production is organised through global value chains in which corporate control, manufacturing, logistics, data, intellectual property, and the location of profits are geographically dispersed. A corporation may own few factories yet control thousands of suppliers through standards, patents, purchasing contracts, and digital platforms. Workers across the Global South can be integrated into the same transnational production networks while remaining divided by borders, wage regimes, migration controls, and differential exposure to environmental harm. Lead firms and financial owners tend to capture disproportionate shares of value and profit while suppliers, workers, and states bear much of the risk.

A polarity framework has no convincing way to capture this kind of power. If a US-based technology firm organises production in East Asia, extracts data globally, holds intellectual property in one jurisdiction, books profits in another, depends on minerals from Africa and Latin America, and relies on US courts, diplomacy, export controls, and military protection, to which pole does its power belong? Calling it simply ‘American’ misses the transnational organisation of accumulation. Treating it as ‘stateless’ misses the state-backed legal and coercive systems that make its property claims enforceable. The Marxist point is therefore not that corporations have replaced states. Capital requires states to define and enforce property rights, discipline labour, absorb losses, build infrastructure, police borders, and defend access to markets and resources. Multinationals operate across borders yet remain dependent on territorial authority. They exploit differences among national tax, labour, and regulatory regimes while relying on powerful home states and bargaining with host states. The relation is neither a world of sovereign billiard balls nor a borderless empire of corporations. It is a differentiated unity of states and capitals – distinct forms of power that are nevertheless deeply interdependent.²³

This relational understanding of power has deep roots in Marxist theories of imperialism. Vladimir Lenin’s theory of imperialism already points beyond a simple map of great powers. Lenin defined imperialism through five connected features: the concentration of production and capital into monopolies; the merging of bank and industrial capital into finance capital; the growing importance of the export of capital; the formation of international monopolist capitalist associations; and the territorial division of the world among the major capitalist powers. For Lenin, the division of the world among great powers mattered not simply because several powerful states existed but because this division was bound up with capitalist accumulation and uneven development. A contemporary theory must extend this insight to global production, dollar-centred finance, digital monopolies, the expansion of extraction into new ecological frontiers, and the legal architecture of intellectual property and investment – not retreat to counting poles.

At the Second Congress of the Communist International in 1920, the debates over the Theses on the National and Colonial Question – including Lenin’s draft theses and the supplementary theses presented by the Indian revolutionary M. N. Roy – asked how communists should distinguish the oppressed from oppressor nations, analyse the class composition of anticolonial movements, relate workers and peasants to the national bourgeoisie, and combine temporary alliances with the political independence of the proletariat. In this framework, interstate relations, national liberation, and internal class struggle formed interconnected parts of a single strategic problem.

Mao Zedong’s On Protracted War (1938) offers a masterful example of this method in the context of China’s war of resistance against Japan. Mao’s analysis began from the concrete balance of forces: China’s social and political conditions and vast geography, Japan’s military and economic strengths and weaknesses, and the changing international situation. The Vietnamese revolution likewise defeated French colonialism and US imperialism by combining an analysis of domestic classes, the social and political contradictions inside the United States, and the possibilities created by relations and divisions among states with different social systems and political alignments. Cuba’s revolutionary diplomacy – from its defence of sovereignty to its internationalist activity – has repeatedly demonstrated the same capacity to work within interstate contradictions without reducing politics to relations among states.

This map was created in 1956 as an advertisement by a major US bank to show its expansion overseas, especially in Latin America, Africa, and Asia, highlighting international financial circuits and capital accumulation.
Sourced from an advertisement by the First National City Bank of New York in the article ‘How Bankers “on the Scene” Make the Difference Overseas’, the PJ Mode Collection of Persuasive Cartography, Cornell University Library via Wikimedia Commons.

States, Internationalised Capital, and the Imperialist Chain

Nicos Poulantzas developed Lenin’s account of imperialism in a direction that is particularly useful for understanding the relationship between national states and internationalised capital. Although Poulantzas did not produce a separate theory of geopolitics, his writings on the capitalist state, social classes, and the internationalisation of capital provide important elements for such a theory. In State, Power, Socialism (1978), he described the state as a ‘specific material condensation of a relationship of forces among classes and class fractions’.²⁴ For Poulantzas, the state does not stand outside of class struggle: the balance of forces among classes and competing sections within them takes material form in its institutions and policies. The state is neither a neutral actor nor simply the instrument of a unified capitalist class. Capital itself is divided into competing fractions – such as industrial, commercial, and financial capital – whose interests do not always coincide. The state consequently has a degree of ‘relative autonomy’ from any one fraction, allowing it to organise what Poulantzas called a ‘power bloc’: the alliance of dominant classes and class fractions whose competing interests are held together under the leadership of a hegemonic fraction. Foreign policy should therefore be understood neither as an automatic expression of capitalist interests nor as the pursuit of a timeless national interest. It is shaped by the changing balance of class forces expressed through the state.

In Classes in Contemporary Capitalism (1975), Poulantzas applied this approach to the internationalisation of capital. The expansion of multinational corporations, he argued, did not abolish national states or generate a single borderless capitalist state. International capital still had to be reproduced within concrete social formations and through national states, which guarantee property, regulate labour, provide infrastructure, manage crises, and organise political consent. The internationalisation of capital therefore takes place not outside or above national states but through changes within them. Foreign capital becomes embedded in domestic production, ruling-class alliances, administrative institutions, and state strategies. Imperial relations are thus reproduced not only through external pressure between states but also through the class relations and institutions within the states they connect.

Poulantzas developed the category of the ‘internal bourgeoisie’ to describe fractions of the domestic bourgeoisie whose accumulation remained nationally grounded but had become entangled with foreign capital. In post-war Europe, he argued, important sections of the bourgeoisie were neither autonomous national capitalists nor passive agents of the United States. Their capital remained rooted in their own social formations, but its reproduction had become intertwined with US corporations and international circuits of accumulation dominated by US capital. These bourgeoisies could bargain with US capital and sometimes conflict with it while remaining structurally dependent on the wider imperial order. Imperial hierarchy was consequently reproduced not only through pressure exerted by one state on another but also within the power blocs of subordinate states themselves.²⁵

This brings us back to the Leninist conception of the imperialist chain, now understood through Poulantzas’s theory of the state. Its links are not identical sovereign states distinguished only by how much power they possess. Each link is a social formation with its own combination of classes, institutions, political struggles, and forms of dependence. The chain is marked by uneven development: states occupy different positions within global production, finance, military power, and technological control. But these positions cannot be determined by ranking states according to national capabilities alone because the relations of imperialism also run through the internal class structure of every state.

A Poulantzian analysis would therefore ask questions that the theory of polarity does not. Which class fractions constitute a state’s power bloc? Which fraction exercises hegemony within it? How is foreign capital embedded in domestic production and state institutions? Which corporations depend upon which states to protect their ownership, markets, technologies, and supply chains? How do popular struggles alter the balance of forces within the state? Such questions do not make interstate conflict irrelevant – they explain its social content.

Poulantzas helps us move beyond the false choice between a world of sovereign billiard balls and a borderless empire of multinational corporations. Capital is internationalised, but political power remains unevenly organised through a plurality of national states. States are nationally constituted, but their internal class relations are also shaped by international capital and imperial hierarchy. Geopolitics is not an autonomous contest among territorial states but the international organisation of class power through a differentiated system of states. This also helps us understand the power of transnational corporations. Global supply chains and international finance allow large monopoly firms to operate across jurisdictions, but their existence and expansion depend decisively upon states. US digital and technology monopolies rely on public research, procurement, subsidies, courts, patents, export controls, diplomatic pressure, intelligence systems, and military power. Other monopoly capitals likewise depend upon the states in which their ownership and strategic command are anchored. At the imperialist stage, the integration of state and monopoly capital does not end when finance crosses borders; it is reproduced internationally through an unequal system in which capital operates across national boundaries while retaining privileged relationships with particular state apparatuses.

The erosion of particular forms of sovereign power under the pressure of international finance must therefore not be confused with the disappearance of the state. Even a subordinate capitalist state remains a central apparatus through which the dominant classes organise exploitation, secure property, discipline labour, manage social conflict, and reproduce their power. Integration into the imperialist chain changes the balance of forces within the state and may compel its ruling bloc to cede policy space, but it also makes the struggle against imperialist subordination part of the class struggle within that state. The relationship between class power and national sovereignty is consequently concrete rather than abstract. Sovereignty exercised by a capitalist ruling class is not popular sovereignty, yet breaking or loosening mechanisms of imperial command can enlarge the terrain on which workers and other popular forces struggle.

The 1980 Brandt Line, overlaid onto a Gall-Peters map projection, depicts the division between a wealthier industrialised North and a poorer, largely formerly colonised South.
Sourced from the cover of North-South: A Programme for Survival, the report of the Independent Commission on International Development Issues chaired by Willy Brandt, Pan Books, 1980.

A Contradictory World Order

Neither unipolarity nor multipolarity adequately captures the present world order. US predominance has eroded, but it has not simply disappeared. The United States remains the world’s largest military spender, sits at the centre of an unparalleled system of alliances and overseas military bases, and benefits from the dollar’s central place in reserves, trade, debt, and payments.²⁶ Its power is institutional and relational, not merely an aggregate of national resources. At the same time, China has become a major industrial and technological centre. Russia retains formidable military capabilities and vast natural resources. States across Asia, Africa, and Latin America have widened their room for manoeuvre. The expansion of BRICS+ reflects real demands for representation, development finance, and freedom from Western tutelage in the Global South.

Yet these tendencies do not add up to several equivalent poles. Military command, industrial capacity, finance, energy, technology, natural resources, and political legitimacy are distributed in different patterns. A state may be militarily protected by the United States, industrially tied to China, dependent on Gulf energy, exposed to private creditors, and responsive to a domestic popular movement – all at once. Interdependence does not abolish hierarchy, and rivalry does not end integration. US and Chinese capital may compete over technology while firms in both countries remain connected through production and markets. The semiconductor industry illustrates this layered geography: intellectual property, chip design, fabrication, lithography (the specialised technology used to print microscopic circuits onto chips), mineral extraction, assembly, and final markets are concentrated in different territories and controlled by different combinations of corporations and states. No single national ranking captures this structure, even though state power remains decisive through subsidies, export controls, sanctions, infrastructure, and military alliances. Sanctions and so-called ‘de-risking’ policies designed to reduce strategic dependence on rival economies disrupt or redirect some economic circuits while finance and commodity exchange bind others together. Meanwhile, the climate crisis, care work, migration, and social reproduction create material pressures that no great-power ledger measures.

These different distributions of power did not emerge independently of one another: they were reshaped through a connected transformation of global capitalism. From the 1970s onward, US-based capital increasingly relocated labour-intensive production abroad while retaining or strengthening control over functions such as design, intellectual property, standards, branding, finance, and the legal and coercive infrastructures that sustained them. The industrialisation of several US-aligned East and Southeast Asian economies – including the ‘Asian Tigers’ as well as Malaysia, Thailand, and others – and the financialisation of the North American and British economies were two aspects of the same process. What appears in one register as US decline – the loss of manufacturing employment – and in another as continuing US structural power – the dollar, legal and financial institutions, export controls, and military bases – can therefore be understood as related outcomes of the same restructuring. As strategic control over production became increasingly separated from the physical location of manufacturing, major conflicts began to emerge not only over territory but also at the points where this control is exercised, such as export controls on lithography and chip design, sanctions on payment systems, the extraterritorial application of domestic law, and struggles over technical standards and data. A ranking of great powers cannot readily capture these conflicts because it converts the relations that produce them into attributes of individual states.

A Marxist theory of geopolitics should begin not with the number of powerful states but with the historically developed totality of capitalist social relations. It would ask how surplus is produced, appropriated, and transferred; how ownership and control are organised across commodity chains; how finance disciplines states and households; how race, caste, gender, and citizenship segment the global working class; and how ecological costs are displaced onto dominated regions. It would examine the changing balance of forces within each state, locate that state within the unequal relations of the imperialist chain, trace how particular blocs of capital remain nationally anchored while accumulating transnationally, and examine how state institutions pursue the conditions for accumulation, political legitimacy, and security under different historical circumstances.

Applied to the present world order, this method refuses to treat China, Russia, India, or Brazil as socially or politically equivalent simply because they participate in BRICS+ or favour a less US-dominated order. Rather, it asks in each case how each state emerged historically, which classes and class fractions make up its power bloc, how different forms of property and planning shape accumulation, how labour and peasant struggles constrain it, how domestic capital relates to international finance, and whether its external actions reproduce or weaken imperial hierarchy. Only such an analysis can explain why states align in some areas and diverge in others rather than inferring their politics from their position on a geopolitical map.

Such a theory would also distinguish among political struggles and projects too often collapsed into the category of ‘great-power competition’. These include conflicts among fractions of capital, projects of sovereign development, socialist experiments, national liberation movements, and struggles against transnational corporations by workers, peasants, women, Indigenous peoples, and other impacted sectors. Their political content cannot be deduced from whether they oppose Washington, nor can every assertion of sovereignty be treated as emancipatory.

The relevant questions are concrete: does a project weaken mechanisms of imperial domination? Does it expand popular control over production and social wealth? Does it reduce dependency and hierarchy, or does it reproduce exploitation under a new flag? A state can increase its strategic autonomy without transforming its domestic relations of exploitation. Conversely, greater sovereign control over finance, resources, technology, or infrastructure can create conditions for popular forces to pursue more fundamental transformations. Neither outcome can be deduced solely from a state’s opposition to the United States.

This map, created by the US government, divides the world into the geographic ‘areas of responsibility’ assigned to US military commands.
Sourced from the United States National Imagery and Mapping Agency, The World with Commanders’ Areas of Responsibility, 1999, via the Library of Congress Geography and Map Division and Wikimedia Commons.

For this reason, struggles over national sovereignty cannot be dismissed as distractions from class struggle. In states subordinated within the imperialist system, control over finance, resources, food, technology, infrastructure, and foreign policy can be indispensable to any popular programme. A loosening of links in the imperialist chain may be valuable even when initiated by a capitalist state because it can weaken external constraints and alter the conditions of domestic struggle. However, its political meaning depends on which class forces lead it, who gains control over sovereign capacities that are recovered, and whether popular organisation can convert state autonomy into social transformation.

Multipolarity, then, has a limited but real political meaning. A less concentrated interstate order can make unilateral coercion more difficult and create openings for the Global South. Those openings can assist national liberation projects and advance class struggles within capitalist states. However, alignment with other states against US primacy does not alter a state’s class character or abolish its exploitative social order. There is no automatic path from several centres of state power to equality among peoples. A world of competing capitalist powers can remain profoundly imperial, exploitative, and ecologically catastrophic.

The task is not to choose the correct number of poles. The image of a balance treats power as a set of comparable weights whose distribution can be rearranged. A chain, by contrast, is composed of unequal and interlocking relations. It is transformed not simply by rearranging its links but by breaking the relations that bind them at the points where those relations are weakest. The decisive question is therefore not where a state is strongest but where imperial relations are most vulnerable to rupture.

For Marxists, the horizon cannot be a better balance among ruling states alone. It must be the international power of working peoples: popular sovereignty against imperialism, cooperation against zero-sum rivalry, social ownership against monopoly, ecological repair against endless accumulation, and internationalism rooted in popular struggle.

On 4 September 2026, the UN General Assembly voted 164 to 1 to promote equal-area projections (shown in red) that more accurately represent the relative size of continents, challenging distortions such as Mercator’s enlargement of northern landmasses (shown in blue). The only opposing vote was from the United States.
Mercator and Equal Earth projection comparison by Tricontinental: Institute for Social Research.

Notes

About Studies on the DDR

The German Democratic Republic (Deutsche Demokratische Republik, DDR) was a socialist state founded in East Germany in 1949 as a democratic, anti-fascist reaction to the Second World War and the subsequent restoration of monopoly capitalism in West Germany. The DDR represented a new Germany – one in which the land was redistributed, the means of production socialised, agricultural production organised into cooperatives, and significant advances made in gender equality, education, healthcare, and other areas of society. The DDR also cultivated friendly and close-knit economic relationships with other socialist states and exercised political and material solidarity with countries and movements fighting for their independence in Latin America, Asia, and Africa. In this series, we use the German acronym DDR instead of the English acronym GDR specifically because it has become a positive point of reference in many parts of the world and especially for countries in the Global South.

After the reunification of Germany in 1990, the DDR’s economy was dismantled and treated as a prototype for the shock therapy austerity measures that were soon imposed on other countries – not just the former socialist states. At the same time, the DDR was politically, judicially, and morally delegitimised. The publications in this series are a rejection of the narrative propagated by enemies of socialism, both new and old, that the demise of the DDR – and other socialist projects – proves the inevitable failure of any socialist policy and economy.

Through the series Studies on the DDR, the Internationale Forschungsstelle DDR (International Research Centre DDR), and Tricontinental: Institute for Social Research seek to stimulate a renewed engagement with the history, achievements, and principles of the DDR. After all, millions of people around the world are still fighting for advancements that were once a given in the socialist system of the DDR but were lost with its downfall.

1. Introduction

The state of contemporary global agriculture and food production urgently begs for alternatives. If the global food supply were distributed fairly and equally, everyone would have enough to eat. Yet, in 2025, 645 million people – roughly one out of thirteen – worldwide went hungry.¹ Prices are further inflated through financial speculation on food, raw materials, and land, perversely ensuring that profits are reaped not only from basic goods needed for survival but also from growing rates of hunger as rising commodity prices themselves become targets of speculative trading.²

While just over one-third of the world’s food continues to be produced by small farmers,³ massive agricultural corporations such as the ABCD Group – Archer Daniels Midland, Bunge, Cargill, and Louis Dreyfus – dominate global food cultivation, processing, marketing, and distribution.⁴ The result is that companies like those in the ABCD Group have a near monopoly over advanced agricultural technology and increasing control over globally centralised production and processing chains. This concentration of land ownership and the means of food production reinforces patterns established during the colonial era, driving monocultures, one-sided trade relations, and land evictions. These dynamics, in turn, exacerbate small-scale subsistence farming, extreme exploitation of agricultural workers, poverty, hunger, environmental destruction, and rural maldevelopment. While the development of capitalist relations in agricultural technologies and science has unleashed significant potential for food production, this has clearly not been used to satisfy people’s most basic need for food.

Decades after the dissolution of the DDR, we look back at its achievements – and what is to be learned from its shortcomings – for answers. In the DDR’s mere forty-five years of existence, the conditions in rural areas changed fundamentally. This change took place in three stages:

• The first stage (1945–1948) was during the shift in the balance of power in East Germany after the Second World War, which marked the decisive starting point for the DDR’s agricultural transformation. During this period, land reform was implemented that broke up the centuries-old land ownership structures in the countryside. This process of democratisation created the first framework for new relations of production based on cooperation.
• The second stage (1952–1960) was characterised by the Genossenschaftsbewegung (cooperative movement), which was able to resolve the contradictions between modern technology and small-scale production in the interests of farmers and agricultural workers without ruining, displacing, or subordinating them to the interests of big business, as occurs under capitalism. By the same token, agricultural cooperatives also provided the rural population with previously unattained social and cultural rights and opportunities.
• The third and final stage (1970s–1980s), built on the foundation established during the previous periods, saw the development of large agricultural production networks and deepening cooperative relationships between the various stages of production, from primary production to processing and distribution.

The DDR’s agricultural reform was among the most successful and comprehensive programmes in the country’s short history, transforming the reality of the peasantry from feudal-like working and living conditions (including economic and political subordination and cultural deprivation) into one marked by confidence, democracy, cultural vibrance, and highly productive industrial agriculture.

Every country and region has its own history and conditions that must be carefully analysed and for which unique solutions must be found. Nevertheless, the decisions made in the DDR’s agricultural sector offer practical insights for an alternative to capitalist market competition and profit-seeking. This study aims to contribute to contemporary debates on the prospects of socialist agriculture by examining the historical process that fundamentally changed the social structures and ways of working and thinking in the East German countryside. It is not possible to cover all aspects of the DDR’s agriculture in this study; important questions around forestry, ecology, agriscience, internationalism, and the inner workings of the Democratic Farmers’ Party (Demokratische Bauernpartei Deutschlands, DBD) could only be touched upon in passing and certainly deserve to be explored further in the future.

This study is structured chronologically, following the concrete historical development of agriculture in Germany: first the centuries-old feudal yoke imposed on Germany’s peasantry, then the period of land reform and, finally, the transition to a cooperative organisation of agriculture and large-scale industrial agricultural production.

2. The Prussian Path to Capitalist Agriculture and the Rise of Fascism

German agriculture transitioned from feudal to bourgeois-capitalist conditions during the nineteenth century, relatively late compared to other European countries like England and France.⁵ Even then, it only occurred very gradually and without a radical overthrow of the feudal aristocracy, who retained their hold over the land. The unification of Germany into a modern bourgeois nation-state was brought about by the political manoeuvring and warmongering of the Prussian ruling class, led by Otto von Bismarck and Kaiser Wilhelm I.

During this period – referred to in Marxist literature as the ‘Prussian path to capitalism’ – the emergence of the capitalist mode of production was tightly bound to relations of feudal subordination. The living and legal conditions of the various peasant working masses changed, but their economic and social dependency on the landed nobility remained. A class of ‘free’ small and medium peasants⁶ emerged, most of whom were heavily indebted to the landlords. Servants and sharecroppers who used to be landless farmhands for the lord of the manor now crystallised as a new class of agricultural workers. The feudal landlord himself became the owner of a sort of proto-agribusiness. Village life continued to be shaped by the large landowners, who maintained an economically and politically dominant role well into the twentieth century that was upheld by laws, customs, traditions, and new capitalist dependency relations. Backward, almost medieval-era conditions thus prevailed in the countryside.

Feudal laws, which had ostensibly been abolished in 1919 after the fall of the German monarchy, remained a de facto reality for agricultural workers and servants who were prohibited from organising politically or in trade unions such as in the German Agricultural Workers’ Association (Deutscher Landarbeiterverband, founded in 1909). Excerpts from the Prussian Servants’ Regulations, which was established in the early 1800s, reveal how labourers were still deprived of basic rights:

Paragraph 77: ‘If the servants provoke the lord’s anger through improper behaviour and are treated by him in return with insulting words or petty abuse, the servants may not seek legal redress’.

Paragraph 79: ‘Except in cases where the life and health of the servant is in present and unavoidable danger due to maltreatment by the lordship, he is not permitted to actively resist the lord’.

Paragraph 74: ‘Without prior knowledge and permission of the lordship, servants may not leave the house, even for their own affairs’.

Paragraph 76: ‘The servants are to accept the orders of the lordship and their reprimands with deference and modesty’.⁷

From the early nineteenth to the twentieth century, Germany’s economic base underwent a fundamental shift from an agro-industrial to an industrial economy. The population nearly tripled, from around 18 million in 1800 to over 52 million in 1900.⁸ While most Germans lived in the countryside in the early nineteenth century, from 1871 to 1910 – in the wake of the urban migration that accompanied capitalist development – the percentage of the population living in cities (defined at the time as having more than 2,000 residents) rose from 36% to 60%.⁹ Meanwhile, from 1882 to 1907, the share of workers engaged in agriculture and forestry dropped from 42.5% to 28.6%. A large class of landless agricultural workers soon outnumbered the owners and tenants of farms in rural areas. In 1907, large estates (greater than 100 hectares) employed an average of 51 labourers, and one-third of the 9.5 million people working in agriculture were wage labourers.¹⁰ Agricultural wage labourers were concentrated in the areas east of the Elbe River, where large agricultural enterprises were most widespread. Foreign seasonal workers, especially from the Polish regions, were particularly exploited. In the mid-nineteenth century, there was a significant concentration of land ownership: less than 1% of landowners held more than 40% of arable land, while around 80% of farmers owned less than 12%, consisting of individual plots under 7.5 hectares.¹¹

Conditions in the German countryside developed unevenly throughout the nineteenth century because of the country’s territorial fragmentation into many small states. West of the Rhine River, serfdom and indentured servitude had been swept away with the French Revolution of 1789. In central and southern Germany, between the Rhine and the Elbe Rivers, small and medium-scale farmers worked parcels of land with leases to large landowners. But when serfdom was abolished by law in the areas east of the Elbe (which would later make up much of the DDR’s territory), peasants were forced to compensate their former masters. Ultimately, to be emancipated from feudal obligations and duties, peasants paid a total of 1.8 billion marks and had to give 425,169 hectares of land to the Junkers, Prussia’s landed nobility.¹² This intensified the already greatly unequal distribution of arable land and gave rise to the large, landed estates that shaped the East German countryside prior to 1945.

