Hard facts and half-truths: The new archival history of China’s Great Famine

Books reviewed

Frank Dikötter, Mao’s Great Famine: The History of China’s Most Devastating Catastrophe, 1958–62. London: Bloomsbury, 2010.

杨继绳 (Yang Jisheng), 墓碑 － 中國六十年代大饑荒紀實 (Mubei: Zhongguo liushi niandai da jihuang jishi) (Tombstone: A chronicle of the Great Famine in China in the 1960s). Hong Kong: Tiandi tushu, 2008.

Yang Jisheng, Tombstone: The Great Chinese Famine, 1958–1962, Introduction by Edward Friedman and Roderick MacFarquhar, trans. Stacy Mosher and Guo Jian. New York: Farrar, Straus and Giroux, 2012.

Zhou Xun (ed.), The Great Famine in China, 1958–1962: A Documentary History. New Haven, CT: Yale University Press, 2012.

In 1962, Joseph Alsop was invited by the editor of The China Quarterly to elaborate on a thesis introduced in his widely syndicated column that ‘Communist China has somehow been caught in a remorselessly descending spiral’. ¹ Successive bad harvests and heavy grain procurement quotas had reduced the Chinese countryside to a condition of creeping starvation, Alsop wrote. He presented some statistical data in support of his thesis, derived from surveys of mainland Chinese refugees in Hong Kong:

The most reliable data … derived from great numbers of refugee interrogations and collected and analysed with extreme care, showed an average food intake for mainland China of 1,300 to 1,600 calories per person per day. ²

This was a shockingly low range, much lower than common estimates of average food intake in China before the Great Leap Forward, and below the level classified as a starvation diet by the US Occupation administration in post-war Japan. ³ Alsop proposed that through careful rationing the Chinese government had managed to avoid a dramatic episode like the Ukrainian famine of 1933, but that a poor harvest in 1962 might push the Chinese population beyond the limits of its endurance.

The following issue of The China Quarterly published responses to Alsop’s article from 10 leading scholars of contemporary China. While several of the respondents were sympathetic to elements of Alsop’s thesis, all but two raised fundamental questions about his use of statistical evidence. In particular, the calorific intake figures were presented without citation, and were substantially lower than more commonly cited estimates of average per capita food intake in the refugees’ home counties of between 1600 and 1800 calories per day. It is likely that Alsop derived the lower figures himself by applying a seemingly reasonable discount to these established estimates to allow for the place of origin bias in the sample of refugees, most of whom came from the relatively prosperous Guangdong counties adjacent to Hong Kong. ⁴

Kenneth Walker, an agricultural economist who studied conditions behind the bamboo curtain with the aid of Chinese newspapers, transcripts of radio broadcasts, and rare compendia of official statistics, insisted that further information was needed if Alsop’s data were to be taken seriously. ‘How representative was the sample? What did tests of significance reveal?’ Walker then drew upon official grain production data to present a very different portrait of the food situation in China. He argued that per capita food consumption amongst China’s agricultural population had risen in most years between 1952 and 1957; that ‘all the evidence’ pointed to a bumper harvest in 1958; and that large stocks put down in 1958 ‘help to explain why China has not suffered from widespread hunger’ despite three subsequent years of poor harvests. ⁵

In this instance, Alsop’s creative interpretation of an unrepresentative sample of anecdotal evidence provided better witness to the reality behind the bamboo curtain than Walker’s careful analysis of data gleaned from the official Chinese government sources available at the time. (Official retrospective fertility surveys conducted in the 1980s bear out Alsop’s claim that living standards declined sharply in 1958 and fell further in 1959, taking most of rural China into conditions of severe famine rather than creeping starvation. ⁶ ) Walker was not the only leading exponent of empirical scholarship on China who failed to detect the largest famine in modern human history. G. William Skinner saw a breakdown of the rural trading network but not famine; ⁷ Dwight Perkins acknowledged that certain parts of China experienced serious ‘weather problems’ between 1959 and 1961, but that famine had been averted thanks to the expansion of transport infrastructure over the previous decade. ⁸

Between the mid-1980s and the late 1990s, a loose consensus of knowledge on the famine emerged out of the work of demographers, economists, and political scientists. It was premised on a limited set of official source materials released by the State Statistical Bureau (SSB) and other central government information agencies in the early Deng era. These included provincial crude mortality rates and grain production figures, collections of high-level policy documents, official biographies, and the memoirs of leading personalities of the reform era such as Bo Yibo. The loose consensus, challenged in the 1980s and today by a small but resilient circle of radical academics, bloggers, and activists, embraced estimates of excess deaths of between 24 and 30 million, and a set of plausible causes of the famine that included radicalism of local Party leaders, mismanagement of communal dining halls, urban biases in economic and welfare policies, and poor weather.