Furthermore, the Junkers east of the Elbe were a central pillar of the reactionary, militaristic Prussian state and alongside influential industrial monopolists, would later become the facilitators of Hitler’s fascism. As historian Siegfried Kuntsche points out, the aristocratic Junker class used the key positions it retained in Germany’s military and state bureaucracy to openly support reactionary forces during the Weimar era (1919–1933), with ‘representatives of several long-established noble families join[ing] the Nazi movement’ as far back as the early 1920s.¹³ In the region of Mecklenburg, one of the heartlands of the Junkers, a quarter of all landowners became members of the Nazi party.¹⁴ Half of the generals in the high command of the fascist Wehrmacht (armed forces) came from nobility.¹⁵ During the Second World War, Junkers used their already dominant role in village social life to promote the theft of Eastern European land, pitting peasants against their counterparts in the East and diverting attention from the Junkers’ own history of appropriating peasant land.

Under Hitler (1933–1945), German fascism pursued a policy of the complete subjugation of peasants and the rural population to enable its imperialist war policy. Organisations and unions of agricultural labourers were banned. Everyone involved in the production, processing, and trade of agricultural produce was forced to become a member of the Nazi Reich Food Estate, or Reichsnährstand, under the command of large landowners. A total delivery obligation was enforced during the war, and peasants’ savings and credit balances were confiscated to pay for the cost of war. The fascist state compensated for the rapid decline in crop yields during the war years by imposing delivery quotas on produce from occupied countries in Eastern Europe and enforcing a merciless starvation policy against local populations. Only through the barbaric exploitation of millions of forced labourers were the fascists able to keep German agriculture afloat.

Nazi Germany’s wartime policies created a major setback for the country’s agricultural development. As early as the mid-1930s, the war economy was prioritised over agriculture. ‘If necessary, we will be able to cope without butter, but never without cannons’, as Nazi propaganda minister Joseph Goebbels put it in 1936.¹⁶ The damage incurred by the war left machinery and large tracts of land in ruin and livestock decimated. This was especially true in East Germany, where the fascists sought to ensure that nothing could be left behind for the Red Army and proceeded to systematically slaughter livestock and raze the soil – what Gerhard Grüneberg, the Socialist Unity Party of Germany (SED) agriculture secretary from 1960 to 1981, described as the continuation of the Nazis’ ‘criminal scorched-earth policy’.¹⁷

3. Land Reform: Junkers’ Lands in Peasants’ Hands

Post-War Destruction and Hunger

East Germany was the main European front for the final battles of the Second World War, leaving large areas of agricultural land, facilities, and processing plants in ruin; means of transport practically non-existent; and production capacities set back by decades. As the Red Army advanced on German territory, many Junkers fled to West Germany, often taking their livestock, technical equipment, and other private possessions with them and thus further worsening the supply situation. Swine livestock was reduced to 20% of its pre-war levels, cattle to 66%, and poultry to 25%.¹⁸ As a result of all these factors, the domestic food industry collapsed completely in 1945.

Image 1. Farmers remove debris and stones from a field in the East German village of Podelzig, which was almost destroyed in the final days of the Second World War, 1945.

After the war, Germany was divided into four military occupation zones. The Eastern Zone (known as the Soviet Occupation Zone, SOZ) was under the control of the Soviet Military Administration, and the Western Zones were controlled by the United States, the United Kingdom, and France. Berlin, which was located in the SOZ, was similarly divided among the four allied powers. In the SOZ, the German Administration for Agriculture and Forestry (Deutsche Verwaltung für Land- und Forstwirtschaft) was set up to deal with the urgent task of securing the first post-war harvest. On 18 June 1945, the Soviet Military Administration abolished the Nazis’ total delivery obligation by passing Order No. 40, which established delivery targets for grain, potatoes, vegetables, and oilseeds, followed shortly thereafter by Order No. 18, on 29 July 1945, which put an end to the compulsory delivery of livestock products. These were among the first measures implemented that promoted a democratic agricultural policy and helped spark the peasantry’s interest in increasing production.

Soon thereafter, a two-price system was established: farmers were paid a lower price for produce that they were required to sell to the state to ensure that there was sufficient food for the population and a higher price for surplus produce, which they could sell freely. Administrative bodies, trade unions, and newly permitted parties – which were united in the anti-fascist democratic bloc – called on the population to help with the harvest. Engineers and industrial workers formed crews to repair agricultural machinery as townspeople went into the fields to help with the harvest and peasant committees were formed in the villages to coordinate the work, combat hoarding, and ensure that the population received their food rations.

The Land Reform Programme

The Soviet Occupation Zone

During the war, German communists in exile in the Soviet Union developed a land reform programme for the post-war period. Edwin Hoernle was one of the main architects of this programme. Hoernle, an agricultural expert and communist parliamentarian during the Weimar era, returned from exile in May 1945 and took over as president of the Administration for Agriculture and Forestry (the main administrative position for agriculture and forestry in the newly created German Economic Commission). The meticulous planning and viability of this programme garnered support among the Communist Party of Germany (KPD) and its base, which in June 1945 issued a public appeal for democratic land reform under the slogan Junkerland in Bauernhand! (‘Junkers’ lands in peasants’ hands!’) and campaigned for a swift deliberation and implementation by the four-party bloc. On 22 August 1945, the KPD’s central committee adopted a concrete proposal to implement a land reform programme by the end of the year. Its central points were:

1. To expropriate, without compensation, large private estates (over 100 hectares in size).
2. To expropriate, without compensation, all property belonging to Nazi functionaries and war criminals, regardless of size.
3. To create a state land fund consisting of the expropriated property as well as state-owned assets.
4. To allocate the largest part of the land fund as private property in small, individually owned plots (five hectares in the case of high- and medium-quality soil and eight to ten hectares in the case of poor- and very poor-quality soil).
5. To distribute livestock and basic production equipment to ‘new farmers’, i.e. former agricultural workers, landless peasants, or refugees.
6. To transfer tractors and large agricultural machinery as well as workshops and agricultural processing plants to peasants’ mutual aid committees for collective use.¹⁹

Image 2. On 11 September 1945, three municipalities near Dresden, East Germany, expropriated and divided up the former Royal Manor. The sign (front left) reads ‘Free farmer on free soil’, and the banner (upper right) reads ‘The master’s time is up’.

The land reform initiative was promoted in a variety of settings, from the press to churches and party committees, and found increasing support among the population.

Hoernle and the KPD argued that compensating the large estate owners for their expropriated land would be irresponsible given the supply shortages and general economic destitution. As they pointed out, by no means would large estate owners be reduced to beggars if their estates were expropriated without compensation, as they still retained their private belongings and accumulated wealth. Furthermore, this was considered an appropriate punishment for war crimes and an act of historical justice.²⁰

Since the Christian Democratic Union of Germany and Liberal Democratic Party of Germany – both members of the anti-fascist democratic bloc alongside the Social Democratic Party of Germany and the Communist Party of Germany – opposed the proposal for expropriation without compensation, there was a series of lengthy negotiations before the legal ordinances calling for expropriation without compensation were ultimately passed by the five state governments within the Soviet Occupation Zone in early September 1945.

Image 3. A 1945 poster from the KPD’s central committee which reads: ‘Away with Junker rule: Junkers’ lands in peasants’ hands’.

The Western Zone

Popular pressure and the approval of land reform in East Germany in early September 1945 prompted the US and British military administrations in the West to draw up legal proposals for land reform in their zones. Yet the legal provisions they drafted offered politicians and large landowners enough leeway to ultimately prevent redistribution measures or to delay them until the balance of political power had shifted in their favour. The bourgeois parties forming in West Germany at this time sought to ensure that the old private capitalist ownership relations were preserved, despite popular referenda demanding the democratisation of the economy through the nationalisation of industry and other measures.

The West German region of Hessen is an example of this desire for change. Here, the social democrats and communists held a majority in the state government and drafted a state constitution that, in Article 41, outlined plans to socialise key industries and place the banking sector under public administration. Though a public referendum in December 1946 showed that 72% of voters were in favour of Article 41, the US military administration nevertheless opposed it, and US General Lucius Clay prohibited its implementation.

The land reform laws being drafted in West Germany were indicative of the rural population’s desire for change. Ultimately, however, the rural population was left disappointed, as the few measures enacted in the West applied to only 5% of the total arable land. In the DDR, meanwhile, 3.2 million hectares were expropriated and redistributed – fourteen times more than in West Germany. ²¹

The Political and Economic Background of Land Reform

Neither the Soviet Military Administration nor the German communists pursued the goal of socialist construction in East Germany in the immediate aftermath of the war. Rather, from 1945 to 1952, their main objective was to foster an anti-fascist, democratic transformation of a united Germany; eliminate German imperialism and militarism; and ensure that post-war Germany would become a neutral demilitarised zone in the international rivalry between the capitalist and socialist states.²² As such, land reform – i.e., the parcellation of land as private property – cannot be seen as a policy of the DDR’s socialist construction, which was first initiated in 1952, seven years after the war but as an effort to address pressing practical issues in the post-war era. Among them was the need to rebuild the collapsed food supply in East Germany by promoting small-scale agriculture, which motivated farmers to produce more and offered them the opportunity for self-sufficiency. Edwin Hoernle summarised the main reasons for land reform under these specific conditions:

The great agrarian reform is above all a decisive means for the complete eradication of Prussian militarism, whose main supporter was the Junker caste. Economically, the handing over of the land to the self-employed peasant is also urgently needed. In light of the current ruin of agriculture and the enormous lack of technical equipment, only proverbial peasant diligence, the fanaticism with which the smallholder farmer clings to his land, can bring about a rapid increase in agricultural production.²³

Nonetheless, land reform was a decisive step that later enabled the eventual transition to socialism. The democratisation of the countryside and the encouragement of cooperation among the peasantry laid the socioeconomic foundations for the eventual creation of socialist relations in agriculture. This policy was first and foremost the fulfilment of the peasants’ demand for the redistribution of land and an end to feudal oppression. Walter Ulbricht, the leader of the Socialist Unity Party of Germany (SED), which emerged from a merger between the Communist Party of Germany (KPD) and Social Democratic Party of Germany (SPD) in April 1946, summed up this connection at a party conference in 1947:

To be frank, some people asked us at the beginning of the land reform: What kind of socialists are you? You’re not even in favour of the ‘collectivisation of farmland’. We responded to them: That might suit you, gentlemen, keeping the large estates together so that they could later be easily returned to their former owners! No, our course is different: we are acting to ensure that the large estates are completely divided up, so that the large landowners in Germany can never again reclaim this land.²⁴

In the short term, the structures created by land reform were able to significantly improve the food supply, both for the individual farmers and the whole country. However, it was clear from the outset that the small-scale farming facilitated by land reform would not be able to meet the increasing demands of food production in the long term, as it faced challenges resulting from the size of the allotted agricultural land, the limited availability of equipment, and other factors.

Image 4. Members of a land reform commission in the DDR measure a parcel of land in September 1945.

Image 5. Young people help dismantle an old estate house to obtain building materials for new farmhouses.

The Implementation and Impact of Land Reform

In East German public village assemblies in September 1945, agricultural workers, landless and smallholder peasants, and refugees²⁵ elected members from their ranks to land reform commissions that were tasked with coordinating the implementation of the reform. The commissions determined which land was to be expropriated; documented and secured an inventory of associated buildings and stocks; drew up lists of smallholder and landless peasants, agricultural workers, and refugees; and assessed land allocation applications submitted by individual peasants. Redistribution plans were discussed and ratified in public meetings in which all applicants for land allocation participated. In order to avoid favouritism, decisions regarding the allocation of plots, stock, and inventory were made by drawing lots. The land reform commissions were also responsible for creating the first peasants’ mutual aid committees, established in the third quarter of 1945. These committees ensured that agricultural machinery was used efficiently and that mutual aid was organised among new farmers, from providing each other with advice on managing pests to lending helping hands during the harvest. Solidarity was to replace competition or obedience in the countryside.

More than 52,000 people²⁶ participated in the 9,500 land reform commissions, 38% of whom were agricultural workers, 35% of whom were smallholder and tenant farmers, and 12% of whom were resettled refugees. The remaining 15% were industrial workers and village craftsmen. The responsibility for implementing land reform was thus placed in the hands of the social forces that had the most vested interest in the reorganisation of land ownership. The executive boards of the commissions, consisting of five to seven members, were usually chaired by the district’s mayor. By the end of 1945, most of the land marked for individual reallocation had been redistributed, and the landed nobility was generally expelled from the estates and villages over which they had reigned.

Since the implementation of the land reform was overseen by the local population, the commissions came to act as schools of democracy. For instance, the handover of the title deeds for allocated land became a day of celebration in the villages. Local party member Hermann Wesemann’s account of a public meeting in Torisdorf, Schönberg county, paints a vivid picture of how land reform was implemented at the village level:

Torisdorf was an estate of around 400 hectares. It belonged to Junker Axel Bunger, a staunch militarist who let ‘his people’ address him as ‘Lord Captain’. There were some class-conscious workers in the village who helped us prepare a [monthly] village meeting. … Everyone was there, including the resettled refugees. Some agricultural workers nervously tried to give the impression that they had just come across the meeting by chance. They held buckets of water or milk cans in their hands and stood at the back. This was no doubt because of [their fear of] ‘Lord Captain’, who kept tabs on everything in the village. …Opinions went back and forth for a while until, finally, everyone agreed to divide up the estate. A land commission was formed and was to be headed by an agricultural worker. But now the landowner had to be officially informed of the meeting’s decision. When we went to see him, he was already shouting and swearing at us. I told him briefly about the decision and asked him to hand over the keys and all estate documents to the land commission immediately, to stay in his room until further notice and to refrain from any interference. He initially tried to intimidate us, declaring that the meeting did not have the authority and that he would not recognise oral decrees at all. In response to the question of authority, I asked him whether he thought this was a question of political power. For in this regard, he was too late: it had already been decided in favour of the working people. He and his peers had been outplayed. We kept him under constant watch until the estate was divided up, which happened soon afterwards, and prevented him from sending communications.²⁷

Torisdorf was among the 3.3 million hectares of land – around one-third of the agricultural and forestry land – and 14,000 buildings that were expropriated without compensation in the Soviet Occupation Zone as part of the land reform. Of the 3.3 million hectares of expropriated land, 76% consisted of 7,160 private estates (farms with more than 100 hectares of arable land); 4% of 4,537 farms owned by active Nazis and war criminals with less than 100 hectares of land; and the remaining 20% of state property and Nazi party property.²⁸

Two-thirds of this land (2.2 million hectares) was transferred to a central land fund, which was then distributed to individual applicants as identified and listed by the land reform commissions. The expropriated land was distributed to small farmers, agricultural workers, peasants, and refugees as privately owned plots and could be passed down within a family on the condition that the recipient family members use the land. It could not be sold, leased, or mortgaged. Forty-three percent of the expropriated land was distributed to agricultural workers and landless peasants, while 35% went to refugees.²⁹ As a result of the land reform, the share of agricultural land cultivated by smallholder farmers doubled.

Various measures were implemented to improve the quality of life of the new farmers and support nascent businesses. These measures included flexible delivery quotas, loans, and the provision of seed and livestock. Efforts were made to promote mutual cooperation and assistance during harvest times, as well as a special construction programme initiated by the SMA which led to the building of approximately 95,000 houses, 104,300 stables, and 38,470 barns by 1953.³⁰ Due to the shortage of labour, agricultural machinery, and draft animals, it was important and necessary to ensure cooperation between the peasants and the local authorities, organisations, and businesses. For example, the urban population was mobilised to help new farmers harvest their yields.

The remaining expropriated land (1.1 million hectares) became state property, resulting in the creation of 500 state-owned estates.³¹ Initially, these estates served as supply farms for the Red Army as part of the post-war reparations to the Soviet Union. Later, they were converted into Publicly Owned Estates (Volkseigene Güter, VEG), which became important mechanisms to support other agricultural enterprises, both publicly and privately owned, such as farms for seed and animal breeding, training and education, and experimental and research purposes. They also supported economically struggling farmers in an advisory capacity and played a pioneering role in the introduction of new farming methods.

Data from Klaus Schmidt, Landwirtschaft in der DDR [Agriculture in the DDR] (Clenze: AgriMedia, 2009), 16.
* Beginning in the early twentieth century, non-profit companies bought large estates and divided them into plots for lease to sharecroppers.
** Individuals and organisations associated with the Nazi party.

Data from Klaus Schmidt, Landwirtschaft in der DDR [Agriculture in the DDR] (Clenze: AgriMedia, 2009), 16.

This means, for example, that in 1939 only 1.1% of all farms had more than 100 hectares of land. However, these few farms accounted for 28.3% of the total agricultural land. Land reform not only greatly increased the number of small and medium-sized farmers but also enormously increased their share of the total utilised agricultural land.
Data from a Collective of authors, Früchte des Bündnisses – Werden und Wachsen der sozialistischen Landwirtschaft der DDR [Fruits of the Alliance – The Development and Growth of Socialist Agriculture in the DDR] (Berlin: Dietz Verlag, 1985), 33.

In early 1946, East Germans went to the polls to elect local commissions for the Peasants’ Mutual Aid Association (Vereinigung der gegenseitigen Bauernhilfe, VdgB), thereby turning the organically grown peasant mutual aid committees into official administrative bodies. The Peasants’ Mutual Aid Association was tasked with establishing machine-lending stations (Maschinen-Ausleih-Stationen, MASs), where peasants could rent agricultural machinery and equipment, and with setting up businesses that processed agricultural products and built repair workshops and cattle rearing centres.

By 1950, there were 32,621 state-owned and 617,886 privately owned agricultural farms of at least one hectare, the latter of which cultivated around 88% of all agricultural land in the DDR (5.7 million hectares).³² In 1948, 40,000 farms were run independently by women farmers, a result both of land reform and of the significant number of soldiers who died during the war.³³ With an average size of 9.3 hectares, privately owned farms were divided as follows:

Data from Klaus Schmidt, Landwirtschaft in der DDR [Agriculture in the DDR] (Clenze: AgriMedia, 2009), 131.

At the time, roughly two of the eight million people employed in the DDR worked in agriculture. A quarter were agricultural workers (workers with contracts); another quarter were farmers (self-employed owner or tenant of a farm), and half consisted virtually entirely of family workers (the family of the owner or tenant working without contract, social security, or income tax), which were almost exclusively women.³⁴ Of the approximately 500,000 farmers, 210,000 were new farmers.³⁵

In sum, the implementation of land reform in East Germany marked a historic turning point in the democratisation of agrarian relations and radically improved the living conditions of the rural population. With the economic and political power of the landed nobility finally broken, peasants and agricultural workers increasingly developed into self-confident and socially influential classes. They began to actively shape society through new democratic structures such as the Land Reform Commission, the Peasants’ Mutual Aid Association, and the Democratic Farmers’ Party of Germany³⁶ Cooperation and participation in decision-making processes became central features of peasants’ lives, bringing the question of power to the fore in the countryside. This was a new beginning, uncharted territory for a peasantry that had known only centuries of feudal oppression and then twelve years of fascist warmongering.

Image 6. A celebration of the 100th tractor built in the Brandenburg tractor factory in October 1949. The sign (right) reads: ‘Fighting for peace is fighting for Germany’s future’; the mural (left) reads: ‘Peace nourishes – war destroys’.

4. Building Cooperatives: ‘From Me to We’

In Germany, cooperatives and cooperative farmers’ associations existed prior to the DDR in the form of agricultural enterprises that offered bank loans and organised purchases and sales of agricultural products (seed, fertiliser, etc.). However, since Junkers and large-scale farmers held leading positions in the associations, they were able to use them to strengthen their dominance over small and medium-sized farms. By 1950, because of the land reform programme, these organisational units had been restructured and integrated into the Peasants’ Mutual Aid Association as the Farmers’ Trade Cooperative (Bäuerliche Handelsgenossenschaft), replacing a logic of subjugation with one of solidarity.

On the production side, the first Agricultural Production Cooperative (LPG) was founded in 1952. From the outset, LPGs were conceived not merely as institutions for the collective production of agricultural goods, but as broader social organisations tasked with economic, social, and cultural functions. The socialist character of cooperative production established itself through the LPG’s stated purpose, operations, and, above all, its integration within the DDR’s political and economic structures. Yet establishing this socialist character and continuing the fundamental transformation in the countryside proved to be a complex process.

For the DDR’s agricultural policies to meet the material needs of the population, especially in a context facing post-war ruin and an economy short of labour power, the question of scale and efficiency were key. Bruno Kiesler, head of the SED Central Committee’s Agriculture Department, explained how ‘democratic land reform was a strategically correct decision, though its initiators were aware that large-scale production was ultimately superior to a fragmented, small-scale economy in agriculture’. In order to ‘take the rural masses along this path’, he continued, rather than imposing the idea of large-scale production on them from above, the peasants ‘first had to find their own footing and gain trust in the political leadership. Through their daily working of the land, they gradually came to recognise that communal, cooperative work was more effective’.³⁷

The economic pressure to overcome small-scale agricultural production increased at the beginning of the 1950s with the heightened demand for agricultural products, both as raw materials for industry and as foodstuffs for the population as post-war scarcity gave way to peacetime demands. At the same time, there were signs of stagnation in agricultural production and labour productivity, and the difficult economic situation led to farm closures, especially among new farmers.³⁸ Furthermore, parcelled agricultural production on small private farms hindered the rational distribution of labour and made it difficult to efficiently use agricultural machinery, which was already scarce. Manual labour remained predominant. In 1950, the 514 machine-tractor stations (Maschinen-Traktoren-Station, MTSs) in the DDR had just 10,834 tractors and 675 trucks for the entire country.³⁹ By 1952, before the cooperative movement began, some 30,000 new farmers had abandoned their farms.⁴⁰ While the MAS – the precursor of the MTS – serviced individual farmers and offered small equipment, the MTS sought both to extend beyond individual equipment rentals, providing large agricultural machinery and organising its deployment in the most efficient way across more and larger LPGs.

Image 7. Courtyard of an MTS in Obhausen, Saxony-Anhalt, DDR (1950).

The socialist historian Kurt Gossweiler described the increasing necessity to make a strategic decision about the political system under which to develop the agricultural sector:

Faced with the urgent need to increase the agricultural yield, the DDR’s leadership had to decide how to open the way to large-scale agriculture: either along capitalist or socialist lines. For this reason, the discussion about the construction of socialism in the DDR was primarily a discussion about whether the formation of agricultural production cooperatives would be approved.⁴¹

A policy oriented toward cooperative agricultural production would have a profound impact on labour and property relations in rural life. It was a far-reaching decision that could not be resolved until the national question in Germany had been settled. In 1949, when the DDR was founded as an anti-fascist, democratic state in reaction to the founding of the Federal Republic of Germany (Bundesrepublik Deutschland, BRD), the Soviet Union, as well as the communists in West and East Germany, continued to pursue the strategic goal of a unified, non-aligned German republic in which the Four Ds of the 1945 Potsdam Agreement would be implemented.⁴²

Ultimately, the SED and the political leadership of the Soviet Union recognised the necessity to shift strategies. At its 1952 party congress, the SED decided to begin the construction of socialism through the formation of LPGs. Walter Ulbricht, general secretary of the SED, stressed the need for farmers to join cooperatives on a voluntary basis, emphasising ‘the principle of absolute voluntariness in the organisation of such cooperatives and to point out the inadmissibility of applying any kind of coercion towards the peasants in this regard’. Furthermore, Ulbricht continued, ‘Comrade Lenin himself had stressed that communal agriculture cannot be introduced by decrees and laws’.⁴³

Image 8. An SED agitation poster from the 1950s that reads: ‘The future belongs to socialism. Individual farmers, join the LPG’.

Building Socialism One LPG at a Time

Though the need to develop cooperatives and the structures for large-scale agricultural production was clear, in many respects the subjective and objective conditions to carry out such a project had not yet developed. How would the DDR bring about a shift in peasants’ way of thinking, working, and living?

In some cases, it was peasants and agricultural workers who pushed for a form of collective production. Such peasants had in fact urged the SED to support the founding of cooperatives even before the party’s decision to do so in July 1952, and this approach is what led to the creation of the first LPG in Thuringia in the same year. They were in favour of LPGs because they either saw them as a necessary response to the difficult economic situation on their own land or because they knew of cooperative agriculture in the Soviet Union and were convinced of the need for socialism.

Image 9. In January 1953, nine farmers founded an agricultural production cooperative in the municipality of Blochwitz (Grossenhain district). A farmer named Godzina signs a document to register herself as a member of the cooperative.

Yet, for the vast majority of peasants – especially among long-established medium and large-scale farmers – the thought of giving up individual farming was still inconceivable in 1952. Many were dismissive, while others took a wait-and-see approach. An unprecedented reorientation was necessary to promote a more collective way of thinking and working, which came to be known by the slogan Vom Ich, zum Wir (‘From Me to We’). SED agriculture secretary Gerhard Grüneberg summarised some of the major difficulties in this process:

Cooperative work makes it necessary to become a member of a collective. Previously, the peasant alone decided how to organise his work. Now he⁴⁴ receives instructions or, if he becomes head of a brigade, for example, he must organise the work for many people… [and] learn to overcome certain prejudices and respect other cooperative members as his equals.⁴⁵

In the initial phase, new and smallholder farmers who managed economically weak farms of all sizes were the main driving force behind the cooperative movement.⁴⁶ Large-scale farmers were excluded from LPG membership until 1955 in order to ensure that those peasants with closer links to the working class played the decisive role in the cooperatives from the outset.⁴⁷ Generally, smallholder farmers, agricultural workers, and new farmers were easily recruited to join LPGs because they faced the greatest economic difficulties and were less attached to their land.

Image 10. On 19 May 1954, the Freie Erde (Free Soil) LPG in Magdeburg-Salbke builds a large farmyard. The cooperative’s farmers laid the foundation stone together with workers from neighbouring industrial enterprises. The sign reads: ‘How we work today is how we will live tomorrow!’

Image 11. The LPG construction brigade builds a feed house on 19 October 1955. Between April and October of that year, the construction brigade completed three new stables. By 1960, the brigade had built farrowing pens, and combines for pigs, cattle, and poultry. All these major construction projects were completed with the help of volunteer construction workers.

The nascent LPGs faced unfavourable conditions of production in the early years, such as degraded soil, no experience in organising large-scale collective production, and a shortage of labour and lack of available barns, stables, and other farm buildings. The process was made even more difficult by the fact that farmers often joined the cooperatives one by one. Integrating new members and new areas of production were a constant challenge. Crop rotations and the organisation of production had to be adjusted frequently. On average, it took five to ten years to establish solid structures and continuous operating procedures in LPGs.⁴⁸

Despite these and other challenges, by 1952 the first cooperatives were already farming an average of 114 hectares of agricultural land.⁴⁹ This meant that the material prerequisites for large-scale production – such as the new agricultural machinery and technology and funds needed to construct farm buildings – had to be secured as quickly as possible. Even though the manufacturing of agricultural machinery accelerated in 1952, there was still a lack of sufficient technical equipment and training for cooperative farmers. All of this put further strains on the national economy.

In order to facilitate a gradual and accessible transition to cooperatives, the state introduced active development incentives and subsidised newly established LPGs through measures such as the reduction of tax arrears and outstanding debts, the provision of free technical expertise, the priority allocation of quality seeds and mineral fertilisers, and the provision of loans on favourable terms.⁵⁰ For the same reason, three different types of LPGs were created, which introduced a gradual approach away from the private property model of the previous system to the collective approach of LPGs. From type I to III, each stage introduced a different extent of socialising the means of production, requiring participation in communal work, and regulating income distribution:

• In Type I LPGs, arable land, machinery, equipment, and draught animals became collective property while productive and breeding livestock remained the property of the farmers.
• In Type II LPGs, tractors, draught animals, machinery, and equipment were also transferred to the LPG.
• Type III LPGs operated as fully collective and independent enterprises. Here, except for private household farming, the farmers’ entire homestead (arable land, grassland, woodland, livestock, tractors, machinery, equipment, farm buildings) was transferred to the cooperative.

In all three types, the land that peasants brought into the LPG remained their own property rather than state property and was taken into account when wages were distributed.⁵¹ The higher the tier, the less the land contributed by the farmers was taken into account for remuneration.

In 1949, the machine-lending stations, which had been organised by the Peasants’ Mutual Aid Association to share existing machines between farmers, were transformed into state-owned machine-tractor stations (MTSs). This change reflected the growing supply of tractors and agricultural machinery. The MTSs played a leading role in forming and stabilising cooperatives: while larger technical equipment, trucks, and machine systems in the MTSs could be borrowed according to a tariff system that favoured small over large-scale farmers and cooperatives over individual farms. Furthermore, since the technical base of the MTSs and their advisory service on crop production and soil conditions were continually expanded by the state, they effectively determined the direction of development towards large-scale industrial-style agricultural production.

In addition to meeting the economic needs of the peasantry, the MTSs played a central role in the political and cultural development in the villages. Cultural annex buildings were attached to the MTSs, where further training for farmers, agrochemical and zootechnical consultations, film screenings, theatre trips, music and dance groups, libraries and sports clubs, and much more were organised by the MTSs. These cultural houses became an essential hub for agricultural development, social and cultural interaction, and democratisation of society. Furthermore, MTSs served as support hubs and cooperative recruitment bases for the working class in rural areas, connecting mechanics, tractor drivers, and other workers with peasants who were involved in cooperatives.

Image 12. Villagers browsing through the library in the community of Krien, in Mecklenburg-Vorpommern, DDR.

Integrating Cooperatives into the Planned Economy

Though the shift toward cooperative agriculture was initiated with the founding of the first LPG in 1952, the first LPG law was only passed in 1959. Incorporating the experiences of the past seven years and the aspirations of those yet to come, this law and later versions included a comprehensive definition of the social role and task of cooperatives. The 1982 version of the law stated the following:

§1 Principles
(1) Agricultural production cooperatives (hereinafter referred to as LPGs) are voluntary associations of farmers, gardeners, and other citizens that promote joint socialist production to meet the material, cultural, and food supply needs of the population, as well as provide industries with raw materials. LPGs ensure a significant increase in output and high efficiency in agricultural production by constantly improving the ratio of input to output.
(2) In LPGs, cooperative farmers organise their joint work and social relations according to the principles of cooperative democracy and socialist business management, such as equality, comradely cooperation, and mutual aid. Farmers jointly decide over cooperative property and share in the economic gains of their LPG according to cooperative principles of distribution.⁵²

Six years into the transition to socialist agriculture, the SED’s Fifth Party Congress in 1958 set the goal of achieving self-sufficiency in agricultural production as well as adequate access to raw materials for the agricultural industry by 1962. Apart from products that could not be grown in the DDR due to climate and environmental conditions (e.g. tropical fruits, coffee, cocoa, etc.), food sovereignty was thereafter largely achieved.⁵³

In order to plan and develop the national economy, cooperative agricultural production had to be integrated into the DDR’s centrally planned economic system. The primary instrument for planning agricultural production was the Plan of State Production Volume of Agricultural Products, which was worked out through a complex process of assessments and discussions. First, the demand for agricultural goods and products was determined at a central level by the State Planning Commission, the Ministry of Agriculture and Forestry, and the Central State Office for the Registration and Purchase of Agricultural Products. This included priorities for the development of production as well as cultural and social progress. The objectives laid out in this draft plan were then discussed at the district and municipal levels as well as in the individual LPGs and the other socialist agricultural enterprises before being presented again at the central level and then adopted by the People’s Chamber of the DDR. Only then, based on the annual plan, did individual LPGs draw up their final operating plans, working closely with the municipal councils. The plan’s implementation was then monitored by the municipality’s council or mayor.

Each task in the LPGs was assigned a value on a scale of 0.8 to 1.6 work units, a conversion key that served to classify these according to their importance, difficulty, and the qualifications required to complete them. Complicated work was remunerated with more units than simpler work, with members completing a range of tasks in a given period, each assigned a specific work-unit value. A work unit corresponded to a monetary value which was calculated from the total sum of the generated funds available for distribution to the members and the number of work units performed in the LPGs. A minimum value of seven marks (the official currency of the DDR) per work unit was guaranteed by the state. Since the exact value of earned work units could only be determined at the end of the year, a system was used in which approximately 70% of the planned value of work units was paid in advance.⁵⁴ Solveig Leo, a long-standing chairwoman of an LPG, recalled the process of assigning value to work units and the debates that arose:

In the evening, they [the LPG farmers] usually assessed the day’s performance and recorded the work units on a blackboard in the LPG: 1.1 or 1.2 work units or so per day for each name. This meant that every cooperative farmer was able to see how many work units his colleague had been awarded. And, of course, there were debates. For instance, if someone disagreed and thought they had worked much more diligently than what the calculation showed, they asked: ‘Why do I have only 1.1 and he has 1.3?’. Then the chief brigadier had to explain why.⁵⁵

Image 13. LPG bookkeeping was a difficult and often unpopular but necessary task. An agronomist from the Peasants’ Mutual Aid Association (left) trains a farmer (right) who took on this work for an LPG in Neubredersdorf, DDR, while some of the other members check the calculations (October 1954).

When an LPG laid out its production plan, it allocated it to various funds in coordination with the municipal council. These funds were divided into a consumption reserve and an investment reserve with various distinctions, such as the essential fund (which went towards farm buildings, machinery, equipment with a long service life, membership initiation fees from cooperative members, etc.); working capital; investment; remuneration; cultural; and social expenditure.

As was the case in other sectors in the DDR, in order to promote production, competitions were organised within and among LPGs alongside annual plans and everyday brigade work. The aim was to mobilise additional resources in labour time, draw public attention to the importance of agriculture, and make villagers aware of the collective results of all the work in the agricultural sector, including the work of the LPGs. Author Thomas Kupfermann describes how much public attention these competitions attracted:

The mechanised harvesting of crops on a large scale… was known as the ‘battle of the harvest’. The hard-working men – and women – who drove their combine harvesters across the large LPG fields and took their machines to the next job on the roads… were the ‘harvest captains’. Every year, the harvest battle raged, and the media accompanied it with headlines and success stories… The cooperatives participated in district competitions, which were also reported on, and at the end of the harvest battle there were awards for the winners and their pictures were featured in the ‘Streets of the Best’ [blackboards where the best workers were displayed].⁵⁶

Image 14. Colleagues from the Institute for Agroeconomics at the German Academy of Agricultural Sciences in Berlin support the senior accountant of the Schönemark LPG, Anne-Margret Mielke, as she prepares the monthly cash audit.

Image 15. Thirty-two female farmers from the Vorwärts LPG in Groß Beuchow, DDR, harvested 100 tonnes of potatoes, which were immediately taken to nearby shops (September 1961).

Image 16. 1961, the Friendship Milking Brigade of an LPG in Neßdorf produced more than 80,000 kilograms of milk, exceeding the target production plan for that year.

Image 17. Members of the Priborn LPG discuss high crop yields and the efficient use of technology at their annual general assembly (January 1973).

Class Struggle and Policy Disputes

The progress of the cooperative movement in the DDR was shaped both by internal conditions and by the political conditions across Europe in the 1950s. In the DDR and Eastern Europe, the development of socialism was still in its infancy. At the same time, the Cold War waged by the US and its Western allies against the socialist camp intensified. There were many attempts to exploit uncertainties and contradictions in socialist countries and their political leadership, especially following the death of Joseph Stalin in March 1953. The violent political crises and anti-government protests in the DDR in 1953, Poland in 1956, and Hungary that same year can be understood in this context.

As late as March 1952, Heinrich von Brentano, then chairman of the CDU/CSU parliamentary group in West Germany’s Bundestag, openly stated his intentions to regain control of East Germany: ‘We will do anything and everything – and I say explicitly: anything and everything – to get the Soviet Occupation Zone back’.⁵⁷ In particular, the open border between West and East Berlin offered the West a wide range of opportunities for disruptive actions, including espionage, sabotage, and establishing terrorist organisations. For instance, the West Berlin radio station, Rundfunk im amerikanischen Sektor (Broadcasting in the American Sector), used manipulative, negative reporting to increase farmers’ doubts about the prospects of the socialist perspective.⁵⁸

The West exploited existing challenges and discontent in the DDR, which mainly stemmed from the population’s struggle to overcome the consequences of the war, and weaponised them against its government. Seeing how tense the political situation was becoming at the beginning of June 1953, the DDR’s leadership implemented comprehensive relief measures for the population. Nevertheless, protests broke out weeks later on 17 June 1953, which were incited into a counter-revolutionary coup attempt with the help of the West.

In the aftermath of the unrest, the SED leadership reflected critically and publicly on its policy mistakes, including in regard to the cooperative movement that had begun one year earlier. In a newspaper article, the SED Central Committee wrote:

The party, which had taken the right course towards building the foundations of socialism in the DDR, took the wrong path of accelerating the solution of this task without taking into account the real internal and external conditions. This… had detrimental consequences for the population’s food supply and to a certain extent led to the disruption of proper relations between the party and the working masses. Administrative methods were employed where broad and diligent educational work amongst the masses should have taken place.
It was right for the party organisations to support the movement to form agricultural production cooperatives that emerged in the DDR following the initiative of working farmers. However, in some circles, party organisations tolerated the violation of the strict principle of voluntariness and tried to increase the number of cooperatives without devoting the necessary attention to the organisational and economic consolidation of the existing cooperatives, which is the main task of the party when advancing the development of cooperatives in the villages.⁵⁹

Eventually, the effort required to establish new LPGs made further expansion difficult, leading the state to shift toward consolidating existing LPGs, supporting their efforts to raise production capacity, and integrating additional farmers, while also providing greater support to private farms whose productivity was still essential to meeting overall agricultural targets.

During this period, there was a renewed debate about the future of agriculture in the DDR, which largely revolved around the extent to which the state should continue to promote cooperatives. In the mid-1950s, most cooperatives were still less productive than medium and privately owned large-scale farms and had to be supported by state subsidies. Some politicians and analysts used this as an opportunity to campaign against state support for ‘unproductive’ cooperatives and for strengthening privately owned farms. These proposals were sharply rejected by the SED leadership under Walter Ulbricht, who declared in 1957 at a conference of the Central Committee:

It is being said that the orderly development of agriculture and the systematic promotion of socialist agriculture should be abandoned… Priority, they say, should be given to the development of [private] family farms… This proposal paves the path back to a capitalist economy, whereby the large farmers would become the main economic force in the village.⁶⁰

These disputes reveal that, throughout the entire period of the cooperative movement, there were deep conflicts and debates about how to balance the relationship between voluntariness and the active and determined promotion of cooperatives by the state. Administrative exaggerations by government agencies and political leaders that violated the principle of voluntariness were criticised and opposed. Yet so too were ideas of a self-sustaining process that assigned the state a passive role towards the farmers. There was also an active struggle against the forces that set out to sabotage the development of socialist agriculture.

Socialist Springtime: The Height of the Cooperative Movement

By 1958, six years after the first LPG was founded, there had been a gradual consolidation and stabilisation of productive capacity within the cooperative movement, though state support remained necessary in most cases. Yet, as long as private farms existed alongside the cooperative sector, the socialist transformation of the countryside remained incomplete. The coexistence of different socioeconomic structures in agriculture – cooperative production and individual farming – was bound to hold back the growth of productivity. At the same time, the mere existence of private enterprises in the countryside gave political opponents ammunition to argue against the project of socialist agriculture as a whole. Pressure increased to bring private enterprises into the cooperative movement and create a unified system.

From 1958 to 1960, commonly referred to as ‘socialist springtime’, almost all remaining private farmers were brought into LPGs through initiatives led by bloc parties,⁶¹ mass organisations, the Democratic Farmers’ Party of Germany, the Peasants’ Mutual Aid Association, and MTSs. This consolidation was reinforced by economic measures that strengthened the sector’s technical capacity, such a marked increase in the agricultural machinery available. As the LPGs’ productivity visibly rose and state support became more effective, medium- and large-scale farmers who had long held reservations began to

By the time the campaign was declared complete in the spring of 1960, its scale was visible in the statistics: the predominantly small, privately owned farms that had characterised agriculture in 1950 had given way to over 19,000 LPGs, each cultivating an average of around 280 hectares. Eighty-four percent of all agricultural land was now farmed cooperatively.⁶²

Image 18. Industrial workers help with the potato harvest. Cooperatives were assigned industrial enterprises at an early stage to provide support and workers helped with the harvest and repairs.

Women farmers played a key role in the expansion and consolidation of the cooperative movement, as they were often the key actors winning their families over. Women, who had traditionally worked in the time-consuming and physically taxing field of animal husbandry while also managing their households and childcare, certainly stood to gain the most from joining LPGs and the social benefits granted to their members.

Though the vast majority of farmers had agreed to join LPGs by the spring of 1960, in some cases, they only came to truly appreciate the benefits of cooperative production in the years that followed. One farmer, Elli I., reflected on her scepticism towards the cooperative in the early years:

As individual farmers, we had to work from dawn to dusk. In 1957, my daughter was born prematurely because I had to help my husband load thirty sacks of fertiliser. In the cooperative, we no longer had to worry about such things. We had joined the cooperative quite late and, after years, I thought to myself, my goodness, how stupid had we been? You can’t live any better than here. The work got easier. We had medical care, every woman could go to the doctor, we had housekeeping days, and vacations. When I think back, … those were my best years.⁶³

The socialist agricultural sector includes publicly owned estates and cooperatives. Other small state-owned agricultural enterprises are not shown separately, which accounts for the difference between the total share of socialist farms and the combined shares of publicly owned estates and cooperatives.
Data from State Central Administration for Statistics, Statistisches Jahrbuch der DDR 1960/61 [Statistical Yearbook of the DDR 1960/61], VEB Deutscher Zentralverlag, 1962, 419.

Data from Gerhard Seidel, Die Landwirtschaft der DDR [Agriculture in the DDR] (Berlin: VEB Deutscher Landwirtschaftsverlag, 1961), 37; Collective of authors under the direction of Volker Klemm, Von den bürgerlichen Agrarreformen zur sozialistischen Landwirtschaft in der DDR [From Bourgeois Agrarian Reforms to Socialist Agriculture in the DDR] (Berlin: VEB Deutscher Landwirtschaftsverlag, 1985), 220; State Central Administration for Statistics, Statistisches Jahrbuch der DDR 1989 [Statistical Yearbook of the DDR 1989] (Berlin: VEB Deutscher Zentralverlag, 1990), 196.

Data from State Central Administration for Statistics, Statistisches Jahrbuch der DDR 1971 [Statistical Yearbook of the DDR 1971] (Berlin: VEB Deutscher Zentralverlag, 1971), 222; State Central Administration for Statistics, Statistisches Jahrbuch der DDR 1955 [Statistical Yearbook of the DDR 1955] (Berlin: VEB Deutscher Zentralverlag, 1956), 224.

Cooperatives Drive Social and Cultural Progress

From their origin, the tasks of LPGs were not limited to agricultural production but to the improvement of rural life in general. Over time, they developed into hubs of social, communal, cultural-political, and economic development in rural communities. In coordination with the municipalities and urban enterprises, LPGs built apartments, cultural centres, kindergartens, and sports facilities. They improved roads and constructed new ones, installed water pipes and sewage systems, dug drainage ditches, and opened stores and restaurants. LPGs built on and expanded the cultural work of MTSs. Performances by theatre groups, dance nights for the youth, sporting events, and film screenings brought a new level of cultural life to the villages.

Hans Luft, an agronomist and former employee of the DDR’s Ministry of Agriculture, reflects on the cultural transformation brought about by the new production relations of the LPG system:

The joint work in the cooperative and [resolving] the emerging problems such as the use of complex machinery to bring in the harvest as quickly as possible, difficulties with repairs and procuring spare parts for agricultural machinery, completion of construction and irrigation systems, [and] epidemic-free management of large livestock quickly broke down generations of enmity in the villages. The farmers’ children grew into the new forms of production with new knowledge, experience, and, above all, higher levels of qualification.⁶⁴

Cooperatives vastly improved the rural population’s living conditions, reducing the physical strain on farmers through communal work and the use of machinery.⁶⁵ In addition to national programmes that guaranteed access to national social, health, and insurance systems and so on, the LPGs achieved a number of important social advances for their members, which were impossible for previous generations of farmers to even dream of and still unavailable to the few remaining private farms, such as:

• Equal pay for equal work
• Special remuneration for work on Sundays and public holidays
• Sick pay
• Child benefits, such as monthly cash payments and birth grants
• Paid vacation, with the option of spending it in a subsidised LPG-owned vacation site
• Fixed and predictable work hours, except during the harvest months
• The inclusion of all cooperative farmers in a uniform social insurance system that covered benefits such as healthcare, pensions, and accident insurance
• ‘Exemption’ from income tax contributions⁶⁶

In addition, child labour, which had been a common practice on family-run farms, disappeared with the advent of LPGs. Central schools replaced the single-class schools in the villages. Children and young people had access to the DDR’s polytechnic education system, which enabled them to choose an independent career and provided a foundation for agricultural studies. Comprehensive education and training opportunities were created for cooperative farmers, from which female farmers benefited significantly as they were able to gain professional qualifications and take on managerial roles in the cooperative farms for the first time.

Margarete Müller, an agronomist and SED politician, described the progress that her village achieved:

I have lived in Kotelow since 1960. Back then, at the age of 29, I was elected LPG chairwoman. When we started, there were no paved roads and our boots got stuck deep in the mud. Bit by bit, we modernised the village and improved it. We were happy about every new house, every new stable, every new barn, about the dairy farm and the pigsties that we built together. We celebrated together and also got frustrated together when things didn’t go the way we wanted them to. We were a community. Even in the summer, farmers could go on holiday, and youth who qualified at universities and technical colleges sometimes moved back because they saw a future for themselves and their families in little Kotelow.⁶⁷

From 1960, the height of the cooperative movement, until the dissolution of the DDR in 1990, women accounted for roughly 40–45% of LPG members.⁶⁸ While only 8.5% of women working in agriculture had a degree or vocational training in 1963, compared to 18.1% of men, by 1984 this figure had reached 88.6% in the field of agriculture, compared to 90.2% of the general population.⁶⁹ In 1966, 26% of all board members of the LPGs were women.⁷⁰ Concrete structures such as the Women’s Committees and Women’s Development Plans were set up comprehensively in the LPGs and intended to strengthen the role of women, advocate for their interests, and increase their leadership positions. By 1978, the proportion of women in leadership positions had reached 35%, in part due to the increase in their professional and educational qualifications. However, the process of increasing the number of women in leadership stagnated in the late 1980s and women remained underrepresented among LPG chairs until the end of the DDR. ⁷¹

Image 19. Two teachers and an educational assistant look after forty-seven children in the kindergarten of the LPG Walter Ulbricht in Jahna, founded in 1952.

Image 20. The Peasants’ Mutual Aid Association Women Farmers’ Commission in the Stollberg district followed the resolutions of the Farmers’ Congress to support women farmers and girls in their continuing education by offering special training courses. For fifteen weeks, forty-eight women and girls came to the district town for further training, including lectures, seminars, practical exercises, and field trips. The focus was on cattle breeding and husbandry because most of the participants worked in this field (February 1962).

Image 21. Anneliese Hartmann, a mechanisation expert from the V. I. Lenin LPG in Öttersdorf, operates a swather over green fodder areas in June 1978.

Anti-Imperialist Solidarity in Agriculture

As the DDR sought to bring the quality of life in rural and urban areas closer together, another pillar of socialist construction was cultivated among farmers: international solidarity.

On 1 September 1960, the Institute for Tropical Agriculture was founded at the Karl Marx University in Leipzig. Why was such an institute established in Germany, a country situated in a cool temperate climate zone? After losing its colonies at the end of the First World War, there was no longer any state interest in research and education relating to tropical agriculture. This changed with the DDR, albeit with an altogether different premise: young nation states that had emerged from anti-colonial struggles were to be actively supported on their path to economic liberation. Thousands of students from over sixty countries in the Global South were trained to develop agriculture in their home countries at the Leipzig Institute and a handful of other educational establishments, such as the School of Agricultural Engineering for Tropical Agriculture in Altenburg and the VEB Interagrarkooperation.