Over the past decade, this loose consensus has been undermined by historians and journalists who have discovered a new range of sources on the famine. Many detailed local histories of the famine era have been published which draw on the memoirs of local Party officials, oral history interviews, and documents obtained from local Party archives. ⁹ Large collections of internal Party reports on the famine have been published or uploaded onto the Internet. ¹⁰ Detailed data collections compiled for internal use by various government agencies in the 1980s have attracted the attention of several scholars. ¹¹ All of these contain descriptive and statistical information on the famine era that are not readily reconciled with those currently endorsed by the central government information agencies.

The historical writing based on these new sources constitutes the new archival history of the famine. Its proponents generally advocate a much higher famine toll than the social scientists that came before them. They tend to allocate blame for the famine squarely on the shoulders of the central Party leadership and on Mao Zedong in particular, with other factors ascribed a secondary significance. They generally assemble a more compelling narrative of the famine by making greater use of anecdotal material, and by bringing the policy debate at Party Central into dialogue with everyday life in the villages.

This article reviews the two books that introduce the new archival history of the famine to an English-language readership. Yang Jisheng’s Mubei: Zhongguo liushi niandai da jihuang jishi (Tombstone: A chronicle of the Great Famine in China in the 1960s; hereafter Tombstone) was published in Hong Kong in 2008. The book was greeted with enthusiasm in the blogosphere but initially attracted limited attention in scholarly and literary journals. ¹² An abridged English edition translated by Stacy Mosher and Jian Guo was released in late 2012 under the title Tombstone: The Great Chinese Famine, 1958–1962 (US edition), which has attracted a number of extended reviews. ¹³ French and German translations were published in coordination with the English translation. ¹⁴ The original Chinese edition of Tombstone is reviewed here, with only a brief discussion of the abridged English translation.

The second book reviewed here, Mao’s Great Famine: The History of China’s Most Devastating Catastrophe, 1958–62 by Frank Dikötter, was released two and a half years after the Chinese edition of Tombstone. While the book was initially greeted with much acclaim in literary journals and online magazines, it has received more critical attention in academic journals. ¹⁵ A collection of documents on the famine translated by Dikötter’s long-term collaborator Zhou Xun, many of which are cited by Dikötter, was published in mid-2012.

Both Yang and Dikötter present plausible accounts of how Mao and his colleagues established the policies and institutional structures that gave rise to the famine. They both provide many vivid descriptions of the suffering the famine caused at the village level. This review reflects critically on the nature of the sources cited and the methodologies deployed in the two books. Special attention is paid to the representativeness of evidence drawn from archival and other sources, a central problem in research on the Great Leap era from 1962 through to the present.

Tombstone

Yang Jisheng presents Tombstone in memory of his uncle and adoptive father, who starved to death in Yang’s native village in Hubei Province in the spring of 1959, and of the millions of other Chinese people whose lives were cut short by the famine. He also presents Tombstone as a cautionary tale for the living. The caution is for the author himself, who suffered a cancer scare while writing a book that itself carried certain political risks, and for the authoritarian political system that gave rise to the famine. While the pathological changes in Yang’s body turned out to be benign and the publication of Tombstone has not yet deprived Yang of life or liberty, he questions whether China’s body politic will be able to cleanse itself of the authoritarian cancer within. In the boldness of its challenge to authority, the book is deserving of the alternate title proffered by several pundits of ‘China’s Gulag Archipelago’. ¹⁶

Like many intellectuals of his generation, for the first five decades of his life Yang was not aware of a connection between the policies of the Party and the starvation deaths that he witnessed personally and was told about by family and friends. As a student at his local county high school and then in the mechanics faculty of Tsinghua University, Yang’s grain ration was guaranteed by the state. He graduated in May 1966, on the eve of the Cultural Revolution, and was allocated a job in the Tianjin office of the Renmin ribao (人民日报). He remained an employee of the state press until his retirement in 2001, gaining many professional accolades along the way. After 1989, he moonlighted as the author of several books on social and political tensions in the reform era that strayed well outside the constraints of the official media, including an account of the factional struggles within the Party in the lead-up to the Tiananmen massacre. ¹⁷ Since 2003, he has served as one of the deputy editors of Yanhuang chunqiu (炎黄春秋), which since its founding in 1991 has been the most widely read journal on Party history. Under Yang’s editorial guidance, Yanhuang chunqiu has published several dozen memoirs and local histories of the Great Leap Forward and famine, making it the leading vehicle for critical reflection on the Party’s responsibility for the famine.