DDR citizens also attended courses and seminars designed specifically to prepare them for agricultural assignments abroad. Groups of DDR specialists, working in close collaboration with local experts and officials, supported the complex and protracted process of transforming feudal, colonially distorted, and backward agricultural structures into sovereign, state-controlled agricultural production.

Examples include pig specialists in Vietnam attempting to combine the fertility of the Vietnamese pot-bellied pig with the fattening capacity of German breeds, a tropical farmer in the People’s Democratic Republic of Yemen who, together with local officials, carried out the first thorough statistical surveys of livestock production, or a group of thirty-eight DDR specialists who, on the basis of a friendship treaty between the DDR and Mozambique and in close cooperation with local forces, helped to establish a state farm – eight of whom were killed by counter-revolutionary forces in 1984.⁷² Furthermore, the development of the first coffee-growing regions in Dak Lak, Vietnam – supported by 32 million roubles⁷³ – enabled the country to become the world’s second-largest coffee producer.⁷⁴

The dispatch of tractors and agricultural machinery, as well as the setting up of repair workshops, was financed by donations from DDR citizens, including contributions from farmers. In thousands of cooperatives, farmers in the DDR set up solidarity groups and supported campaigns such as ‘Farmers Help the Farmers of Vietnam’, which sent spraying equipment to Vietnam to rehabilitate crops poisoned by the US chemical attacks.

Image 22. Members of the Oskar Schneller LPG in Mansfeld express their solidarity and friendship with the Vietnamese people in the presence of Bui Tru, the second secretary of the embassy of the Democratic Republic of Vietnam in the DDR (July 1972).

The Path to Large-Scale Agricultural Production

In the late 1960s, when land reform had been fully implemented and the cooperative movement had reached its height, a third central phase of agricultural development began. During this phase, strategic goals were established that would guide agricultural planning until the beginning of the 1980s: the consolidation of small- and mid-sized farms into larger cooperatives, specialisation in parts of the production process, and collaboration with processing industries as prerequisites for large-scale industrial-like production in the countryside. Hans Reichelt, long-time minister of agriculture, explained the necessity of industrial agricultural production under the conditions of a socialist society:

This was the path towards creating larger economic units in agriculture – small-scale farming is as ineffective as keeping livestock in small stables. Some may object here: ‘Yikes, large-scale livestock farming!’ But industrial-like production in agriculture is not bad per se. It is objectively necessary to feed humanity – this is not possible with nineteenth-century production methods. What is bad is when all moral and ethical rules, including animal welfare, are ignored for the sake of profit. However, this stems from the character of the society. When corporate wholesale buyers force farmers to sell their produce below the production price, and when shareholders and banks speculate with food on the stock exchanges, this inevitably has an impact on production.⁷⁵

As a result of the increased investment capacities and efficiency that came with years of accumulated experience in cooperative-style production, all LPGs gradually developed into Type III LPGs (the most advanced cooperative structure, through which farmers incorporated all of their land and homestead into the cooperative, which functioned as an independent enterprise). Through a 1963 decree of the Council of Ministers, most MTS machinery was handed over to the Type III LPGs free of charge while the Type I and Type II LPGs were able to purchase it at a low price and in instalments.⁷⁶ Many MTS workers, tractor drivers, and technicians became members of cooperatives. This meant that the use of technology could be coordinated more effectively and integrated into the work of LPGs. Finally, LPGs also merged to form larger ‘joint LPGs’. The total number of cooperatives decreased, while the agricultural area farmed by each cooperative increased. Where previously there were 19,313 LPGs working roughly 5.4 million hectares of land in 1960, in 1988 3,855 LPGs worked the same amount of land.⁷⁷

Data from a collective of authors, 40 Jahre erfolgreiche Entwicklung zu einer leistungsfähigen sozialistischen Landwirtschaft in der DDR [40 Years of Successful Development Towards Efficient Socialist Agriculture in the DDR] (Eggersdorf: agrabuch, 1989), 58.

This increase in agricultural productivity was driven by the need to meet the growing demand for food and raw materials, which were necessary inputs in a number of industries, such as the textile, leather and fur, pulp and paper, chemical and pharmaceutical, woodworking, and cultural industries. In 1988, almost 65% of the DDR’s primary raw materials were produced in the agricultural sector.⁷⁸ This increase in agricultural productivity also freed up urgently needed labour that could be deployed to other sectors of the economy; promoted the efficient use of agricultural machinery; and created the conditions to increase the level of training that farmers and agricultural workers received – all factors that in turn further increased productivity.

Image 23. Specialists Klaus Ulrich and Franz Mally check a computer-controlled irrigation system in an LPG in Dambeck.

Image 24. Workers use a combine harvester in an LPG in Werben.

Agricultural planning took place at the central, local, and farm levels. Until then, LPGs had been relatively self-contained, now they become increasingly integrated into the overall social planning and management of the DDR’s economy. The Agricultural-Industrial Associations (Agrar-Industrie-Vereinigungen, AIVs) that emerged in the mid-1970s were the culmination of this process. AIVs were an institutionalised form of cooperation between various LPGs, Publicly Owned Estates, inter-enterprise facilities used by more than one organisation also known by their German acronym VEGs, and the upstream and downstream enterprises in the service, processing, and manufacturing sectors. By 1989, thirteen such production complexes (AIVs) had formed in the DDR, which together farmed 6.6% of the DDR’s total agricultural land.⁷⁹

By the early 1980s, concrete problems arising from the trend towards specialisation and excesses in large-scale production led to course corrections and debates. For example, the specialisation and separation of plant and animal production disrupted the direct exchange of natural fertiliser and animal feed and led to logistical difficulties. Increased transport expenditures and costs for large estates strained production. In animal production, the effort to tend to veterinary hygiene and disease prevention noticeably increased. Ecological issues caused by the excessive use of mineral fertilisers rather than environmentally friendlier options were corrected over time. Energy-intensive machine systems were susceptible to energy price spikes on the global market, which had a particularly severe impact on the DDR, especially in the beginning of the 1980s.

The effort to improve the integration of LPGs into the DDR’s social and economic plans through AIVs did not signal a plan to change the ownership structure of the land. As an official information brochure on agriculture explained:

Time and again, visitors to our republic ask if and when land will be nationalised. This question is not on the table, because nobody is thinking of changing the forms of property in the countryside. Most of the land is the private property of the farmers. Its cooperative use, and the distribution of its yield to the LPG members according to their share of the work performed, creates every opportunity to further advance the [DDR’s] developed socialist society.⁸⁰

Though private ownership undoubtedly had an influence on the consciousness of individual farmers, on the whole, the increasing integration of the LPGs into the unitary structures of economic planning and democracy in the DDR prevented private land ownership from developing economic or political significance. The relationship between the different actors in rural areas and agricultural production was not without friction. As Klaus Schmidt, former director of the DDR’s Institute for Agroeconomics, pointed out:

Particularly where the AIVs managed the territory of an entire municipality or where their sphere of influence even extended beyond the municipality’s boundaries, there was a tendency for them to increasingly take on the official duties that had hitherto been managed by state organs. A fundamental problem that arose with the development of the AIVs was the question of the prospects and opportunities of individual units within the AIV, specifically to what extent they could maintain their independence.⁸¹

Such contradictions and problems, alongside the misguided separation of plant and animal production in the 1970s, should have resulted in lively debate about the future of agriculture. The issue of balancing land ownership and independence with the further integration and centralisation of LPGs would certainly have remained the subject of political debate had the DDR continued to exist. As long as the SED and the wide-ranging democratic structures in rural areas – from the agricultural councils from the municipal to central levels to the farmers’ organisations, and numerous scientific institutions and agriculture publications of the time – responded to these contradictions and tasks in an agile, self-critical and proactive manner, they were able to drive development forward.

5. Conclusion

The development of socialist agricultural relations was brought to an abrupt end with the dissolution of the DDR in October 1990. Structural changes were imposed at breakneck speed, transforming the social and economic reality of farmers.

At the time of the DDR’s dissolution, 55% of the land cultivated by the LPGs was owned by individual LPG members. The remaining 45% was owned by the state or leased to the LPG by private owners. After 1990, this 45% was rapidly privatised, removing land from the LPGs’ existing economic structures and providing a lever for their dissolution in favour of private farming. LPGs seeking to continue to produce had to lease or purchase this land, often at additional cost, while completing the transition to a new legal form used for cooperatives in the West by the end of 1991. The transformation process placed considerable pressure on existing LPG structures.⁸²

This and other policy shifts in the post-1990 period created a rural exodus, especially of young and qualified workers. Of the former 923,000 agricultural workers in the DDR, 743,900 were laid off by 1993 and 4 out of 5 were pushed to leave the countryside.⁸³ The social and cultural infrastructure was destroyed. Although the transformation in land ownership established by land reform was legally protected by the so-called reunification treaties signed by the last DDR government, the landowning elite that had controlled the countryside before 1945 was able to use loopholes to reclaim parts of their expropriated land.

Yet, despite the radical capitalist turnaround, the history of the DDR has left a visible mark on agricultural structures in Germany. On average, farms in East Germany cultivate much larger areas than those in West Germany and have maintained elements of the structures established by LPGs, such as machine sharing under the new legal categories of the market economy.⁸⁴

In less than fifty years, revolutionary ruptures in the living and economic relations of the countryside were seen in East Germany, from the remnants of a semi-feudal, hierarchical social and property order to a collective way of working and living based on cooperation. The land reform that began in 1945 created the conditions for the transformation in the countryside and the dismantling of the political and economic monopoly of the landowning classes. Only by ending their rule was it possible to establish and develop new economic and social relations in East Germany. With the expropriation of the large Junker estates, land reform in the DDR ended the rule of the feudal aristocracy and agrarian capitalist classes. Agricultural workers, landless and smallholder peasants, and the general village population were liberated from the system of economic exploitation and semi-feudal disenfranchisement that had been in place for centuries. Under the conditions created by the reality of Soviet occupation after the Second World War and the leadership of the workers’ movement and parties, land reform was achieved in this part of Germany, and the oppressed became the protagonists shaping agriculture and village life.

The phase of collectivisation in the post-war years (approximately 1952–1960) illustrates the complexity of social development when a change in the thinking of those involved is crucial to its success. The SED had to manoeuvre between the strategic goals of socialist construction and the adverse conditions on the ground. Between retreat and advance, external and internal contradictions, and the necessities and prospects of economic development, the SED constantly reassessed the speed and force of its leadership in this comprehensive development process.

The complexity of the path to socialism, the protracted nature of some processes, especially in people’s consciousness, and the influence and sabotage of political opponents had to be taken into account and counteracted. This required agility and clarity of purpose, but also the ability to be self-critical, which the political leadership generally demonstrated over long periods of time.

While social hardship and individual cases of injustice cannot be denied, it is necessary to understand such instances in the contradictory and complex context of post-war East Germany and to understand that they represent exceptions, not the rule. Without the leading role of the peasantry, the transition towards cooperatives would have failed. Naturally, from the bourgeois perspective, the transition away from the dominance of private property is to be condemned on principle. Narratives of ‘forced collectivisation’ that dominate the public discourse in Germany today do not seek to understand the processes in the DDR but rather to promote anti-communist prejudices. Individual fates, in which there were also tragic conflicts and entanglements, are used to obscure the view of the broader development and prevent a serious examination based on historical facts. This method of framing can be seen over and over in narratives about socialist states.

Cooperative forms of agricultural production were a cornerstone of East Germany and a decisive building block for the development of the DDR as a socialist state. Existential financial angst, which is inevitable under the ‘freedom’ of private property and competition, was overcome in the DDR through mutual aid among peasants and their integration into state planning. While some see the continued existence of cooperative ownership alongside state ownership as a major obstacle to the deepening of planned economic relations, this was not the case in the DDR. The experiences of the AIVs certainly pointed to contradictions between the different forms of ownership that needed to be resolved, but the cooperatives were already deeply integrated into the overall economic processes of the DDR through jointly developed plans and thus binding agreements on the distribution of funds as well as the fixed-price system. While the LPGs could exercise some economic autonomy regarding their internal workings, there would be no significant effect of these autonomous decisions on the broader economy.

The cooperative way of thinking, working, and living in the DDR democratised the villages by eroding social hierarchies, expanding leisure time, and promoting extensive opportunities for cultural, sporting, and social activities. Cultural centres, kindergartens, village stores, cafeterias, comprehensive medical care, continuing education opportunities, vacation sites, sports teams, music groups, and much more were an integral part of village and community infrastructure. It was the rural population itself that played a key role in building these structures and sustaining them.

The DDR was largely able to produce its own agricultural products, and hunger and poverty became things of the past. Today, in view of soaring rates of global hunger, the omnipresent overexploitation of nature, and the market dominance of agricultural multinational corporations, some political movements seek the solution in small-scale production. The DDR followed a different path, which – despite a decreasing workforce – was able to avoid the negative effects of capitalist farming while increasing yields by industrialising large-scale agricultural production on the basis of socialist relations.

The abrupt termination of this project in 1990 and the developments in the years since elucidate the essential characteristics of capitalist transformation: privatisation, the dismantling of ‘unprofitable’ social and cultural infrastructure, land speculation, and rural exodus. Farmers had and still have to fight their own way through the chaos of liquidation, privatisation, and fierce competition, pitted against a constructed false enemy/scapegoat of ‘cheap’ labour provided mostly by migrants.

The dismantling of socialist infrastructure in the countryside furthered the tendency towards isolation among rural populations. While the renewed reign of private property postured itself as offering increased freedom of choice to farmers, this ‘freedom’ is inherently limited by constraints of the market and comes at the expense of farmers’ roles as co-creators of society. Along with it came a new rule of order: profit as the determining factor in agriculture. While the conditions today differ greatly from those in 1945, the slogan ‘The land to those who work it’ is yet again on the agenda in East Germany and across the world.

Notes

Image Credits

As I sat on a plane from Hong Kong to Beijing, I reread the speeches that Lu Xun – known as the father of modern Chinese literature – had delivered in Hong Kong almost exactly a century earlier. It is commonly said that around 1927 Lu Xun bid farewell to fiction and made a leftward turn ideologically. He increasingly turned to zawen – short, polemical essays – as his principal means of intervening in the political and cultural struggles of his time, awakening the Chinese people, and giving voice to a nation living under semi-colonial and semi-feudal conditions. To mark ninety years since Lu Xun’s passing in 1936, this art bulletin revisits several zawen from the final decade of his life.

One such text is ‘Voiceless China’, a speech Lu Xun delivered at the Chinese YMCA of Hong Kong on 18 February 1927. Lu Xun championed baihua, or vernacular written Chinese, over Classical Chinese. He criticised the use of an ‘archaic language good only for expressing ancient and obsolete ideas’. In a society where illiteracy remained widespread, the distance between Classical Chinese and the language of everyday life placed yet another barrier between most people and the written word, while Classical Chinese itself could no longer adequately express their thoughts, feelings, and struggles. Lu Xun therefore called on young people to ‘transform China into a China with a voice’, insisting that only then could they ‘live in the world with others’. For him, this search was also part of a wider internationalist struggle, with China’s liberation bound up with that of other colonised peoples. ‘Have we heard voices from Egypt?’ he asked. ‘Have we heard the voices of Vietnam and Korea? Aside from Tagore, does India have any other voices?’

Against Despair

Käthe Kollwitz, Sharpening the Scythe, 1905.

The following day, on 19 February, Lu Xun delivered another speech at the Hong Kong YMCA, returning to the question of how China’s new voice might be created. In ‘The Old Tunes Are Finished’, he looked to revolutionary Russia – ten years into its socialist revolution – as a guiding light, a society where a new voice had emerged and the old songs were no longer being sung. For Lu Xun, this process was neither solitary nor individualistic. One person’s awakening could give rise to others, linking individual transformation to a broader collective awakening.

The year 1927 was tumultuous for China. By then, the republic established after the 1911 Revolution, which had brought an end to more than two thousand years of dynastic rule, had failed to overcome warlord rule, feudal oppression, and foreign imperial domination. With the support of the Comintern, the Chinese Communist Party (CPC) and the nationalist Kuomintang (KMT) had formed the First United Front, a short-lived alliance aimed at defeating the regional warlords and unifying China. One of its most important military institutions was the Whampoa Military Academy, founded with Soviet assistance, where Nationalist and Communist cadres trained alongside one another. Chiang Kai-shek served as its commandant, while Zhou Enlai had directed its Political Department. On 8 April 1927, the academy invited Lu Xun to speak on ‘Literature in Times of Revolution’. Reflecting on the persecution of radical students he had witnessed in Beijing, Lu Xun lamented, ‘Literature, oh literature, you are a most useless thing. Only those without power talk about you; no one with real strength bothers to talk, they just murder people. Oppressed people who say a few things or write a few words will be killed’.

Käthe Kollwitz, March of the Weavers, from the series A Weavers’ Revolt, 1893–1897.

Given the situation of China at the time, Lu Xun asked a sobering rhetorical question: ‘What value does this literature have for the people, then?’ Yet he never ceased to write, making it his personal philosophy to ‘resist despair’. In Wild Grass, his collection of prose poems published that same year, he confronted hope and despair not as simple opposites but as illusions that could both paralyse action. Rather than retreat into a lost past or defer everything to a promised future, he insisted on grappling with the present through struggle and creation.

Revolutionary times do not necessarily produce revolutionary literature. For Lu Xun, it was not enough simply to bear witness to and condemn suffering and injustice; it was necessary to cultivate revolutionaries and a revolutionary programme. ‘Only when revolutionaries start writing will there be revolutionary literature’, he stated in his speech. Similarly, it was not enough to produce literature about peasants and workers. ‘The true liberation of the workers and peasants’, he argued, was necessary to make possible a true ‘people’s literature’. Revolution, in turn, could liberate the oppressed classes and, at the same time, liberate literature – a literature that, in Lu Xun’s words, ‘wants to resist, and it wants revenge’. He ended his speech at the military academy by contrasting the force of literature with that of armed struggle: ‘For myself, I’d naturally prefer to hear the sound of artillery. It seems to me that the sound of artillery is a much finer thing than the sound of literature’. Four days later, events in Shanghai would give those words a brutal immediacy.

A Revolutionary Literature Written in Comrades’ Blood

Käthe Kollwitz, Outbreak, from the Peasants’ War series, 1902.

A mere four days after Lu Xun’s speech, in a betrayal of the First United Front, Chiang Kai-shek’s KMT forces and their allies massacred Communist-led workers in Shanghai, initiating both the anti-communist repression that became known as the White Terror and the collapse of the United Front. As the repression spread, the CPC responded with a series of armed uprisings, including at Nanchang, along the Hunan-Jiangxi border and in Guangzhou. Their defeat, together with the White Terror, contributed to the CPC’s turn towards agrarian revolution and rural armed struggle. At the 7 August Emergency Conference, held clandestinely in Hankou, now part of Wuhan, the CPC adopted what was later summarised as the general orientation of agrarian revolution and armed resistance to the reactionary KMT. Mao’s words at that meeting later became the well-known phrase, ‘political power grows out of the barrel of a gun’. The following month, Mao led the Autumn Harvest Uprising, whose forces formed the First Division of the First Workers’ and Peasants’ Revolutionary Army before retreating towards the Jinggang Mountains.

Amid the White Terror, Lu Xun left Guangzhou, where he served as dean and head of the Chinese Language and Literature Department at Sun Yat-sen University. He settled in Shanghai, where he spent the remaining nine years of his life politically active and engaged in the systematic study of Marxist and Soviet literary theory. In 1930, Lu Xun became one of the founding members and leading figures of the League of Left-Wing Writers. In his speech, ‘Some Views on the League of Left-Wing Writers’, he warned against ‘salon socialism’ – radical talk cultivated in the safety of the sitting room but detached from actual social struggle. Revolution, he told the roomful of writers and poets, is ‘painful, inevitably mixed with dirt and blood, and nothing at all interesting and perfect as the poets imagine’. What is asked of the writer is lowly work – not the drama of destruction but the slower, thankless labour of construction. Within a year of the League’s founding, Lu Xun’s warning took on flesh. On 17 January 1931, KMT police in the Shanghai International Settlement arrested a group of Communist organisers at a clandestine meeting. Five young writers connected to the League were among those arrested and, on 7 February, they were secretly executed at Longhua. In response, Lu Xun wrote that ‘the first chapter of revolutionary literature of the Chinese proletariat… has been written with our comrades’ blood’.

Let Them Go on Hating Me

Käthe Kollwitz, Unemployment, 1909.

In the last year of his life, 1936, with failing lungs, Lu Xun wrote more than thirty essays on subjects ranging from the New Woodcut Movement – the socially engaged printmaking movement he had helped launch and champion – to the killing of the young writers at Longhua and the literary debates of the time. ‘Death’, written on 5 September 1936, sits at the centre of this final collection. It opens not with Lu Xun’s own mortality but with that of the German socialist printmaker Käthe Kollwitz, whose work depicted war, poverty, working-class struggle, and mourning. Earlier that year, Lu Xun had prepared Collected Prints of Käthe Kollwitz, a volume of twenty-one reproductions of her prints, published in Shanghai in July 1936, and had asked the US journalist and activist Agnes Smedley, who knew Kollwitz personally, to write a preface. Smedley’s preface recounted a conversation with Kollwitz in which she asked why death had become such a persistent theme in her later work. Kollwitz replied that perhaps it was because she was growing older.

The exchange led Lu Xun to reflect on death and on Chinese attitudes towards it – from the condemned peasant who calmly faces execution to people of modest means preparing materially for the afterlife by choosing their burial place and coffin, burning paper money, and relying on their descendants to make offerings after their death. Later in ‘Death’, Lu Xun left seven instructions as if he were writing a will. The fifth was addressed to his son: if he had no particular talent, he should find an ordinary job and ‘under no circumstances should he become a phony writer or artist’. The seventh warned his family to have nothing to do with people who injure others while preaching tolerance and opposing revenge. From there Lu Xun arrived at the essay’s most famous line: ‘I have a great many enemies, and what should my answer be if some modernised person asked me my views on this? After some thought I decided: Let them go on hating me. I shall not forgive a single one of them either’. The Kollwitz prints reproduced throughout this art bulletin are drawn from the collection Lu Xun prepared in the final months of his life.

Cultivating Soil Is More Important than Simply Demanding Genius

One of my favourite zawen essays by Lu Xun is from 1924, based on a lecture he gave to the alumni association of the middle school affiliated with Beijing Normal University. In ‘Before the Appearance of Geniuses’, he reflects on the metaphor of the soil. Five years after the May Fourth Movement, amid continuing calls for cultural and social transformation in China, one of the loudest demands in literary circles was for the appearance of a ‘genius’ to save culture. Geniuses, Lu Xun said, do not spring from deep forests; they emerge from the masses that nurture them, from the soil that lets them grow. Yet the same society that called for geniuses was destroying the soil from which they might emerge. Antiquarians, nativists, and hack critics trampled tender shoots while still calling for the flower. In a China where destructive critics dominated, ‘the masses in this kind of environment are but dust, not fertile soil’. He closed with a patient instruction: ‘cultivating soil is more important than simply demanding genius’.

It has been ninety years since Lu Xun returned to the soil. He did not live to see the worst of Japanese fascism – which claimed nearly 24 million Chinese lives – nor the eventual Communist victory in 1949, the flower that grew out of the soil of the Chinese Revolution. As he wrote in the last line of ‘Before the Appearance of Geniuses’, ‘cultivating good soil bears its own rewards. When a lovely flower emerges from the soil, people will naturally appreciate it, and the soil itself can also take pleasure in appreciating it’. And the soil, if we take it seriously, may itself have a spirit, a voice, and a new tune of its own.

Warmly,

Tings Chak, co-coordinator of the Art and Culture Department of Tricontinental: Institute for Social Research.

When you use AI to write, read reports, or work with files, you may encounter terms such as token, context, and agent. These terms help explain what happens when you give an AI application a task: what information it can use, how it produces a response, and whether it can take actions such as searching for sources or saving a document.

This guide introduces the main terms related to large language models (LLMs) through practical examples. It begins with how models process information, then explains how to give them instructions and reference material, how tools and agents carry out tasks, and how to organise ongoing work. It also introduces ways to access models through other software or run them on your own equipment.

1.  How Language Models Work

Large Language Models

A large language model (LLM) is an AI model trained on vast amounts of text and other data to learn patterns in language. ‘Large’ refers to its scale, including the number of internal parameters adjusted during training.

When generating text, a typical LLM uses the information available to it – including your instructions and the text already written – to predict what could come next. More precisely, it predicts the next token, a small unit of text such as a word or part of a word. It selects a token, adds it to the text, and repeats the process. In this way, it builds a response step by step.

Many people first encounter LLMs through a question-and-answer conversation. But the same process also allows models to summarise documents, translate, and draft text. A fluent response, however, is not necessarily a correct one: generating plausible language does not guarantee factual accuracy.

Token

A token is a small unit of text that a language model processes. Before your message reaches the model, a tool called a tokeniser splits it into tokens. A token may be a whole word, part of a word, or a punctuation mark. Spaces can also form part of a token.

For example, a tokeniser might keep ‘cat’ as one token but split ‘playing’ into ‘play’ and ‘ing’. This is an illustration, not a fixed rule: different tokenisers can split the same text differently. A token is therefore not the same thing as a word, and counting words will not give you an exact token count.

Each token is assigned a numerical ID so that the model can process the text as numbers. When generating a reply, the model predicts tokens one at a time. These tokens are then converted back into the text you see on screen.

Token counts matter when using a model. Input tokens are the tokens in the material supplied to it, such as your instructions, documents, and any earlier messages included in the conversation. Output tokens are the tokens it generates. Many model providers charge for API use based on these counts, often at different rates for input and output. Models also have limits on how many tokens they can process or generate in one request.

Longer texts generally require more tokens. The language also makes a difference: expressing the same meaning in different languages can produce different token counts. To find the exact count for a piece of text, you need the tokeniser used by that model.

Pre-trained Knowledge

Token prediction explains how a model produces text, but it does not yet explain why a model can answer questions without being given any external files. The answer lies in what it learnt during training. During training, a large language model learns patterns and information from massive amounts of text, code, and other data.

This is also why, even if we do not provide the model with any external files, it can still answer a large number of questions about common sense, history, maths, programming, and so on. This content is not obtained by the model through temporary searching. The model has already learnt patterns related to this information during training. This information is often described as the model’s pre-trained knowledge.

However, the model does not have a perfect database of facts. It learns patterns from its training data, and these patterns can sometimes be incomplete, outdated, biased, or simply wrong. When the model does not have enough reliable information to answer a question, it may still try to produce an answer based on what it has learnt. This is one reason hallucinations can happen.

Multimodality

Pre-trained knowledge is still mainly about patterns learnt from text and other training data. But is there a model that can accept different types of input beyond natural language, or generate content other than ordinary text, such as images, audio, and video? The answer is yes. Such models are called multimodal models.

Imagine uploading a photograph of a flower and asking, ‘What kind of flower is this?’  The system prepares two kinds of input. Your question is split into tokens and converted into numerical IDs, as described above. The photograph is processed separately into a numerical form that allows the model to work with visual details such as colours, shapes, and the arrangement of petals.

The model then uses both the visual information from the photograph and the question you have asked to generate a reply. This ability to combine different types of information is a key feature of multimodal models. Models that support audio or video also convert those inputs into numerical forms they can process, although the methods vary between models.

Of course, this does not mean that models supporting only natural language have no external channels at all for understanding images. A common approach is to use image text recognition or audio transcription tools to convert different types of data into general text, which is then input into the model.

To some extent, however, this approach involves loss. Imagine that, for a human being, reading a textual description of an image is still inferior to viewing the image directly. So multimodal models are generally better at working directly with these different types of information.

2. Giving Models Instructions and Information

Prompt

Understanding how a model works internally is useful, but in daily use, the first thing that matters is how we communicate with it. To make a large language model give a high-quality answer, how you express the task to the model matters greatly, in addition to the model’s own abilities. When interacting with a large language model, the instructions, questions, background information, constraints, and examples we provide can generally be called a prompt.

So a prompt is not merely a simple question. For example, ‘help me write an email’ can count as a prompt. But if you add the email’s use case, goal, tone, format, and specific requirements, you can form a more complete prompt. The process of designing and refining this information so that the model can complete a task more accurately and consistently is usually called prompt engineering.

From the point of view of usage, what we normally type directly into an AI input box is usually a user prompt, that is, a prompt provided by the user. In addition, the system may provide the model in advance with a special kind of prompt called a system prompt. This is used to define the model’s overall behaviour, role, capability boundaries, and answering style.

In the usual instruction hierarchy, the system prompt has higher priority than the user prompt. So even if the user gives an instruction that conflicts with the system’s requirements, the model usually needs to follow the system prompt first. This is also why the same large language model may behave differently in different AI applications. Apart from the model itself, the system prompt and other contextual information provided by the application can also have an important impact on the final result.

For practical examples of how to structure instructions, provide context, and specify the output you need, see How to Talk to AI – A Prompt Engineering Guide.

Context

When a model answers a question, it can draw on more than the latest message you typed. The context is the information available to it when generating that response. This can include system instructions, your current question, earlier messages and replies, and any document passages or search results supplied to the model.

For example, suppose you provide an article and ask for a summary. You then ask, ‘How does the author support the main argument?’  To answer this follow-up question, the model needs access to the article as well as your new question. The application usually includes earlier material in each new request. However, a message remaining visible in your chat history does not guarantee that its full text is still available to the model.

The amount of information a model can handle at once is limited by its context window, measured in tokens. For many models, this window must accommodate both the input and the response being generated. Instructions, articles, earlier messages, and replies all take up space. The limit therefore concerns more than the length of a single question.

As a conversation grows, applications manage this limit in different ways. Some ask you to start a new conversation; others remove older material or replace it with a summary. Summarising earlier material to make room for further conversation is often called compaction. For example, a long discussion about editing an article might be condensed into a short summary of the main decisions and remaining tasks. This allows the conversation to continue, but details or qualifications may be lost.

If a model appears to have ‘forgotten’ something, that information may no longer be in its current context, or it may have been left out of a summary. However, models can also overlook or misuse information that is still available. For important tasks, restate essential requirements, provide the relevant passages again when needed, and check the answer against the source. Useful context is relevant and clearly organised; adding more material does not automatically produce a better answer.

Memory

Context is the information a model can use when generating its current response. But some information may also be useful in future conversations. Memory is a feature that allows an AI system to save information for later use. The application typically stores it outside the model, for example in files or saved records, and adds relevant information to the context when needed. Saving a memory does not retrain the model or increase its context window.

For example, in an application with memory enabled, you might ask it to remember, ‘Please use British English in writing.’  The application can save this preference and include it in the context of later writing tasks, reducing the need to repeat it. Memory can also preserve project information, such as the intended audience for an article, editorial decisions you have agreed on, or tasks that remain unfinished.

What is saved and how it is used varies between applications. Some let you explicitly choose what to remember; others identify useful information from conversations. Saved information may become outdated, may not be brought into a particular response, or may be applied incorrectly. For important tasks, it is still useful to restate key requirements and check that the answer follows them.

Retrieval-Augmented Generation

A large collection of documents may exceed a model’s context window, while only some passages are relevant to your question. The model may also need information it did not encounter during training, such as internal documents or recent reports.

Retrieval-augmented generation (RAG) is a way of finding relevant information and using it to help a model answer. The process has three main steps: the system searches a source or collection of sources for relevant material, adds that material to the model’s context alongside your question, and asks the model to generate an answer using it. The retrieved information provides reference material for that response; it does not retrain the model.

For example, suppose you have a collection of research reports and ask, ‘What was the unemployment rate in South Africa in 2023?’  The system searches the reports for passages or tables that could answer the question. It then supplies the relevant extracts to the model, which uses them to formulate a response. Where the system preserves source details, the answer can also include references so that you can check the figure and its context.

The sources available for retrieval are often organised into a knowledge base: a collection of documents or other information that the system can search. This might contain research papers, institutional reports, or internal guidance. Web search can also form part of a RAG process when the system retrieves information from webpages and supplies it to the model to help answer a question.

RAG can help a model produce answers grounded in specific sources while reducing the amount of material placed in its context. However, it does not guarantee accuracy. The system may retrieve the wrong passage, miss an important qualification, or use an outdated source; the model may also misinterpret the material. For important claims, check the cited source and whether it actually supports the answer.

3. From Answers to Actions

Tools

So far, we have looked at how models generate answers and use information. But many tasks also require actions, such as finding a report online, updating a document, or playing a song. Tools allow an AI application to carry out these operations.

A tool is a function that a model can call to perform a specific operation. For example, a web search tool retrieves information from the internet, while a file tool reads or modifies documents. The model requests an operation, the application executes it, and the result is returned to the model.

Suppose you ask an AI assistant to revise an article and save the changes. It can call a file-reading tool to access the draft, use the model’s language capabilities to revise the text, and then call a file-writing tool to save the revised version. The model produces the edits; the tools provide access to the document and carry out the changes.

Agents and Workflows

Tools allow an AI application to perform individual operations. Completing a task may also require deciding which operation to perform next, checking its result, and adjusting the approach. An agent is an AI system that uses a model to guide this process towards a goal, without requiring the user to specify every step.

An agent typically works through a repeated cycle: the model chooses an action, a tool carries it out, and the result is returned to the model. The model then uses that feedback to decide what to do next. This might mean continuing with the task, trying a different approach, or asking the user for clarification.

Consider the article-editing example above. You ask the assistant to revise a draft according to an attached style guide and save a new version. An agent can read both documents, identify the changes needed, revise the article, and check the result against the guide. If it finds inconsistent spelling or a missed requirement, it can make another revision before saving.

An agent may stop when it judges the task complete, needs your input, or reaches a limit set by the application. Its output still needs to be checked against the original requirements.

Tasks can also be organised into a workflow: a predefined sequence of steps, such as reading a draft, checking its spelling, and saving a revised version. In a workflow, the main steps are specified in advance; an agent can decide which steps to take as the task develops. The two approaches can be combined. For example, an editing workflow might include an agent that decides how to resolve problems found during review.

4. Organising and Extending AI Work

Workspace

Once AI can carry out tasks, the next question is how to organise longer projects. If you only occasionally ask AI a question, a chat window is enough. For longer projects, such as writing a book or developing software, repeatedly explaining the files, background, and progress takes time.

Many AI tools provide something called a workspace. You can think of a workspace as a dedicated working area for a project. It may contain files, instructions, previous outputs, and other information related to the project. The application can access relevant materials from the workspace when needed; this does not mean that every file is included in the model’s context at once.

For example, we can create a ‘writing project’ workspace and put articles, images, and reference materials into it. Afterwards, when we ask AI to revise an article, organise materials, or continue writing the next chapter, it can work with these materials without our needing to re-upload all the materials every time.

For a step-by-step guide to organising project folders and working with an AI agent in Visual Studio Code, see our AI workspace setup guide.

Skill

When you ask AI to carry out similar tasks repeatedly, you may find yourself explaining the same requirements each time. A skill is a reusable set of instructions and supporting resources for a particular task. It gives the application a prepared approach to follow, reducing the need to explain the process from scratch.

For example, suppose you regularly edit articles for a publication. You could create an editing skill that instructs the AI to use British English, preserve the author’s argument, check punctuation, and present suggested changes with brief explanations. The skill could also include the publication’s style guide and examples of previous edits. When you use it with a new draft, the application can draw on these instructions and references to review the article.

You can reuse the same skill with different articles and update it when your editorial requirements change. This helps maintain a consistent approach while allowing you to add instructions specific to each draft.

To learn how to find, install, use, and create skills for your own tasks, see How to Install, Use, and Create Skills.

Model Context Protocol

Model Context Protocol (MCP) is a shared standard that helps AI applications connect to other software and use its tools or information. You can think of it as a common connector: when both sides support MCP, they have an agreed way to work together.

For example, suppose you are editing an article and the reports you need are stored in your organisation’s document library. With a suitable MCP connection, your AI application can search that library and retrieve relevant passages to help check the article. This can save you from finding and uploading each report yourself.

A skill and an MCP connection can support different parts of the same task. The skill provides instructions for reviewing the article; the MCP connection gives the application access to the reference materials.

5. Accessing and Running Models

API Key / Auth Token

An application programming interface (API) is a way for one piece of software to ask another to do something or provide information. For example, a writing application can use DeepSeek’s API to send an article to a DeepSeek model and request ‘Summarise this text’. The model generates a summary and sends it back to the writing application, which displays it for you.

To use that API, the application may need an API key or an authentication token: a credential sent with the request to show that it is authorised to access the service. Depending on the service, this credential may also link usage to an account for billing.

Here, ‘token’ refers to an access credential, rather than the units of text described earlier.

Open-Source Models

An open-source model is released with materials and permissions that allow others to use, study, modify, and share it. Researchers and developers can examine how it works and adapt it for their own purposes, building on existing work rather than training a new model from scratch. Examples include publicly released models from China’s DeepSeek and Alibaba’s Qwen team, which users can download, run, and adapt for their own tasks.

For users, one practical benefit is the option to run a model on their own equipment. For example, a research organisation could use a downloaded model on its own server to summarise internal reports, keeping the documents within its own systems. Smaller models may also run on a personal computer, while larger ones require more powerful hardware.

Conclusion

The concepts in this guide describe different parts of working with AI: how a model generates a response, what information is available to it, and how an application uses tools to carry out a task. Understanding how these parts work together helps explain both the capabilities and the limits of the applications you use.

For your next task, consider what information the model needs, which actions the application must perform, and what requirements the result should meet. These questions can help you provide relevant materials, give clearer instructions, and check whether the work has been completed as intended.

Socialism is the path that the Chinese people chose to enter modernity. As President Xi Jinping stated at the 20th National Congress of the Communist Party of China, ‘Chinese modernisation is socialist modernisation pursued under the leadership of the Communist Party of China’.¹ In China, ‘modernisation’ (现代化) is understood in dialectical opposition to the ‘century of humiliation’ (百年国耻), the period of imperialist subjugation that extended from the First Opium War (1839–1842) until the victory of the Chinese Revolution in 1949. Chinese modernisation refers to a progressive historical process rooted in China’s particular conditions and directed towards socialism. The Chinese conception of modernisation therefore has a genealogy distinct from that of classical modernisation theory, which understands development as a linear transition from ‘traditional’ society towards the economic, political, and social institutions of the industrialised capitalist West.² Xi’s formulation makes this socialist orientation explicit. It is this understanding of Chinese modernisation as a socialist historical process that lies at the heart of this dossier.

Capitalism came into China not in the form of modern industry but through opium, gunboats, and unequal treaties that heralded a century of humiliation. Capitalism – as imperialism – did not socialise production and labour across China on a national scale. Instead, it smashed and atomised Chinese society, propping up the landlord class, warlords, and comprador elites tied to foreign capital. The historical task of the Chinese Revolution was to overcome this humiliation and create a movement towards scientific reason and human dignity that rejected the political subordination and social polarisation associated with capitalism. We call that movement socialism. But how could socialism be built in a poor country?

At the time of the revolution, China was one of the poorest countries on the planet. Only ten countries had a lower per capita GDP.³ There was no way to build a classless society, short of socialising poverty. That was no solution at all when modern industry was needed to provide a dignified life and to build the material and technological capacity required to defend national sovereignty against war, blockade, and sanctions. The communists of China therefore had the dual task of transforming social relations and developing the productive forces – the two legs of China’s revolutionary path towards socialist modernisation. The Chinese Revolution had to advance on both legs, with the emphasis shifting between social transformation and technological development at different historical moments.

The Chinese Revolution is one of the great transformations of the modern era. In less than a century, China moved from the humiliation of semi-colonial domination and semi-feudal mass poverty to becoming one of the principal forces shaping the contemporary world. This transformation did not emerge from the spontaneous operations of the market, the benevolence of the capitalist class, or the guidance of foreign powers. It came from a protracted revolutionary process led by the Communist Party of China (CPC) and rooted in the struggles of workers and peasants.⁴ Chinese modernisation unfolded through distinct historical steps, each responding to the contradictions inherited from the previous stage and each attempting to advance the material and social conditions of the Chinese people.

There have been three major steps in China’s socialist modernisation since 1949. The first, the period of socialist revolution and construction from 1949 to 1978, focused on laying the foundations in an impoverished and unstable country so that the productive forces could develop and national sovereignty could be secured. The second, the period of reform and opening up from 1978 to 2013, concentrated on rapid industrialisation, technological advancement, and the creation of social wealth. The third, the new era inaugurated under President Xi Jinping’s leadership in 2013, has sought to place common prosperity and high-quality development at the centre of socialist construction. These steps are not isolated from one another – they form a continuous historical process shaped by the logic of socialist modernisation.⁵

Step 1: Revolution, Stability, and the Foundations of Development (1949–1978)

When the People’s Republic of China was founded in 1949, the country was among the poorest in the world. More than a century of upheaval had devastated China’s social and institutional fabric. This included imperialist intervention, the Opium Wars and widespread opium addiction, unequal treaties, warlordism, the Japanese invasion and occupation of China from 1931 to 1945, and civil war. Modern industry was underdeveloped, agriculture was backward, infrastructure was shattered, literacy rates were extremely low, and life expectancy hovered around thirty-five years.⁶ China faced not only economic underdevelopment but also political fragmentation and social chaos. The first task of the revolution was survival and stabilisation. The revolutionary government had to consolidate national sovereignty, dismantle feudal structures and the comprador networks that linked domestic elites to foreign capital, and create the institutional basis for long-term development. Without this initial step, no later modernisation would have been possible.

One of the earliest achievements of the new government was agrarian reform. By 1953, nearly 47 million hectares had been confiscated from the landlord class, with some land also requisitioned from rich peasants. Although they constituted distinct rural classes, landlords and rich peasants together accounted for less than 10% of the rural population but had previously controlled 70–80% of the arable land.⁷ The land was redistributed to approximately 300 million landless and land-poor peasants, amounting to a near-universal redistribution of land among the rural poor. This transformed the countryside and broke the power of feudal elites who had dominated rural China for centuries. Land reform was not merely an economic policy; it was a social revolution that gave peasants dignity and political agency. In the language of the revolution itself, it was a process of fanshen (翻身): ‘to turn over’ or ‘turn the body’. The revolution’s first great achievement was therefore measured not only in hectares redistributed but in the ‘turning over’ of hundreds of millions of peasants, who became active participants in national development. ⁸ This understanding of development as both material redistribution and human transformation has defined Chinese modernisation ever since.

Alongside land redistribution, the government moved to stabilise the economy on new foundations. Through the public control of essential commodities and the elimination of feudal rentiers, the government broke the speculative networks that had driven one of the most severe episodes of hyperinflation in the modern era. In April 1949, monthly inflation exceeded 5,000%. The renminbi, first issued in December 1948, gradually replaced the multiple currencies then in circulation. By the early 1950s, the government had established a stable, unified national monetary system.⁹ This action stabilised prices and facilitated exchange throughout the country, creating the conditions for productive economic life. The government also took the radical step of repudiating the illegitimate debts that previous regimes had accumulated through more than 700 foreign loans contracted between 1852 and 1949.¹⁰ All of these initiatives helped wipe the slate clean for Chinese society and unlocked the resources needed for future development.

The new Chinese state also launched mass literacy and public health campaigns. Before 1949, illiteracy was widespread and medical services were inaccessible to most people. Through mass mobilisation, reforms to the Chinese written language, and the rapid expansion of rural educational infrastructure – including 182,960 rural reading rooms by 1956 – literacy rates rose from under 20% at the time of the 1949 revolution to 57% by 1959.¹¹ Meanwhile, the ‘barefoot doctor’ programme – which directly inspired the World Health Organisation’s 1978 Declaration of Alma-Ata on primary health care – extended basic healthcare to rural villages, greatly reducing infant mortality and the incidence of infectious diseases. Life expectancy rose, as the World Bank data shows, from roughly thirty-five years in 1949 to over sixty-three years by 1978. This rise in life expectancy alone represents one of the greatest improvements in human welfare in modern history. By 1978, China had a healthier and better-educated population than other countries at comparable income levels – an important foundation for the rapid development that followed.

Industrialisation was at the heart of the First Five-Year Plan (1953–1957). Supported in part by technical assistance from the Soviet Union, China began to establish a heavy industrial base through the programme known as the 156 Projects, which covered sectors including coal, steel, power, and machinery and was carried out by state-owned enterprises. Although China remained poor, these projects created the industrial foundation needed both for future development and for national defence by providing the materials, machinery, and energy needed to equip and sustain its armed forces. During this period, China expanded its iron and steel production capacity, with steel output quadrupling from 1.3 to 5.2 million tonnes over the course of the First Five-Year Plan. The country also built its first automobile and tractor factories and constructed major hydroelectric and thermal power facilities.¹² Together, these efforts laid an integrated national industrial and infrastructural foundation, including bridges over the Yangtze River and the restoration and expansion of rail links into remote parts of northwest China. New industrial zones were deliberately built in the interior rather than in the coastal treaty ports of the colonial era. Socialist planning enabled the mobilisation of scarce resources towards national priorities rather than private profit.¹³

Critics often focus solely on the difficulties and errors of this period, including the consequences of the Great Leap Forward (1958–1962) and the turmoil associated with the Cultural Revolution (1966–1976). These events undoubtedly caused suffering and political instability. Yet it is historically incomplete to reduce the entire period to these tragedies. The broader significance of the first step lies in the fact that China emerged from decades of humiliation as a sovereign and unified state capable of independent development. What is remarkable is that many of these advances were achieved under severe external constraints and the constant threat of war. In 1951, the United Nations General Assembly imposed an embargo on strategic goods destined for China and the Democratic People’s Republic of Korea (DPRK). The United States reinforced these restrictions through the Mutual Defence Assistance Control Act of 1951, commonly known as the Battle Act. These restrictions and broader US sanctions, which were designed to wreak havoc on the economy, continued throughout the 1960s, limiting the supply of essential goods and machinery, constraining the pace of development, and imposing a high degree of autarky on the economy.¹⁴ Moreover, after US-led forces invaded the DPRK and advanced towards the Chinese border in October 1950, China intervened militarily in support of the DPRK and in defence of its own national security. As a consequence of these threats, China held its first nuclear test in 1964 to establish a shield against attack by its enemies.

The collective leadership of the first generation of the CPC, led by Mao Zedong, must be understood in this historical context. The revolution achieved what earlier reformers and nationalist governments could not: territorial unification, political independence, the liquidation of feudal relations, and the creation of a state committed to the welfare of the masses. The revolutionary ethos fostered a culture of collective sacrifice and national purpose. Even Western development economists later acknowledged that the foundations for China’s economic rise were laid during this socialist period through land reform, basic industrialisation, education, and public health.¹⁵ The emphasis on universal education produced a disciplined and increasingly skilled population. Rural industrialisation and grassroots governance helped create a decentralised and responsive administrative structure. When economic reforms began in 1978, China already possessed a cohesive state apparatus, a healthy labour force, and basic industrial capacity.

Step 2: Reform, Growth, and the Creation of Social Wealth (1978–2013)

By 1978, Chinese leaders recognised that although the country had achieved sovereignty and stability, it remained economically underdeveloped. The productive forces were still too weak to provide prosperity for the population. Moreover, because many resources were scarce in relation to the size of the population, China needed to trade with the world. During this period, Deng Xiaoping, the principal architect of reform and opening up, and other reform leaders initiated a new step of socialist development aimed at accelerating industrialisation, technological upgrading, and economic growth. This period is often misunderstood in Western accounts as a simple abandonment of socialism in favour of capitalism or, worse, as the adoption of neoliberalism.¹⁶ The reform era represented an adaptation of socialist strategy to concrete historical conditions. The CPC maintained political control while gradually allowing market mechanisms to operate and private capital to accumulate. This process was monitored and mediated through the structures and institutions of the state built in the first step. The objective was not private accumulation as an end in itself but the rapid expansion of the productive forces to enable the eradication of poverty.¹⁷

Deng Xiaoping famously argued that ‘poverty is not socialism’.¹⁸ The revolutionary state had secured national independence, but socialism also required material abundance. Without economic growth, China would remain vulnerable and incapable of improving the lives of its people. The results of this strategy were extraordinary. Between 1978 and 2019, China experienced one of the fastest economic transformations in human history. Over 765 million people were lifted out of poverty.¹⁹ Manufacturing expanded rapidly, cities modernised, and infrastructure projects transformed the physical landscape of the country. Special Economic Zones such as Shenzhen became laboratories for industrial experimentation. In 1980, Shenzhen was a small fishing town. Within decades, it had become a global technological centre with advanced manufacturing industries. Major firms such as Huawei and Tencent were founded there, and by 2018 its GDP had surpassed Hong Kong’s – an international hub of global capital.²⁰ This transformation reflected the strategic use of foreign capital and technology under state guidance.