The two-volume Chinese edition of Tombstone was published in Hong Kong to mark the 50th anniversary of the commencement of the Great Leap Forward campaign. While distribution of the book is banned on the mainland, pirated copies are available at street-side booksellers and circulate on the Internet. Volume One consists of a dozen province-level case studies, based primarily on investigatory reports commissioned by central and provincial Party committees between 1960 and 1962. In Volume Two the author examines the broader causes of the famine. In a sequence of thematic chapters, he refutes the stock official explanations for the ‘hardships’ of the Great Leap Forward, and argues that the fundamental cause of the famine was the top-down radicalization of the Chinese Party-state. In this second volume, Yang draws on the standard published documentary collections and memoirs of retired senior leaders, as well as select meeting minutes and internal reports sourced from provincial archives. Throughout the two volumes he cites the work of historians published in mainland China and Hong Kong over the previous 10 years, including articles from Yanhuang chunqiu. ¹⁸

The first volume of Tombstone opens with a 50-page account of the ‘Xinyang Incident’, the term used within the Party in 1960–61 to refer to occurrences of famine and related violations of Party discipline in one prefecture of Henan Province. This is a superb case study of the interplay between local and central Party leaders during different phases of the mass mobilization campaigns of the Great Leap era. Yang presents a rich synthesis of the documents produced by the Party-state as it first nurtured the famine then, in a different mode of revolutionary governance, responded with some effectiveness to the immediate causes of the famine. Aside from the special investigatory reports commissioned in the latter cooling-down phases of the Great Leap, Yang draws on other official documents, interviews with retired local Party officials, and a long memoir on the incident written by Zhang Shufan, the former head of the Xinyang prefectural administration. ¹⁹

The Xinyang local Party leadership were the first to respond to the March 1958 instruction issued by the Politburo that agricultural collectives should be combined into larger units. The leaders of four townships in Suiping county, under the close guidance of the prefectural Party committee, merged their jurisdictions to form the Chayashan Large Collective. The Henan provincial leadership promptly renamed this new administrative unit as the Chayashan Sputnik Agricultural Collective (嵖岈山卫星集体农庄), inspired by Soviet initiatives in large-scale farm management and aeronautical engineering. When the Minister of Agriculture came to Chayashan on an inspection tour, he likened the high degree of collectivization to that of the Paris Commune, a pre-Soviet model praised by Lenin at the height of the Civil War. After further deliberations between central and local officials, the Chayashan Sputnik People’s Commune came into being. ²⁰

These changes in the organization of agricultural production brought instant results, as reported by the state media. On 8 June 1958, the Renmin ribao announced that No. 2 Production Team of Chayashan had produced an experimental plot of wheat yielding 15.8 tonnes per hectare, more than 10 times the normal yield. Four days later the Renmin ribao reported that No. 1 Production Team had achieved a yield of 25.6 tonnes. Even greater successes were soon reported from other counties in Xinyang prefecture and further afield, as the Sputnik-era Lysenkoist propaganda exercise to perform miraculous feats of agricultural production came to life. The Xinyang Party secretariat confidently announced a record harvest for the prefecture as a whole, and the corresponding record grain tax quota was duly delivered to central government grain stores. The reports of new record yields continued through 1959, despite severe drought, and grain deliveries increased apace. As explained in the memoir of Zhang Shufan, the reports of agricultural successes had little to do with reality. The actual level of grain production in 1959 was well below average, and with a record amount of tax grain taken out of this reduced harvest, a large section of the local population starved to death. ²¹

The dozens of survey reports on conditions in the villages of Xinyang cited by Yang, commissioned by various Party and state organizations at different phases of the Great Leap campaign, systematically distorted reality in different ways depending on the particular Party mobilization efforts that the reports were intended to serve. Those produced in the summers of 1958 and 1959 tabled spectacular agricultural successes in support of very ambitious grain collection campaigns. In the autumn of 1959, the reports focused on the ‘hoarders and class enemies’ who sought to prevent the state from collecting its claimed lawful share of the grossly inflated harvest, thereby justifying greater use of extraordinary measures. From early 1960, the focus of the reports shifted from enemies outside the Party to enemies within the Party itself, as the central Party leadership sought to limit the political and economic fallout of the famine.