Throughout the reform period, the state retained a central role. Unlike in neoliberal economies, where private capital dominates the state and markets incentivise short-term speculative gains, China’s reforms preserved public ownership in strategic sectors such as banking, energy, telecommunications, and transport. This enabled long-term planning and prevented private capital from capturing natural monopolies and extracting rents from them. Instead, the state led a process of market construction in which private capital was directed towards politically determined goals and subjected to intense competitive pressure, forcing firms to upgrade production and expand industrial capacity. Control over the financial system, particularly the allocation of credit, gave the state a key mechanism for steering investment. The market was treated as an instrument rather than as an absolute principle.²¹

This period saw major investment and advances in science and technology. Jiang Zemin, who served as minister of the electronics industry from 1983 to 1985 and later as president from 1993 to 2003, oversaw strategic developments in electronics and information technology that helped lay the foundations for later advances in semiconductors, computing, and artificial intelligence.²² By the time China joined the World Trade Organisation in 2001, its domestic industrial base was dynamic enough to attract significant volumes of foreign capital and its goods were competitive enough to dominate the world market. China had already begun to move from labour-intensive production to increasingly advanced manufacturing and innovation. High-speed rail, renewable energy systems, telecommunications infrastructure, and digital platforms developed on a vast scale. China opened its first fully high-speed passenger line in 2007 and by 2012 had built more than 10,000 km of high-speed rail from virtually nothing – the longest network in the world.²³ By the early twenty-first century, China was no longer simply ‘the factory of the world’ but an emerging technological power.

At the same time, this step generated new contradictions. Rapid growth produced significant regional inequalities, corruption, and disparities between urban and rural areas. Wages did not keep up with rising productivity. Coastal regions advanced more quickly than inland provinces. Private wealth expanded and social stratification became increasingly visible. Environmental degradation also became a serious concern. Yet these contradictions were increasingly recognised by Chinese leaders themselves. Under Hu Jintao, general secretary of the CPC from 2002 to 2012, the party began to respond more directly to the contradictions produced by rapid growth, including labour unrest, regional inequality, and environmental degradation. Minimum-wage standards were raised and labour protections strengthened in the 2000s, while the ecological costs of rapid industrialisation pushed environmental questions higher on the policy agenda. This shift was reflected in the Scientific Outlook on Development, introduced in 2003 and written into the CPC constitution in 2007, which called for people-centred, comprehensive, balanced, and sustainable development.²⁴

The achievements of the second step are undeniable. China became the world’s largest industrial economy, dramatically improved living standards, and built the technological capabilities necessary for sovereign development. No other country has achieved development on such a scale in such a short period. Nor has any country managed to sustain double digit growth rates without eventually entering into a crippling recession. Importantly, China avoided the fate experienced by many post-socialist societies after the collapse of the Soviet Union. Rather than engaging in shock therapy, dismantling the state, and surrendering to oligarchic privatisation, China retained the leading role of the CPC and preserved public ownership of the commanding heights of the economy – banking, energy, telecommunications, and transport. This continuity allowed the country to maintain national coherence and long-term planning.²⁵ Ultimately, the reform period was never presented as the final form of socialism. It was understood as a necessary stage in developing productive capacity.

Step 3: Common Prosperity, Ecological Civilisation, and High-Quality Development (2013–Present)

By 2013, China had largely accomplished the historical tasks associated with developing the productive forces. The country possessed advanced infrastructure, technological sophistication, and substantial national wealth. New questions began to emerge. How should this wealth be distributed and what social purpose should development serve? How can the productive forces and the growth imperative be brought into harmony with ecological limits? What is the place and role of China’s traditional culture in a technologically advanced and atomised world? What is China’s place in a churning world order and what is its relationship to the Global South? These are not easy questions, and none has a clear-cut answer. The third step of Chinese modernisation has sought to address them.

Equality and common prosperity have re-emerged as central socialist principles under Xi Jinping’s leadership. First, the Chinese state has increasingly emphasised the social dimensions of development rather than growth alone. One of the most significant achievements of this step was the campaign to eliminate extreme poverty. Launched in 2013, the Targeted Poverty Alleviation (精准扶贫) campaign sought to identify and address the remaining pockets of extreme poverty that the previous stage of development had not been able to eliminate. Its objectives were summarised as ‘one income, two assurances, and three guarantees’: a minimum income, assurances of adequate food and clothing, and guarantees of compulsory education, basic healthcare, and safe housing. The campaign mobilised more than three million party cadres alongside state institutions and local governments and expanded infrastructure and public services in impoverished rural areas, including access to drinking water and electricity. More than 9.6 million people who lived in areas with harsh environmental conditions were provided with improved housing, infrastructure, public services and access to employment in resettlement communities. In February 2021, the Chinese government announced that it had eradicated absolute poverty in accordance with its multidimensional national standard.²⁶ The campaign formed part of a broader effort to strengthen public services and social protection, including the pension system. Rather than relying solely on market outcomes, the state intervened directly to improve living conditions in the poorest communities, with party cadres playing a leading role and demonstrating the continuing capacity of socialist organisation. This is a reflection of China’s conception of development – one that is centred on collective welfare.

Second, the concept of common prosperity has also entailed measures to prevent extreme concentrations of wealth. Beginning in late 2020, the Chinese government intensified antitrust enforcement across the platform economy, targeting abuses of market dominance, exclusive arrangements, and mergers and acquisitions that had not undergone required government reviews. Alibaba, Tencent, and Bilibili were among the large technology companies investigated or penalised, with Alibaba receiving a record fine of 18.2 billion yuan (approximately US$2.8 billion).²⁷ These measures sought to restrain market concentration and subject the power of large technology companies to greater public regulation.

The government has also imposed sweeping restrictions on private tutoring companies to reduce educational inequality and the financial burden such practices impose on families.²⁸ In response to the emergence of a potential real estate bubble, President Xi Jinping’s slogan ‘houses are for living in, not for speculation’ signalled a shift towards greater regulation of the property market.²⁹ The state has also used its control over the financial system to restrict lending to the property sector and channel credit towards strategic high-technology industries – something the capitalist systems in the Global North have been unable to do since the Third Great Depression (2007–2008). These policies signal the political will and ability to subordinate capital to broader social and developmental goals.

This evolution reflects an important theoretical understanding. Socialism cannot be measured solely by rates of economic growth or technological achievement, however impressive those achievements may be. The decisive indicator is the relationship between productive development and the social good embodied in the concept of common prosperity. Economic modernisation acquires socialist significance only insofar as it expands the material, cultural, and democratic capacities of the people themselves. The continued strengthening of public education, healthcare, infrastructure, scientific research, ecological planning, and cultural renewal therefore represents an effort to ensure that the extraordinary productive advances of previous decades increasingly serve broader social objectives.

Third, environmental sustainability has become a major priority. Earlier stages of industrialisation often relied on environmentally destructive practices. In fact, the offshoring of polluting industries to China meant that the country bore the brunt of the ecological costs associated with consumption in the Global North. In the current step, however, China has invested heavily in renewable energy and electric vehicles (EVs) and is now a global leader in solar and wind power as well as EV production. By mid-2024, China accounted for two-thirds of all utility-scale solar and wind capacity under construction globally. It also produced over 10 million EVs in 2024.³⁰ The shift from the severe urban smog of a decade ago – widely described as an ‘airpocalypse’ – towards the goal of ‘ecological civilisation’ illustrates the scale of this policy change.³¹ The concept of ‘ecological civilisation’ and the slogan that ‘clear waters and green mountains are as valuable as mountains of gold and silver’ have been integrated into official planning discourse as part of socialist modernisation, signalling China’s intent to tackle pollution and climate change.³² This commitment has taken concrete form in large-scale efforts to rehabilitate environments devastated by industrial pollution. One example is the restoration of Erhai Lake in Yunnan Province through the coordination of residents, party cadres, and the local and central government.³³

Fourth, China has sought to move beyond poverty alleviation and narrow the divide between town and country through its rural revitalisation project, such as by transferring technology and resources to rural areas.³⁴ This represents a return to the longstanding Marxist and socialist imperative to abolish the distinction between town and country.³⁵ It also reflects a return to the CPC’s revolutionary roots in the rural masses who provided the backbone of the Chinese Revolution. Rural revitalisation is therefore not only a poverty reduction strategy but an attempt to narrow the urban-rural divide and build forms of development that are less dependent on the consumerism and social atomisation associated with rapid urbanisation. By strengthening rural infrastructure, public services, local industries, and ecological restoration, the project seeks to make the countryside a site of modern socialist development rather than a reservoir of displaced labour for the cities.

During this third step, China’s role in the international community has also evolved. Projects such as the Belt and Road Initiative, which began in 2013, reflect efforts to reshape global development through infrastructure investment and South-South cooperation. The country’s mastery over the productive forces has also made it an essential partner in the industrialisation of Africa. For instance, the Addis Ababa-Djibouti Railway connects landlocked Ethiopia to the Port of Djibouti and supports regional trade, while the Kenya Standard Gauge Railway has significantly reduced transport costs. This stands in contrast to the record of the International Monetary Fund and World Bank, whose policies have perpetuated underdevelopment across the continent.³⁶ While Western critics often interpret these initiatives through the lens of geopolitical rivalry, many governments in Asia, Africa, and Latin America see China as an alternative partner to the unequal structures historically dominated by Western powers. China has also proposed three broad frameworks for international cooperation: the Global Development Initiative, which supports development projects and capacity building; the Global Security Initiative, which sets out principles and areas for cooperative security; and the Global Civilisation Initiative, which promotes dialogue and exchange among civilisations. Together, they represent efforts to offer an alternative to the polarisation of neoliberal globalisation and the hypocrisy of the post-war liberal world order. Their success depends on participating countries of the Global South shaping their priorities and implementation according to their own needs and agendas.³⁷

The significance of China’s experience for the Global South lies not in its replicability but in what it reveals about the structural preconditions for sovereign development.³⁸ China’s model presupposes a unified state with the capacity to direct strategic sectors, plan across generations, and subordinate capital to long-term national priorities – capacities that are not technical but political, grounded in a revolutionary rupture that fundamentally reshaped the relationship between state, class, and market. Where other nations of the Global South have articulated comparable visions – such as Kwame Nkrumah’s argument for a continental African political union in 1963 – efforts to implement them have repeatedly been thwarted by conditions that China has largely overcome: fragmented sovereignty, external debt and conditionality, dependence on commodity exports, and strategic sectors owned and directed by foreign capital. China’s trajectory does not offer a ready-made blueprint. It poses a question: what political and institutional transformations would other countries need in order to exercise similar strategic control over their own development?

Equally significant has been China’s determination to achieve technological self-reliance at a moment when the United States and its allies have intensified attempts to contain its development through export controls, sanctions, restrictions on access to semiconductor technologies, and an expanding military presence across East Asia.³⁹ These measures have only reinforced the party’s conviction that scientific and technological sovereignty has become inseparable from national sovereignty. Investment in advanced manufacturing, artificial intelligence, renewable energy, aerospace, biotechnology, and quantum technologies is therefore understood not simply as economic policy but as an essential component of socialist development under conditions of intensified hyper-imperialist pressure.⁴⁰

The significance of this third step lies in its attempt to synthesise economic modernisation with socialist ethics. China now possesses a substantial productive base. The challenge is to ensure that it continues to transform social relations and serve collective welfare rather than private accumulation alone. This represents a return to foundational socialist concerns under new historical conditions.

The Road Ahead

The socialism of Chinese modernisation cannot be understood as a simple linear progression. It is a historical process shaped by changing material conditions and strategic adaptations. Yet across its different steps, a central continuity remains: the attempt to coordinate the transformation of social relations and the development of the productive forces in order to build a socialist society.

Karl Marx and Friedrich Engels understood socialism as a historical process rather than a ready-made template to be imposed from above.⁴¹ Progress from one stage of development to another requires confronting and resolving contradictions, but this is not a mechanical or economistic process. Nor is it necessarily linear: it may comprise a cycle of probing advances and tactical retreats. It is an experimental process that requires conscious direction guided by theory, tested in practice against concrete conditions, and rooted in the experiences and struggles of workers and peasants.⁴²

Unlike conventional electoral parties, whose activity within bourgeois democracies is largely confined to contesting elections or administering existing institutions, the CPC has consistently understood itself as the organiser of a historical project whose horizon extends across generations. Its legitimacy has rested not only on its ideological commitments but also on its demonstrated capacity to identify changing historical contradictions and to formulate strategies capable of resolving them in ways that improve the lives of the majority of the population.

What steps come next remain to be seen. Within Chinese modernisation, substantive democracy has been pursued through both improvements in living standards and consultative and responsive governance – what the CPC calls ‘whole-process people’s democracy’. This system combines electoral and consultative democracy and extends participation through elections, consultation, decision-making, management, and oversight, including through people’s congresses, political consultative bodies, and grassroots self-governance. Together, these social and political developments form part of the process towards China’s centenary goal of becoming a modern socialist country by 2049.⁴³

Yet the road ahead is not open-ended. At its 19th National Congress in 2017, the CPC set out a two-stage path: to achieve basic socialist modernisation by 2035 and to build China into a modern socialist country by the centenary of the People’s Republic in 2049. This path was reaffirmed at the 20th National Congress in 2022. The 2035 intermediate target is not a vague aspiration but a concrete programme: to substantially raise per capita GDP to that of a mid-level developed country (which would be roughly double the current figure); strengthen self-reliance and capacity in science and technology; guarantee equitable access to basic public services; and ensure that carbon emissions decline steadily.⁴⁴

The 15^th Five-Year Plan (2026–2030), adopted by the National People’s Congress on 12 March 2026, is the first of the two remaining five-year plans leading to the 2035 target. The CPC has described the 2026–2030 period as a ‘crucial link’ on the path to achieving basic socialist modernisation by 2035.⁴⁵ The plan marks a dialectical shift from extensive to intensive development – from growth driven mainly by expanding inputs to growth driven increasingly by productivity, innovation, and efficiency. For much of the reform era, China’s extraordinary growth was driven by the accumulation of inputs: a large supply of relatively low-cost labour moving from the countryside into industry, large-scale fixed investment in infrastructure and manufacturing capacity, and deepening integration into global export markets. That model, which lifted hundreds of millions out of poverty and built the material foundation of contemporary China, has reached its historical limits. Labour costs have risen, many of the country’s major infrastructure networks are now substantially developed, and access to some technologies and markets has been deliberately restricted through US-led controls and supply-chain decoupling. The 15^th Five-Year Plan registers this transition structurally. The headline growth target has been reduced to 4.5–5% annually – below 5% for the first time since 1991 – while research and development spending is mandated to grow at over 7% per year. The signal is explicit: the volume of output is no longer the sole measure of development; its quality and technological content are also crucial. The fact that this lower growth target represents a strategic choice rather than economic failure is itself historically significant. No major developing country has voluntarily subordinated GDP growth targets to quality metrics on this scale.

The central theoretical framework for this transition is the concept of ‘new quality productive forces’, first articulated by Xi Jinping during an inspection visit to Heilongjiang Province in September 2023. It was established as a major policy priority during China’s annual ‘two sessions’ in March 2024 (the meetings of the National People’s Congress, the national legislature, and the Chinese People’s Political Consultative Conference – the country’s highest political advisory body). Applying a qualitative emphasis to Marx’s category of the productive forces, the concept refers to advanced productivity freed from traditional growth patterns and characterised by high technology, high efficiency, and high quality. In practice, this means fostering a new generation of strategic industries – artificial intelligence, biotechnology, advanced manufacturing, new materials, quantum technology, hydrogen energy, and aerospace – while simultaneously upgrading traditional industries through digitalisation and green transformation. Under the 15^th Five-Year Plan, the digital economy’s share of GDP is targeted to rise from 10.5% in 2025 to 12.5% by 2030, while grain production capacity is set to reach approximately 725 million tonnes, comprehensive energy production capacity is set to increase by 13%, and non-fossil energy’s share of consumption is set to rise from 21.7% to 25%.⁴⁶

A crucial dialectical reversal is at work in China’s response to US-led technology containment. The export controls, semiconductor restrictions, and supply-chain pressures that Washington has imposed since 2018 and tightened in 2025 were designed to arrest China’s technological advance. Yet their effect has been the opposite: by closing off access to foreign technology at the frontier, the United States has compelled China to invest with extraordinary urgency in precisely the capabilities it sought to deny the country. The 15^th Five-Year Plan’s target of at least 7% annual average growth in R&D spending and its emphasis on achieving breakthroughs in semiconductors, industrial software, and advanced materials can therefore be understood in part as a domestic productive response to external containment. External pressure has been internalised as a driver of self-reliance, transforming an obstacle into an accelerant.

At the same time, the plan attempts to resolve a structural contradiction that has accumulated across the reform era: the disproportion between China’s extraordinary productive capacity and the relative weakness of domestic consumption as a driver of growth. The ‘dual circulation’ strategy, first proposed in 2020 and incorporated into the 14^th Five-Year Plan (2021–2025), takes the domestic market as the mainstay of the economy while seeking to ensure that domestic and international circulation reinforce one another. This is not merely a technical adjustment. It reorients the relationship between production and social reproduction – the conditions and services that sustain workers and households – as well as the relationship between the national economy and the world market. Where the export-led model of the second step required China to insert itself into the global division of labour on terms partly set externally, the dual circulation model seeks to make domestic demand the primary driver of economic growth. Rising wages, expanded public services, and a broader middle-income stratum are conditions both for stronger domestic demand and for common prosperity. In this way, the strategy links the distributional and macroeconomic dimensions of socialist development.

At a time when much of the Global South seeks new paths of industrialisation, technological development, and economic sovereignty, China’s experience has reopened questions that many believed had been settled forever after the destruction of the Soviet Union.⁴⁷ It has shown that planning remains possible, public investment can transform societies, technological leadership need not remain the monopoly of the Global North, and the future need not belong exclusively to finance capital.⁴⁸

Chinese modernisation represents a distinct path in world history. It rejects both colonial dependency and neoliberal orthodoxy. It is a socialist project that walks on two legs: transforming social relations and developing the productive forces. This project has progressed through different historical steps while maintaining the central objective of national sovereignty and human dignity. Whatever contradictions remain, the Chinese Revolution stands as one of the most consequential emancipatory processes of the modern age.

Notes

A skill is a reusable workflow for an AI agent. Its main SKILL.md file explains when the skill should be used and what the agent should do. A skill may also contain scripts, templates, and other reference materials. The agent may select a skill automatically when a task matches its description, or you can invoke it directly.

This guide demonstrates how to find, review, install, and use skills with an AI-agent extension in VS Code. If you have not yet set up this workspace, follow Set Up Your AI Workspace (macOS) or Set Up Your AI Workspace (Windows) before continuing.

Although interfaces and installation methods may vary, the same basic workflow also applies to other AI agents that support skills.

Installing Skills in VS Code

Installing a Skill through a Skills Platform

Numerous platforms host skills, so searching for and downloading them is usually straightforward. This section uses https://www.skills.sh/ as an example. The platform provides a skill called find-skills that makes searching more convenient, and its official installation method is to run a specific prompt.

Copy the prompt below into the chat box to install find-skills.

Run `npx skills add https://github.com/vercel-labs/skills -g --skill find-skills` and follow the generated skill instructions now. Read its complete output, redirecting it to a temporary file first if necessary. Resolve relative paths from the supporting-files directory it provides.

Once installed, you can search for skills through the agent. Ordinary wording still works, but commands such as /find-skills are recommended because they are clearer, faster, and more cost-effective: the model can read the specified skill directly. For example:

/find-skills please help me find a skill that can help summarise papers and books.

The agent uses the skill and returns a large number of matching results. With so many options, it is not always obvious which one to choose. Useful criteria include installation counts, star ratings, and the description.

Consider the two skills shown below. Both are hosted on GitHub, and the repositories associated with them have attracted substantial attention. The find-skills results also show that the skills have been installed many times.

Star and installation counts can help narrow a long list of options, but they are indicators of popularity rather than guarantees of quality or safety. Before choosing a skill, check whether its description matches your task, review its SKILL.md file, and examine any scripts or permissions it relies on.

Next, examine the description. You can click the link to learn more about the skill. Read the article that begins with SKILL.md; this is the file mentioned earlier, and it is what enhances your prompts.

The first skill is used to search for and read articles, as well as to write a literature review. The second establishes a primary question and traces claims back to their sources.

After comparing them, we choose the first one, systematic-literature-review, as our example. The skill can both conduct systematic literature reviews using online sources and summarise papers that you provide.

On the published page, click Prompt, copy the command, and paste it into the chat box to install the skill.

If that feels cumbersome, you can also copy the skill’s name or other identifying information and simply tell AI, ‘I want to install this skill.’

Installing a Skill from a Link or Description

Sometimes others may share a direct link to a skill, such as https://github.com/bytedance/deer-flow/blob/main/skills/public/systematic-literature-review/SKILL.md, or merely give an ambiguous reference such as ‘bytedance’s systematic-literature-review skill’. You may not know where it resides or how to install it.

Thanks to the agent’s search capabilities, you only need to provide as much information as possible and let AI handle the installation. For example:

Please help me install a skill. Here is the information:

bytedance's systematic-literature-review skill

Installing a Skill from a Package

Sometimes you will receive a skill package from someone else. For example, the files may end with extensions such as .zip, .rar, or .7z. We also recommend using the agent to install the skill so that the required environment is configured correctly. For example:

Please help me install the following skill and configure the required environment: @systematic-literature-review.zip

Choosing Where to Install a Skill

Many compatible agents can automatically load a skill when a task matches its description, although the exact triggering behaviour varies between agents. This mechanism can substantially change the prompt context that the model ultimately receives. For example, a user may submit a simple request such as:

Please help me make a literature review for the topic of
"Social Science: Reclaiming Research Sovereignty in the Age of Artificial Intelligence."

After skills are activated, however, the effective prompt may be augmented with one or more skill definitions before the original user input, allowing agents to determine which skill to use:

User's computer has the following skills available:

name: systematic-literature-review

description: Use this skill when the user wants a systematic literature review, survey, ...

name: xxx

description: xxxxx

...

User's input:

Please help me make a literature review for the topic of
"Social Science: Reclaiming Research Sovereignty in the Age of Artificial Intelligence."

If several installed skills cover similar tasks, AI may be unclear about which one should run. Expanding the skill set can also increase the prompt context, making it more expensive and leaving AI with less capacity to process other information. In conclusion, installing too many overlapping skills is not cost-effective.

We therefore recommend project-level installation for skills related to a particular body of work. This means that AI can only see the skill within the relevant workspace. If you open another folder, skills installed in the original folder cannot be seen. Only skills that are genuinely useful across unrelated projects should be installed globally.

It is easy to ask an agent to install a skill, but you should state the intended scope explicitly.

Please help me install a skill **as a project skill**. Here is the information:

bytedance's systematic-literature-review skill

After installation, we can find this skill in the Explorer. Its path is .claude/skills/systematic-literature-review. Agents can recognise skills in folders of this type, although some may use .agent/skills/ instead, and can load them automatically.

After learning the location of the project skill, you may wonder where global skills are stored. Open Finder and click Go -> Home.

If you’re using Windows, C:\Users\[your user name] is your home directory.

If you cannot see any folder names beginning with ., you can use Command + Shift + . to make hidden files visible.

Compare the contents of .claude in the Home directory and the project directory. Although .claude contains more files, the structures are similar. The key point to remember is that skills installed in the Home directory can be seen everywhere, whereas project skills can only be seen in the current folder.

Using Skills in VS Code

Before using the skill, start a new session to refresh the skill list. Otherwise, the skill may not appear in the extension.

In the new session, type the first few letters of the skill’s name. When it appears, press Tab to complete the command.

We use the article ‘AI for Social Science: Reclaiming Research Sovereignty in the Age of Artificial Intelligence’ as an example. Type the following:

/systematic-literature-review Please summarise this article @inputs/AI for Social Science-Reclaiming Research Sovereignty in the Age of Artificial Intelligence.md

Then, regenerate a summary without the skill as a point of contrast. Compared with the version produced without the skill, the first summary is more concrete and operational because it explains how AI4SS framework is implemented through technical architecture, skills, agents, workflows, and specific use cases.

It also provides stronger evidence of feasibility and practical impact, rather than remaining mainly at the conceptual or theoretical level.

We can also test its basic ability to find information online and generate a literature review on the topic.

/systematic-literature-review Please help me make a literature review for the topic of Social Science-Reclaiming Research Sovereignty in the Age of Artificial Intelligence

After generation, it returns a link named slr-ai-social-science-20250902.md. Use Command + Click to open the file.

It collects articles online and writes a literature review with references, which can be helpful for quick research.

Creating Your Own Skill

If you have crafted a useful prompt or completed a difficult task and wish to save or share it with others, creating a skill yourself is an excellent choice. Although it may sound daunting, a skill is essentially a set of files that records your experience of working with AI.

Creating a Skill through a Prompt

As mentioned earlier, a skill can be as simple as a short prompt. This is the simplest form of a skill, and it can be saved and used in your agent with ease.

Here is a prompt that we used in the previous article to summarise an article. It provides a useful example.

You can simply type the prompt below and upload it to the agent. If you want to define the name of the skill, you can add the name you prefer, which can make the skill easier to trigger.

Please convert this file into a skill and install it in this project. @inputs/summarise-papers-revised.md

If you do not assign a name to the skill, AI will create one for you. In this case, we can trigger the skill by typing /summarise-paper. A skill’s name is often the same as the name of its folder, although this is not always the case. The best way to determine the trigger word is to open the skill and inspect its header.

Use Command + Click to open the file.

The name written in the header is the trigger word for this skill.

Start a new session and we can find this skill.

Creating a Skill through a Task

However, we cannot always start with a completed prompt file. A more common situation is that, after completing a difficult task, we want to record what worked in order to avoid failure and save the workflow for future use. This can save both time and tokens when a similar task needs to be completed later.

For example, suppose you want to understand the news that happened today. After several turns of conversation, you have completed the task in a way that meets your needs perfectly. You certainly do not want to repeat the entire conversation the next day. In this situation, you can save the conversation history as a skill.

For example, you can type:

Please convert the conversation history into a reusable skill and install it in the current project. The skill should be triggered by `/news-collector`.

Because we have assigned the name news-collector to the skill, there is no need to open the saved file again. Simply use /news-collector to collect the news the next day.

Sharing Your Own Skill

After you have created an excellent skill, you may face the question of how to share it with others. The easiest approach is to package it and transfer it as a single file. For example:

Please package the skill below and provide me with a shareable package that is easy to transfer:

/systematic-literature-review

AI will provide you with a .zip file. It contains everything the skill needs, so on a new computer, you can simply tell AI, ‘Please install this skill through systematic-literature-review.zip.’

Advanced: Understanding the Structure of a Skill

Normally, there is no need to decompress systematic-literature-review.zip. Treat it as an installer for the skill, executed by the agent.

Nevertheless, let us unpack it in this section to understand how a skill works internally. Inside the folder, we can see systematic-literature-review.md, alongside two subfolders named scripts and templates.

The file systematic-literature-review.md is the primary document AI encounters first. It contains the conditions that trigger the skill and outlines the main workflow.

Sometimes, to make a skill easier to maintain, its creator places certain resources in a subfolder because a very long file can be difficult to search and read. For example, this skill stores different citation formats in a folder named templates. When AI needs to reference citation rules, it can look up the relevant file.

Some tasks are more easily accomplished by a program. The skill records how to use the program so that AI knows how to invoke it.

In summary, a skill decreases uncertainty through a set of prior prompts and programs. If your skills do not perform well, you can open the folder, review the workflow, and then ask AI to modify it. Skills often perform better when you take an active role in controlling the workflow.

Treat Unknown Skills with Caution

Skills have spread quickly. They are more flexible than ordinary programmes, and they can be difficult for traditional antivirus software to inspect. Therefore, install them with caution. If your computer suddenly becomes slower, grows unusually hot, or begins producing strange files, investigate immediately.

You can also ask AI to inspect and protect your computer at regular intervals. Skills are not the only risk; many newer attack methods exist, so anything unfamiliar should be treated with caution.

Please conduct a comprehensive security audit of this computer, with a particular focus on detecting malware, trojans, backdoors, persistence mechanisms, suspicious scripts, abnormal network connections, or unauthorised remote access.

Useful Skills in This Series

Several Bandung Circuits how-to guides introduce skills that can be installed in an agent and reused across research tasks. The table below lists these skills and the guides in which they appear.

These skills cover a large part of a research workflow, from reading and critique to figures, slides, and audio transcription. They represent a practical starting set rather than a complete catalogue. When a new task appears, /find-skills remains the most useful way to look for alternatives.

Conclusion

Skills have already shifted everyday prompting from ‘typing a request from scratch each time’ towards a reusable workflow in which you install a set of instructions, call it with a slash command, and let the model follow the prescribed procedure. For researchers and other users of VS Code, this means that summarising a paper, generating slides, or transcribing a recording no longer has to begin with a blank chat box.

An efficient approach is to search first with find-skills, then select skills with high star ratings and installation counts, read the SKILL.md file, and only then install and invoke them. After installation, start a new session so that the skill appears in the list. Used in this way, skills can save time and make the output more concrete, as the literature-review example in this chapter demonstrates.

The goal is not to collect as many skills as possible, but to build a small set of trusted workflows for the tasks you perform regularly.

The Red Army crossed these mountains in 1934, on the march that would take it north. Football arrived a decade later, brought by students of Guangxi University, evacuated to Rongjiang to escape Japanese aggression. Villagers have organised their own matches nearly every year in the eight decades since.

In May 2023, that tradition found a new expression in Cun Chao, the Village Super League: a competition organised by villagers, played by villagers, and watched by villagers. Within weeks of its launch, the stadium’s stands were filled to the brim, and footage carried by the short-video platforms drew the eyes of the whole country and then of the world. None of it ran itself. Packing a stadium every weekend, keeping order, feeding the broadcast — all of that took an organisational capacity the county had not required before.

This organisational capacity became an engine for real economic outcomes. Tourism revenue rose from 8.4 billion yuan in 2023 to 10.8 billion in 2024, and the county’s gross domestic product reached 10.496 billion yuan in 2024, a rise of 9.5 per cent. The league brought more than 2,700 new market entities into being, over thirty per cent of the whole prefecture’s new total, and village collective income climbed from 58.9 million yuan in 2022 to 121 million in 2024.

It also opened a door to something else entirely. Since 2023 the county has set itself a less likely task: to put artificial intelligence (AI) into the hands of the villagers. It has built no large language model, no data centre, and no semiconductor plant. What it has built instead is its own people, organised across government offices, villages, and enterprises, so that residents who had never written a line of code, among them elderly farmers and village cadres, now use AI in their everyday work.

Building the Institutional and Material Conditions

The experiment was made possible by conditions built at multiple levels.

Cun Chao was no technology project but a feat of organisation, and the organisational structure was standing before the first ball was kicked. In October 2021, a year and a half before that match, the county set up a leading small group for new media and rural e-commerce, headed jointly by the county Party secretary and the county governor. In a Chinese county the Party committee leads and the government departments administer, and those departments — agriculture, commerce, culture and tourism — each holds responsibility for part of any new undertaking; a leading small group is how a Party committee concentrates the efforts of different departments on a common task. Here the two highest officials in the county took the headship themselves, one from the Party and one from the government, which put new media at the top of the agenda.

Beneath the group the county placed a working team at the county level, a new-media service centre in every township, and a new-media service station in every village, and it set up a new-media company besides, a county state enterprise holding shares with private partners. The result was a single organisational chain linking three different but complementary capacities: Party leadership, grassroots mobilisation, and market activity. The Party committee provided political leadership; grassroots cadres reached every household; private firms and traders connected production to the market. When Cun Chao began in May 2023, the county set a dedicated Cun Chao Office on top of that standing structure, to mobilise cadres, coordinate security, and keep the media channels fed across all three levels at once.

When the county later turned toward AI, the mechanism had already put thirty-five thousand person-times — a tally of training instances, not unique individuals — through new-media training and built more than two thousand local livestream teams and over twelve thousand new-media accounts.

Rongjiang calls the shift the ‘Three New Rurals’: let the phone become the new farm tool, data become the new farm input, and livestreaming become the new farm work. As Party secretary Xu Bo puts it, the aim is to turn Cun Chao into new quality productive forces for local industry. New technology only raises the productive forces once the relations of production change to accommodate it — technology alone doesn’t do it; the social arrangements have to shift first to let it work. Led by the Party committee and carried by the villagers themselves, the county has therefore pursued this transformation by reshaping who is trained to use the tools, who holds the account they feed, and who keeps what a sale earns.

The material foundations for all this had been laid at national level. Through the poverty alleviation campaign the state carried roads, power, and broadband into places the market had never reached, in order to connect them to it: by the end of 2020, every eligible township and administrative village in the poor counties had a paved road, a bus service, and a postal route; reliable electricity, fibre broadband, and 4G had been extended across the countryside, with fibre and 4G covering more than 98 per cent of poor villages; and e-commerce services reached every poverty county. Only once these material conditions existed could digital transformation become a realistic strategy for rural development, e-commerce first and AI after it. One of the instruments that carried this into counties like Rongjiang is paired assistance, which binds a coastal province or work unit to an inland county and moves people, money, and technology along the link. It dates from a central decision of 1979, at the start of reform and opening up, following Deng Xiaoping’s strategy of letting some regions grow rich first and then bring the rest along. Rongjiang later applied the same logic to AI training, pairing those with greater digital capacity with those who had less experience.

Once these material foundations had been laid, AI became the next stage of national development strategy. This transition built on an earlier digital rural agenda: the central Digital Village Development Outline (16 May 2019) had already named ‘the improvement of peasants’ modern information skills’ as a strategic direction. On 26 August 2025, the State Council issued its Opinions on Deeply Implementing the ‘AI Plus’ Action, proposing six lines of action across science, industry, consumption, people’s wellbeing, governance, and global cooperation, and undertaking to strengthen the unified planning of intelligent computing power. Relying on the variable of the people’s adoption of digital tools, rather than the capital or hardware more often treated as decisive, the county organised its people to take up that opening on its own terms. It translated these national priorities into a local mobilisation plan, set out in a mnemonic it calls the ‘1-2-3-4-5’ target system, which illustrates how a local state translates a national technology strategy into mass participation:

• One benchmark. By the end of 2027, to have built Rongjiang into a national benchmark county for the universal learning and application of AI.
• Two pillars of support. A corps of three hundred home-grown technical backbones and thirty thousand ‘application pacesetters’, cultivated across at least five concrete AI application scenarios. The three hundred are not an elite but come from across the county’s cadres, civil servants, teachers, enterprise staff, village cadres, and returning graduates. They are seeded rather than bought: each of the three hundred trains ten, each ten a hundred, until the thirty thousand are reached.
• Three full coverages. One hundred per cent training coverage of government offices; one hundred per cent basic-AI-literacy coverage of every administrative village; one hundred per cent scenario-application coverage across every sector.
• Four-tier linkage. A four-tier AI application and promotion network running from county to township to village to enterprise.
• Five categories of benefit. Measurable gains in the efficiency of government services, in tourist satisfaction, in the coverage of quality educational resources, in the accuracy of medical diagnosis, and in the reduction of enterprise costs, summed up in the slogan: every village has an AI hand; every trade has an AI exemplar; every task has an AI enablement.

AI on Four Fronts of Rural Revitalisation

The county puts AI at the service of its people, their work, and their days; it turns it to the plain business of rural revival on four fronts.

1. AI Lowers the Cost of Being Seen

Before AI and new media reached the countryside, a village’s produce rarely left the mountains, for the only real route to market was to carry the harvest down to the nearest town and sell it there. Farmers had neither the capital nor the skill to present their goods attractively or to advertise them further afield, and a competent poster or promotional video meant hiring a professional crew few could bear. To make AI accessible beyond specialists, the county first trained villagers in practical AI-based content production, inviting outside professionals to teach simple workflows using mainstream AI tools and organising county-wide competitions to encourage experimentation. AI lowers the cost of being seen. With little more than a smartphone, villagers can now generate and refine the promotional material themselves – a poster, a short video, an animation – and distribute it on the same platforms that carried the league. Making a short film to promote Cun Chao, until lately beyond an average villager’s means, is now affordable and easily learned – and that same footage now carries other villagers’ goods to market too.

Producing enough good video by hand, across thousands of accounts, turned out to be slow and costly, and the traffic it won came unevenly; so the county integrated AI editing and large-scale distribution into its new-media system, operated by the villagers it trains as the league’s new-media promoters, allowing them to produce and circulate promotional content at a speed and volume the old way could not match. When a match is on, they film it and upload the footage, and the system writes the captions and edits it into a finished clip, ready in under half an hour. Distribution is made as simple as production: a tap on the on-site posting device, or a scan of a QR code, puts the clip on Douyin, China’s TikTok, within seconds. Promotional videos now carry built-in distribution links to local products, so that the same circulation that spreads the county’s stories also generates sales and commissions for those who share them.

2. AI Eases the Grassroots Cadre’s Burden

The chronic burdens of the grassroots cadre – the drafting of reports and notices, the compiling of statistical tables, weak policy publicity, slow response to grievances, the sheer weight of administrative toil – are eased by pressing AI into the workflow itself. It drafts the reports and assembles the meeting records, and, more valuable still, it recasts the official register of policy into plain language villagers can follow, which makes policy far easier to communicate. The county reports that a cadre who once spent half a day on a draft now finishes in half an hour, left to localise and refine rather than to assemble from scratch. The figures from Zhongcheng Town are plain: in the first two and a half months of 2025 alone, its village-level monitoring network flagged disputes early enough for staff to resolve sixteen of them before they escalated, AI-assisted forecasting fed into plans for nineteen projects, and cadres saved roughly two working days each. The county audit office, too, runs its own learning routine on cheap, off-the-shelf domestic models.

3. AI Turns Local Talent Into a Reason to Return

Where rural policy typically laments the outflow of the young, the programme attempts to reverse the valuation, so that mastery of AI, and the short films and animations that draw hundreds of thousands of views, become an attractive and remunerative occupation worth returning home for and worth staying for, aimed at returning youth, the women who remain in the villages, and individual traders. Liu Qinlan is the county’s own example: a kindergarten teacher earning a little over 2,000 yuan a month, she and her husband gave up their city jobs and came home when Cun Chao went viral in May 2023. She started out trying to sell the county’s own produce — luohan fruit, green and white tea, pickled fish and meat — and learned filming, editing, and livestreaming from scratch through the county’s training. Known online as Cun Chao‘s ‘Miao sister Lan Lan’, she now runs a batik studio by the Cun Chao ground, drawing on the region’s indigenous textile traditions and giving more than 180 embroiderers and dyers work without leaving their villages. In 2025 her batik and embroidery lines sold more than 2 million yuan.

4. AI Carries Local Culture Out of the Mountains

Cun Chao draws on its own peasant, working-class, and ethnic cultures. Made up of vendors, farmers, tilers, butchers, and factory workers, its villages name their teams after what they grow or do – a Monk Fruit team, a Passionfruit team, a Bayberry team, a Homestay team, a Rafting team, a Rice-Noodle team, even one that calls itself simply ‘the commoners’. The side that wins takes home not cash but the county’s own produce: the 2026 champions carried off a Guanling cow, the runners-up a Guizhou sturgeon, third place a crate of local duck. More than eighty per cent of Rongjiang’s people belong to the ethnic minorities of Miao, Dong, Shui, and Yao, home to heritage forms such as the Grand Song of the Dong (choral singing), traditional indigo dyeing, and the drum-tower. But these traditions – peasant, working-class, ethnic minority, even the Cun Chao festival itself – survive only insofar as they are seen and used. As such, AI animation, short video, and digital-human generation are turned to the work of carrying all of it – produce, trade, and heritage alike – out of the mountains and into national circulation, and from there into the commercial logic of cultural-tourism branding. At the provincial level this is already a working layer: a Cun Chao digital-human and smart-companion platform provides AI tour-companion services keyed to the league and to the Grand Song of the Dong, while the county’s enterprise training uses the Cun Chao brand, special crops, and heritage as its worked teaching cases.

Tailored Mass Training

The Rongjiang experiment rests on its teaching, the means by which an unfamiliar tool reaches farmers, shopkeepers, and cadres with little formal training. Its logic holds throughout: teach each person only what their own work requires, keep every lesson pared down to what can actually be done, and let those who have learned teach those who have not.

Nothing is taught one-size-fits-all. The county sorts the people it trains into groups, each with its own syllabus: for cadres those of governance, for merchants those of income, for the young and the heritage-keepers those of cultural creation. The same care goes into matching the task to the person, so that no one, the county insists, is made to learn what they cannot use. The county teaches AI as a practical skill rather than a technical discipline. Algorithms, programming, and the inner workings of the models are left aside in favour of using domestic AI applications and writing simple prompts. A visiting expert may bring a thousand possible applications of AI, but the local trainer’s task is to reduce them to the two or three that match local needs and market demand. The county therefore standardises its training around three practical courses — in AI document drafting, AI short-video production, and AI visual creation. The same philosophy shaped one of the county’s most recent courses, on AI comic drama in June 2026, for sixty returning youth, heritage artisans, and small entrepreneurs: it was designed for complete beginners, emphasised hands-on production, and connected local cultural heritage directly to market-oriented digital content creation.

Learning is treated as a rhythm rather than an event. Every Friday evening a trainer takes a livestreamed shift, a class that may draw two or three hundred viewers or, some nights, only one. Even one is considered worthwhile, because one person helped is one more person empowered. Since 2025, the weekly stream has not lapsed.

The teaching applies the same logic of paired assistance: those who learn first bring the rest along, with young cadres matched to middle-aged and senior staff to close the digital gap. The records bear it out. Zhongcheng Town designated twelve technology-promotion officers and established forty-two mentoring pairs, and a single round of centralised training covered ninety-eight cadres. At Bakai Town more than sixty cadres trained on four working scenarios, smart document handling, population-data modelling, new-media outreach, and emergency-response dispatch, with those trained to teach the rest, so that village cadres could run the systems on their own.

If pairing carries the skill from one person to the next, a chain of four tiers carries the plan itself from the county down to each level, with county, township, village, and enterprise each holding a distinct job.

1. The county designs. It sets the plan, marshals the teachers, builds the demonstration cases, and drives the work through weekly scheduling meetings that hold the lower tiers to account.
2. The township executes. It runs the routine mass training and the cadre drills.
3. The village is the front line. It is where villagers are taught and the locally useful applications are found.
4. The enterprise turns capability into revenue. At the end of the chain, it welds AI to culture-tourism, agriculture, and handicraft.

It is this chain, as much as the curriculum, that turns a directive issued in the county seat into a lesson under way in a village.

All of this rides on a project discipline recognisably that of a campaigning state, set out as a five-step closed loop for landing the programme in any village or unit.

1. Survey and preparation. An inventory of local industry, governance pain points, population structure, and the digital baseline yields a bespoke plan, using the ‘brainstorming meeting’ — the county’s instrument, inherited from the league campaigns — to identify concerns and potential challenges before they become real obstacles to implementation.
2. Cultivation of the backbone. The core group of technical backbones, the same three hundred named in the ‘1-2-3-4-5’ plan, are trained first.
3. Universal training by the layered method. That core group then cascades the training outward, each backbone teaching the next tier down, until it reaches the wider population.
4. Scenario landing. Training and application proceed together, so that each lesson attaches to a real task rather than staying abstract.
5. Review and quality-raising. What worked and what did not is assessed, and the next round of the loop is adjusted accordingly.

The Long March of the AI Era

Much of the Global South watches the spread of generative AI with unease: that it will deepen the dependence of poorer countries on a handful of firms in the wealthy core, that it will drive up the carbon cost of computation, that it will hasten the displacement of labour. None of these worries is misplaced. But fear on its own settles nothing, and the harder question is what a place without capital, advanced hardware, or a model of its own can actually do with the technology as it stands.

Rongjiang’s answer begins with cost. The computation does not run on the user’s device. The heavy capital sits in the cloud, paid for by firms such as Doubao and DeepSeek that built the models and by the infrastructure of states far from Moon Mountain, and what the villager needs is only an ordinary smartphone and the knowledge of how to speak to the model.

That the state should carry the compute is a matter of policy, and the ‘AI Plus’ Action names ‘strengthening the unified planning of intelligent computing power’ among its basic supports. The base model, moreover, has become a commodity: each is substitutable for the next, and the rural user is bound to no single vendor, so that should one tool withdraw or deteriorate, another takes its place. Substitutability leaves the monopoly rent nowhere to lodge, and what remains is an ordinary competitive cost — the production diffusion that China’s own development path has set against the rent-extraction of technological monopoly.

Even a smartphone, though, cannot everywhere be taken for granted. Rongjiang has met that difficulty directly. When it first carried new media into the villages, many people’s handsets were too old to run the apps. Its answer was not to wait until everyone was equipped, but to begin with the willing minority who already were, let them earn a living from it, and let their example draw the rest in.

Returning youth who used AI and short video to sell heritage crafts that had never before found a buyer, some of them earning a considerable income, pulled in the older women who made the goods. As those women began to earn, they bought phones of their own and began to film. The circle widens with the income.

What, then, could travel? Not the machine, since Rongjiang owns none of it, but the method: a way of organising the people around new forces of production. The institutional vehicle already exists. On 10 November 2025, the county and East China Normal University opened the South School in Toutang village, in Guzhou town, as a platform for the international communication of poverty alleviation and rural revitalisation. It opened into a wider moment: the 2025 Global South Academic Forum (Shanghai, 13–14 November 2025), co-hosted by Tricontinental: Institute for Social Research, ran a panel on digital sovereignty and AI in the Global South, and drew extensive involvement from organisations such as Brazil’s Landless Workers’ Movement (MST). The method has begun to circulate where movements, and not only governments, can take it up.

The Global South’s fight for digital sovereignty is usually pictured as building its own stack – its own chips, its own models, its own data centres. For most of the periphery, that whole edifice is out of reach at the moment, and a sovereignty of that kind, if it comes at all, will be won collectively rather than by any poor country alone.

Rongjiang works a different layer of the same struggle. Ownership of the machinery is not something one county can secure on its own; what it can secure, at its own scale, is the power of its people to use the tool on their own terms and for their own ends, and to keep the value they create rather than let it drain upward to a handful of firms in the North. That – a people able to use the tool, and to hold on to what it earns – is a sovereignty a poor county can actually build.

Ninety years ago, the Chinese Red Army carried the Long March to victory, won not by superior arms but by organised people, their endurance, and the peasants it mobilised as it passed, rousing their energy, their initiative, their creativity. Mao Zedong called that march a seeding-machine, for the revolution it sowed along the way.

Rongjiang has set out on a long march of its own, one for the AI era, and it moves in the same spirit: carried by the people, drawing out that same creativity, sowing among them a new capability as that march sowed revolution, and letting the gains of the technology flow back to them, to the many and not the few. It is not a slogan handed down but the patient work of turning a new productive force into a capacity held by the people, in the tradition of mass work. The technology will keep changing, and fast. What does not change is the insight at the centre: for a place that cannot build the stack itself, the question that remains is who, among its people, has made the model their own.

Havana, August 2026

From 6 to 9 August, more than 180 delegates from 27 countries representing more than 130 organisations gathered at the University of Havana for the IV Continental Assembly of ALBA Movimientos, a space for articulating popular movements around the project of socialism, anti-imperialism, and continental integration. As part of the assembly’s cultural programme, FIDEL x 100 commemorated the centenary of Fidel Castro’s birth with more than 200 posters created by over 100 cultural workers from 32 countries.