In February 1960, a branch secretary of the central internal affairs department, responsible for producing statistics on the number of households affected by spring food shortages, reported to the central Party supervision commission that 200–300 people had died of starvation in Xinyang over the winter months. This opened the way for a flood of reports that documented harvest failure, misconduct of local cadres, and mass starvation. These reports, tabled between April 1960 and November 1961, were used to guide a process of internal self-criticism that moved from brigade-level Party organizations up through the provinces, culminating eventually in the famous 7,000 Cadre Conference convened in January 1962.

Yang quotes the reports on the number of people in different production brigades, communes and counties who were beaten or starved to death in the winter and spring of 1959–60. (In the final report approved by the Henan provincial Party committee, the Xinyang death toll was 549,171, or 6.5 per cent of the prefectural population.) He also draws on interviews with officials involved in the compilation of the reports, describing behind-the-scenes negotiations over the scale and nature of the ‘mistakes’ that should be reported. Mao Zedong personally characterized the excesses carried out in Xinyang as a counterrevolutionary restoration (反革命复辟), an extreme verdict, that led in some counties to the wholesale removal of the leadership of most production brigades (two administrative tiers below the county) and of a quarter of the production teams (a further tier below). ²² The purge was carried out through a mass campaign facilitated by the security agencies, in which both former Party leaders and the usual undesirable elements were publicly beaten and in many cases killed.

The Central Supervision Commission, after close consultation with several central Party leaders and the Henan provincial Party secretary, formulated the following stern judgment on the responsibility of the prefectural Party leadership for the famine:

In Xinyang prefecture over the past winter and spring, [prefectural Party secretary] Lu Xianwen and other counterrevolutionary and bad elements usurped control of Party and state offices, and in the guise of opposing rightist deviationism, zealously carried out the anti-smuggling campaign. They used methods of the landlords and the Nationalists, including indiscriminate beatings, arrests and killings, to exact large-scale class revenge on the 8 million residents of the district, creating a shocking scene of terror. ²³

The same report on the Xinyang Incident recommended that at least 10,000 of the perpetrators be executed, including at least two county Party secretaries, with quotas of 800 executions given to large counties within the prefecture and 400 to small counties, an average of 3 to 5 executions for each production brigade.

This was the judgment of a committee of experts, not a final sentence. The influence of the Central Supervision Commission’s report was curtailed by the personal intervention of Mao, who was concerned about allocating blame too high up the chain of command. He pencilled on the report the comment: ‘I haven’t yet killed a county Party secretary; give them commuted death sentences.’ ²⁴ This released the county Party secretaries off death row, though eight of the 18 county Party secretaries as well as the prefectural Party secretary were expelled from the Party, and some were given gaol sentences of two to three years. (Yang does not provide a full tally of the punishments meted out further down the chain of command.) The same principle of leniency was respected as the internal self-criticism process moved higher up the decision-making hierarchy. Several provincial Party secretaries were demoted due to the neglect they had shown to the populations under their leadership, but none of the senior provincial Party secretaries were formally reprimanded. ²⁵

The remainder of Yang’s opening chapter describes the broader political context of the Xinyang Incident: the dismissal in May 1958 of the First Party Secretary of Henan, Pan Fusheng, and his ‘Anti-Party Clique’; Pan’s replacement by Wu Zhipu; and the promotion under Wu’s leadership of Henan as a model Leap province. Yang’s account maintains a sharp focus on the active role played by central Party leaders in organizing, rather than merely sanctioning, these exemplary local political struggles. He concludes the chapter with a discussion of the famine in Henan as reflected in official grain and demographic data. Here he draws on a set of provincial grain statistics obtained from the central Party archives that is similar to the series produced by the Ministry of Agriculture in the early 1980s, and vital statistics from a provincial statistical almanac which are similar but not identical to the SSB series. ²⁶ While Yang is transparent in his use of sources and his arithmetic is sound, he does not make any major breakthroughs with the numbers. Like much of the quantitative research on the famine, this part of Yang’s analysis is premised on official data on grain production and crude death rates, statistical series that were highly politicized both during and in the aftermath of the Great Leap campaign, and the various distortions and corrections applied to these data are as poorly understood today as when they were first made public 30 years ago. ²⁷