Left: Abhinav (India, Young Socialist Artists/Tricontinental); right: Daniela Ruggeri (Argentina, Tricontinental).

The exhibition emerged from an open call organised by the International Peoples’ Assembly, ALBA Movimientos, Utopix, Young Socialist Artists, Pan Africanism Today, and Tricontinental: Institute for Social Research. The diversity of the works – in style, technique, language, and political tradition – reflected the international reach of Fidel’s legacy. Yet the exhibition was not simply an act of commemoration. It was a recovery of Fidel’s ideas as tools of struggle – the ideas that helped make the Revolution possible and continue to sustain it today.

The full archive is available as a free downloadable collection to be printed, exhibited, studied, and used by organisations around the world. Like the posters that decades earlier travelled folded inside Tricontinental, the magazine founded by the Organisation of Solidarity with the Peoples of Asia, Africa, and Latin America (OSPAAAL), these works are intended to circulate: from assemblies to political schools, from union offices to classrooms, from movement headquarters to neighbourhood walls. We invite you to use them in this way.

FIDEL x 100 belongs to a much longer history. The Cuban Revolution did not treat culture as something secondary to politics, nor as mere decoration for revolutionary achievement. It understood cultural work as part of the material infrastructure of revolution itself: a means of forming political consciousness, cultivating technical and artistic capacities, sustaining collective memory, and extending international solidarity beyond the island’s shores.

Havana, Sixty Years After the Tricontinental Conference

That tradition acquired one of its clearest institutional expressions in January 1966, when Havana hosted the first Tricontinental Conference. More than 500 delegates from liberation movements, revolutionary governments, communist parties, and popular organisations across Africa, Asia, and Latin America gathered to forge a common anti-imperialist strategy against colonialism, capitalism, and war.

Left: Ian Matchett (US, Swords into Plowshares Peace Center and Gallery); right: Ignacio Andrés Pardo Vásquez (Chile, Utopix and Quinmantú).

Out of the conference emerged OSPAAAL, one of the most important institutions of revolutionary internationalism in the twentieth century. Through its publications, solidarity campaigns, educational work, and especially its extraordinary poster production, OSPAAAL helped build a shared political language for struggles taking place across three continents.

Over three decades, the organisation produced around nine million posters distributed to more than sixty countries, many folded inside the pages of Tricontinental. The magazine and posters travelled to places where revolutionary diplomats could not, making the struggles of Vietnam, Palestine, Angola, Mozambique, Puerto Rico, South Africa, Chile, and dozens of others visible to each other as part of a common front against imperialism.

Many of the graphic designers who created these posters had learned their craft in Havana’s advertising agencies before the Revolution. After 1959, they turned those same techniques against empire. Typography, photography, colour, montage, among other visual techniques – once employed to sell commodities – were repurposed to build international solidarity, make anti-imperialist struggles legible across languages, and forge a revolutionary visual language capable of crossing borders.

As the Cuban designer Félix Beltrán later reflected, ‘The poster could circulate in countries where a functionary was not allowed to speak about the ideas of the Revolution’.

The poster thus became more than a medium of communication. Under conditions of blockade and diplomatic isolation, it became an instrument of international organisation. Alongside the revolutionary cinema of the Cuban Institute of Cinematographic Art and Industry (Instituto Cubano del Arte e Industria Cinematográficos, ICAIC), the editorial work of Casa de las Américas, and the international educational institutions built by the Revolution over subsequent decades, OSPAAAL formed part of the cultural infrastructure through which Cuba internationalised its revolutionary project.

A Revolution Can Only Be the Child of Culture and Ideas

By the end of the twentieth century, the Cuban Revolution had survived what many believed would be its final test. The fall of the Soviet Union plunged the island into the Special Period, while Washington tightened the blockade expecting that socialism would finally collapse under economic isolation. Across much of the world, the triumph of neoliberalism was proclaimed as the ‘end of history’, as political scientist Francis Fukuyama famously declared, and Cuba was widely treated as a relic whose disappearance was only a matter of time.

It was during this conjuncture, speaking at the Aula Magna (Great Hall) of the Central University of Venezuela on 3 February 1999, that Fidel returned to a principle that had occupied the Revolution since its earliest years, ‘Una revolución solo puede ser hija de la cultura y las ideas’ (A revolution can only be the child of culture and ideas).

Left: Ingrid Neves (Brazil, Tricontinental); right: Isabela Molin (Brazil, MST).

This was not a defence of culture in the abstract, nor a call to privilege the battle of ideas over material transformation. It was a reaffirmation that revolutions do not simply inherit revolutionary human beings, including the cultural and intellectual workers needed to sustain them; they must help create them. Political consciousness, technical capacity, collective discipline, and internationalist commitment are not spontaneous products of economic change. They must be consciously organised, cultivated, and reproduced across generations.

The National Literacy Campaign of 1961 (Campaña Nacional de Alfabetización) was one of the Revolution’s earliest demonstrations of this principle. More than 100,000 young brigadistas of the Conrado Benítez Brigades travelled to the countryside as part of a nationwide literacy army. By the end of the campaign, more than 700,000 adults had been taught to read and write, reducing illiteracy in Cuba to 3.9%. Yet literacy was only one outcome. Those who taught were transformed alongside those who learned, encountering rural poverty firsthand and participating directly in the remaking of Cuban society. Education became not simply the transmission of knowledge but the formation of a new revolutionary subject.

The same principle shaped the Revolution’s publishing houses and journals, film institutions, art schools, scientific centres, medical training programmes, and educational exchanges with the peoples of Africa, Asia, and Latin America. Together they formed the cultural infrastructure through which the Revolution sought not only to defend itself but to reproduce itself and contribute to the wider anti-imperialist struggle.

The Cultural Congress of Havana, 1968

For eight days in January 1968, around 500 participants from more than seventy countries gathered at the Hotel Habana Libre for the Cultural Congress of Havana. The congress is often remembered as a meeting of writers and artists, but this barely captures its scope. Participants included novelists, poets, economists, physicians, scientists, architects, filmmakers, publishers, educators, sociologists, critics, cultural administrators, and militants from national liberation movements. Dependency theorists shared commissions with painters. Teachers worked alongside guerrillas. Representatives of newly independent states met organisers operating clandestinely under colonial rule. Delegations included participants from the Democratic Republic of Vietnam and the South Vietnamese liberation movement, as well as from the liberation struggles in Guinea-Bissau, Angola, and Cape Verde, alongside delegates from across Latin America, Africa, Asia, Europe, and North America. They met not as representatives of isolated national cultures but as participants in a common anti-imperialist project. The congress reflected the Cuban Revolution’s expansive understanding of culture – not as artistic production alone but as the entire field through which societies produce knowledge, values, institutions, technical capacities, and political consciousness.

Left: Kael Abello (Venezuela, Utopix/Tricontinental); right: Olfer (Perú, Utopix).

The congress advanced three interrelated arguments.

The first concerned the role of intellectual labour. Revolutionary struggle required far more than sympathetic artistic representation. Writers, scientists, physicians, teachers, economists, filmmakers, publishers, and cultural workers all possessed specialised forms of knowledge that could either reproduce imperial domination or be organised in the service of liberation. Their task was not to observe the Revolution from outside but to place their work within the collective project of national liberation and socialist construction.

The second was organisational. In his ‘Message to the Tricontinental’, published in April 1967 while he was fighting in Bolivia, Che Guevara called for ‘two, three, many Vietnams’ to stretch imperialism across multiple fronts. By the time the congress met in January 1968, Che had been killed in Bolivia three months earlier. Closing the congress, Fidel extended that strategy onto another terrain, calling for ‘un Vietnam en el terreno de la cultura’ (a Vietnam in the field of culture). The struggle against imperialism could not be confined to the battlefield or the factory. It also required institutions capable of contesting the production of knowledge, the circulation of images, Western scientific and technical authority, publishing, education, and the international media.

The third argument challenged the geography of history itself. Cuban poet and later president of the Casa de las Américas, Roberto Fernández Retamar, rejected the familiar division between ‘developed’ and ‘underdeveloped’ countries, insisting instead that the modern world was divided between countries that had been underdeveloped and those that had underdeveloped them. The implication reached beyond economics. If colonialism had relegated Asia, Africa, and Latin America to the ‘periphery’ of history, then the liberation struggles unfolding across the Third World had begun to overturn that geography. The principal engine of historical transformation no longer lay in the imperial metropoles but in the peoples fighting colonialism, imperialism, and capitalist domination. As Fernández Retamar would later crystallise in his essay Calibán, it was the Third World that stood at the centre of the revolutionary conjuncture and of history.

This legacy is the subject of Tricontinental’s latest dossier, A Vietnam in the Field of Culture: The Cultural Congress of Havana, 1968 (August 2026). Read it here.

FIDEL x 100

Nearly a decade after Fidel Castro’s death, the Cuban Revolution confronts one of its most difficult moments since the Special Period. The United States has once again tightened the longest economic blockade in modern history. Cuba’s continued inclusion on the US State Sponsors of Terrorism list has further restricted the island’s access to international finance and trade, while Washington has moved to impose a total blockade on oil deliveries to the island, barring one Russian tanker that got through in March 2026. The effects of the sanctions are the intended consequences of an economic war designed to make everyday life increasingly difficult and undermine the social foundations of the Revolution.

Economic war has always been accompanied by ideological war. The blockade seeks not only to constrain Cuba materially but also to isolate its revolutionary experience from new generations and from the peoples of the world. Against this strategy, the defence of the Revolution cannot be reduced to economic resilience alone. It also requires the continued organisation of culture, political education, historical memory, and international solidarity. In this conjuncture, propagating Fidel’s ideas is not an exercise in remembrance but part of defending the revolutionary process itself.

Left: Tings Chak (China, Tricontinental); right: Vanshika Babbar (India, Tricontinental/Young Socialist Artists).

It is here that FIDEL x 100 finds its place.

The exhibition is one contemporary expression of the international cultural front that took shape in Havana nearly six decades ago. It brought together contributions from more than 100 cultural workers, many of them activists in people’s organisations – from the Young Socialist Artists in India to the Landless Workers’ Movement (MST) in Brazil to the Party for Socialism and Liberation in the United States.

The posters approach Fidel from different angles. Some return to the young revolutionary of ‘History Will Absolve Me’, his self-defence after the Moncada assault in 1953. Others revisit the battle of ideas, literacy, agrarian reform, medical internationalism, and anti-imperialist internationalism. Their diversity reflects the many fronts on which Fidel’s legacy continues to speak to present struggles.

The exhibition has already travelled to the Mariátegui School in Argentina, MST encampments in Brazil, mobilisations in Mexico commemorating the 26 July Movement, Hands Off Asia gatherings in Sri Lanka, and political education courses in Pakistan.

Download the archive. Print the posters. Hang them in political schools, trade union offices, community centres, party headquarters, libraries, universities, movement spaces, and neighbourhood assemblies. Use them to teach, to debate, to commemorate, and, above all, to organise. Let us multiply the images and ideas of the Cuban Revolution, carry forward Fidel’s legacy, and continue the task of creating ‘a Vietnam in the field of culture’. A century after Fidel’s birth, the task remains before us.

In solidarity,

Kael Abello, member of the Utopix art collective and the Art and Culture Department of Tricontinental: Institute for Social Research.

Tings Chak, co-coordinator of the Art and Culture Department of Tricontinental: Institute for Social Research.

On 13 August 2026, DeepSeek released DeepSeek Harness, an open framework for building and customising AI agents. Its philosophy can be summed up in two words: freedom and openness. Instead of giving users a finished agent and asking them to work within its limits, Harness opens up the system itself, allowing users to inspect, modify, and extend it.

What Makes DeepSeek Harness Different

Customisable

It is not a finished product, but a box of building blocks.

Tools such as Claude Code and Codex are largely finished products: you can change their settings and add instructions, but the underlying system, including how the agent works with the model, uses tools, manages sessions, and presents its interface, is largely determined by the developer.

DeepSeek Harness takes a different approach. Its basic principle is ‘everything is a plugin’. The model, tools, skills, agent loop, session system, sandbox, and even the user interface are all built as replaceable components. There is no single ‘official’ core that users are expected to leave untouched.

This is made possible by Cordis, the plugin framework underneath DeepSeek Harness. In simple terms, Cordis is what allows these different pieces to be added, removed, connected, and replaced without rebuilding the whole system. It also keeps track of which pieces depend on which others, and cleans up the changes made by a plugin if it is removed.

For users without a technical background, this has a simple meaning: you are not limited to the features the original developers decided to give you. You can add new capabilities, replace existing ones, or even change how the interface works. The growing plugin ecosystem already includes tools that add new functions and new interface components.

Later in this article, we will see what this means in practice: using natural language to change the interface, and then turning a research skill into a visual dashboard with a timeline and information cards.

Traceable

The agent’s work leaves a detailed trail. Instead of seeing only the final answer, you can follow the whole execution record: what the agent was instructed to do, what tools it called, what those tools returned, and what information was passed into the next step.

This distinction matters when something goes wrong.

Imagine an AI spends several minutes searching files, reading documents, and making changes. If you only see a condensed account of what it did, you may know that the final result is wrong without knowing where the mistake happened. With a detailed execution record, you can go back through the sequence and ask: What did the agent see? What did it do? What did the tool return? At which step did the wrong turn happen?

That means debugging does not have to begin by simply running the entire task again and hoping for a better result. You can inspect the path that produced the result and identify where it went wrong.

This is also what makes the visual dashboard later in this article possible. The dashboard is, in effect, turning the agent’s execution record into something a human can read and understand at a glance.

How to Install It

At the time of writing, the official installation process requires using a terminal at least once. Although there are ways to turn DeepSeek Harness into a standalone application, as shown later in this article, we will begin with the official terminal-based method so that you can see how it works. The process is straightforward.

Install DeepSeek Harness on Mac

Press Command + Space to open Spotlight, search for Terminal, and open it.

Copy the following command into Terminal and press Enter. This installs Node.js, starts DeepSeek Harness, and opens its web interface. If you restart your computer, you will need to start it again.

Note: If something goes wrong, first check your network connection. If you see ‘zsh: command not found: brew’, install Homebrew first by running:

/bin/bash -c "$(curl -fsSL 
https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"

Then run:

brew install node && npx -y 
@deepseek-ai/dsh web > ~/.dsh/dsh-web.log 2>&1 & sleep 3 && open http://127.0.0.1:3080

This is only a test statement. Press the button below to continue.

The next window asks for an API key. If you do not have one yet, simply use the first button. We will create one in the following section.

Occasionally, the browser may show ‘This site can’t be reached’, especially during the first launch. This usually means the local service has not finished starting. Wait a moment and refresh the page. After closing the tab, you can reopen DeepSeek Harness at http://127.0.0.1:3080.

Install DeepSeek Harness on Windows

Press the Win key, search for PowerShell, and open it.

Copy the following command into PowerShell and press Enter. It installs Node.js, starts DeepSeek Harness in the background, and opens the local web interface. After restarting your computer, you will need to run it again.

winget install OpenJS.NodeJS.LTS -e --accept-source-agreements --accept-package-agreements; $env:Path="C:\Program Files\nodejs;$env:Path"; Start-Process cmd.exe -ArgumentList '/c','npx -y @deepseek-ai/dsh web > "%USERPROFILE%\.dsh-web.log" 2>&1' -WindowStyle Hidden; Start-Sleep -Seconds 3; Start-Process "http://127.0.0.1:3080"

This is only a test statement. Press the button below to continue.

The next window asks for an API key. If you do not have one yet, simply use the first button. We will create one in the following section.

If the browser shows ‘This site can’t be reached’, especially on the first launch, wait a moment and refresh the page. After closing the tab, you can reopen DeepSeek Harness at http://127.0.0.1:3080.

Basic Configuration

Click the Settings button in the lower-left corner to open the DeepSeek Harness settings page. It is a useful way to explore and better understand its features.

General

The General tab contains several basic settings. For example, you can change the language if necessary. Currently, English and Chinese are the two officially supported languages.

Agent Preset determines what the AI can do. Standard Mode is for most users. Code Mode is useful for repetitive tasks. Minimal Mode is mainly for testing. Creator Mode lets you build your own DeepSeek Harness setup.

Permission Mode determines where the AI can act. Read Only lets it read files without changing them. Workspace Write limits file changes to the folder you selected. Full Access allows it to modify files anywhere on your computer.

Models

Go to the Models tab to configure the model that powers your Harness. The language model handles the reasoning and planning, while DeepSeek Harness provides the tools and execution environment. Since DeepSeek Harness is designed to work particularly well with DeepSeek models, we will mainly use DeepSeek as the example here.

To use a DeepSeek model, you first need an API key. Go to https://platform.deepseek.com/api_keys and sign in. Then click Create new API key.

Give the key a name and click Create API key.

Click Copy to copy the API key, then return to DeepSeek Harness and paste it into the provider settings before applying it.

DeepSeek Harness also supports many other model providers. Click Add provider to browse the available options. You only need an API key from the provider you want to use.

DeepSeek Harness also provides a convenient way to discover the models currently available from a provider. For example, after adding moonshotai and entering its API key, click Customised settings, then use Fetch available models to retrieve the latest model list.

Plugins

Plugins are at the core of DeepSeek Harness, but you do not need to understand them in detail at the beginning. For now, it is enough to know that DeepSeek Harness is highly customisable. You can turn your own experience into reusable instructions, create your own programs, or change the default functions and settings of the Harness.

DeepSeek Harness also provides instruction files that can be used by AI agents. You can use DeepSeek Harness itself to create new plugins or modify existing ones.

Agent Presets

The Agent Presets tab shows the presets available for the Harness. You can click the view button for a preset to inspect its configuration and instructions.

You can also create your own customised preset. In the next section, we will use DeepSeek Harness and explore how it differs from other agent tools.

Using DeepSeek Harness

Dialogue Page

Click the dialogue box in the centre to choose a folder as your workspace. DeepSeek Harness will work within this folder. When Permission Mode is set to Workspace Write, it can only modify files inside the selected workspace.

You can change the model used for the current task by clicking the model selector next to the submit button.

Effort determines how deeply the model investigates a task. Higher effort does not always mean better results; it can also increase both execution time and cost. For most tasks, the default setting is a good starting point.

You can also adjust the other task settings directly from the dialogue page.

Transform Your Main Interface

As mentioned earlier, DeepSeek Harness allows us to create a new plugin by ourselves. Let us start with a simple example.

Have you ever forgotten the name of a skill installed on your computer, or found it troublesome to type /xxx every time you wish to invoke it? DeepSeek Harness allows you to build a convenient plugin to solve exactly this problem; all you need to do is describe your needs to the AI.

Create a Skill Library plugin integrated into the main interface,
allowing users to quickly browse, preview, and insert skills.

After completing the task, click the Skills tab to open the Skills Library. This page displays basic information about each skill, making it easier to review and manage our skills.

Click Insert into composer to insert a reference to the selected skill into the chat box. This saves time and makes the skill easier to use.

Because this model’s capabilities are strong and this requirement is relatively straightforward, there is no need for us to adjust the result. If any problems do exist in your plugin, simply describe them and ask the AI to solve them.

Make a Plugin for Your Skill

In an era when AI is often used together with skills, many people simply type in the name of a skill and wait for the output. The execution process, however, remains a black box: if the output is not satisfactory, we have no choice but to run the skill again, which is a waste of time and money. Although the AI may occasionally ask us to make a choice, such prompts are random and far from sufficient.

By using DeepSeek Harness, we can upgrade our skills with visual interfaces that allow us to monitor and intervene in the operation of the AI. The section below uses POMASA as an example.

POMASA is an open-source skill that creates a multi-agent research system to investigate a given topic; the code is available at https://github.com/eXtremeProgramming-cn/pomasa. Its defining feature is that it separates the research process into a variety of stages and executes them step by step based on the given topic.

We can therefore build an interface for the various stages, see which stage is currently running, and see which files each stage generates, thereby keeping control of the implementation. For example, we can view the intermediate files while the AI is running, without having to search for them in the Finder, and spot errors in good time.

To get started, we need to install POMASA in our project folder.

Please install the following skill as a project skill in the current project.
npx skills add eXtremeProgramming-cn/pomasa

To begin our visualisation, it is better to understand roughly how this skill works. It is perfectly fine to simply tell the AI to create a visualisation plugin for it, but if you know what your skill will require from human operators in the long term, the visualisation can be much better.

Simply speaking, POMASA separates a complicated task into several stages, provides methodology, generates initial ideas, and builds a reusable research system according to your study. We can use the official prompt as an example.

Help me create a multi-agent research system for analysing AI trends in healthcare.

After the AI has completed its work, we can see the working folder it has produced, which contains several subfolders:

• agents: Sets up the workflow for the research.
• library: Contains the external resources collected by the AI for writing.
• references: Contains the primer, initial ideas, and methodology that serve as the basis of the study.
• wip: Records the state of the study, such as when each session started and ended.
• workspace: Contains the outputs generated at different stages of the workflow.
• _output: Contains the final report generated by this skill.

Here is a simple example of a workspace folder generated by a workflow of ‘collect sources -> verify sources -> analyse arguments’. Many skills use similar structures to break complicated tasks down into a series of steps.

However, if the sources searched are wrong, the analysis will probably produce the wrong arguments and cost a great deal of time and money in an unexpected direction. Since the workflow has been separated clearly and has files that can be reviewed by humans, it would be ideal to build an interface that guides the user to read them.

A clear interface can also decrease the difficulty of usage: there is no need for other people to read the specifications and documents before using a skill.

To begin with, we can simply describe our needs.

/pomasa Design a research-process visualisation interface
called POMASA Observatory for DSH that provides a real-time
view of a POMASA multi-agent research workflow.

The interface is an observer:
its purpose is to visualise the research process,
agent activity, and files.

Click the tab named POMASA next to the chat. The first version may not be perfect, nor conform to the vision in our minds, but do not worry, we can complete it step by step. As you observe the interface generated by the AI, you may also come up with new ideas for improving it.

For example, I thought we needed a timeline instead of showing one line for each stage, so that we could see which stage the AI is in and what it is doing. I used the prompt below to modify it.

The RESEARCH STAGES should be displayed as a timeline from left to right.
Below, users can see the actual execution in the chat in real time,
while the same process is simultaneously translated into
a clear visual language through structured graphics.

After several rounds of modifying the style and adding functions, it has completed the task of converting the skill into a visualisable plugin. The whole process does not require professional technical skills; you only need to describe your needs in natural language.

Let us look at the generated result. Click Tab 1, Trend Signal Collection. In this stage, the AI searches for data online, summarises them, records their sources, and classifies them. In the example run, it collected 19 cards for this research.

Open one card, and we can see the content of the source. Each source has been clearly organised into three sections: Core Content, Key Quotes, and Credibility Assessment, making it easier for humans to read and audit.

It will be updated as the research system runs, so people can check it and find errors immediately.

Tab 2, Source Verification, checks whether the source of each card is credible.

Tab 3, Trend Analysis, puts forward arguments on the basis of the cards collected earlier.

Since it is related to the cards collected earlier and forms the basis of the final report, it plays an important role in shaping the final result. If something goes wrong at this stage, the quality of the final article will be affected.

The last stage shows the final report, which is assembled from the artefacts created earlier.

Plugins serve as a visual representation of the thought process behind skills, which means that users do not need to understand a skill deeply to intervene in its process.

This makes the skill more complete and opens up possibilities for users without advanced technical skills to customise their experience.

Conclusion

DeepSeek Harness is more than a tool for running AI agents. Its significance lies in opening up the agent itself: users can add, replace, and customise the components that make it work, rather than being limited to a fixed system. Built around a plugin-based architecture, and supported by the Cordis framework, Harness gives users greater freedom to shape both the agent’s capabilities and its interface.

Just as importantly, Harness makes the agent’s work more inspectable. By retaining a detailed record of the execution process — including the instructions, tool calls, intermediate results, and other information passed between steps — it becomes easier to understand what the agent actually did.

The larger promise of DeepSeek Harness, then, is not simply that it can run more capable AI agents. It is that it gives users more freedom to shape the system and see how it works — making AI workflows more customisable, inspectable, and controllable.