Yang presents chapter-length accounts of the famine in 11 other provinces, each describing how the provincial Party leadership launched and later reined in the Great Leap campaign, followed by a tally of the associated human and economic costs. These provincial case studies each emphasize different aspects of the famine. The chapter on Gansu presents a case study of the Great Leap irrigation campaign, in the form of an extended account of the brutal and economically fruitless effort to redirect one of the headwaters of the Yellow River. The Sichuan chapter describes in detail the measures taken by the provincial Party leadership to suppress information about food shortages and famine deaths. Yang describes investigations similar to those that defined the Xinyang Incident that were carried out in Gansu (the ‘Tongwei Question’), Shandong (the ‘Jining Incident’), and Yunnan (the ‘Luliang Incident’), and elsewhere. He highlights some of the unique contributions made by different provinces to the national Leap programme. For example, the anti-hoarding campaign, implemented on a national level during the harvest season of 1959, was trialled the previous winter in Guangdong under the oversight of Zhao Ziyang, then secretary of the Guangdong provincial Party secretariat. ²⁸ The remaining chapters of Volume One provide a summary of the remaining ‘incidents’ through which the Party came to terms with the famine in other provinces, a comparison of the severity of famine in different provinces that draws heavily on Cao Shuji’s book-length overview of population data in prefectural and county gazetteers, and a discussion of life in the cities during the famine period.

In Volume Two of Tombstone, the focus shifts from the provinces to Beijing, where Yang systematically presents the case that the central Party leadership has a high degree of responsibility for the famine. He opens with a forceful rebuttal of the two main officially sanctioned explanations for the ‘three years of hardship’: its underlying cause was not bad weather (rainfall and temperature levels during the Great Leap Forward were unexceptional); nor was it due to the Soviets reneging on their contractual obligations (the Soviets abandoned their technical assistance programme only after the famine had taken hold). ²⁹ He then describes the institutionalized practices that transformed a series of central Party leadership decisions into mass famine. These included the twin policies of ambitious economic planning and anti-Rightism promoted in successive waves of mass campaigns from late 1955; the over-rapid expansion of the urban population that accompanied the pursuit of ambitious planning targets; the stress this placed on the national grain distribution system; the radical policies implemented to resolve shortages in food and other key commodities (the establishment of communes and communal dining halls, and promotion of the ‘five winds’ once criticized by Lenin under the rubric of ‘Communist conceit’); the ability of the Party leadership to ignore and suppress bad news; and the code of personal loyalty to Mao that was cultivated by an authoritarian Party system. ³⁰ Volume Two concludes with chapters on the economic, social, and demographic toll of the famine, and its political repercussions within the Party.

Yang criticizes the authoritarian excesses of the Party system using the critical language and source material developed by the system. He does not spare any of the Party leaders whose political fortunes in the post-Mao period were built on a selective denunciation of China’s Maoist heritage. As mentioned earlier, Yang highlights the contribution to the anti-smuggling campaign of Zhao Ziyang, the reformist hero of one of Yang’s earlier books. Many examples are given of the crucial role in the formation and execution of the Great Leap programme played by Deng Xiaoping as head of the Central Party Secretariat, as well as Liu Shaoqi, Zhou Enlai, and other supposedly moderate members of the Central leadership team. ³¹ In his capacity as the senior Party leader with personal and military ties to the Southwest, Deng also encouraged Sichuan Party secretary Li Jingquan to disregard popular welfare in pursuit of central government grain quotas, and shielded Li from criticism during the 7,000 Cadre Conference. ³²

Yang’s history is very long (1,200 pages of small type, over 800,000 characters) and is structured in a way that will seem convoluted to readers unfamiliar with Chinese geography and the Chinese communist practice of developing national policies through local pilot programmes. The editors of the English translation have produced a much shorter book that is easier to read from cover to cover, interspersing the first four chapters of Volume One with the main narrative chapters of Volume Two. The four provincial case studies selected for translation, on Henan, Gansu, Sichuan, and Anhui, cover the districts that bore the brunt of the grain collection campaigns in each of the four economic coordination regions most severely affected by famine. ³³ The translators have done their work well, and have succeeded admirably in rendering the operational jargon and campaign slogans of the Great Leap era in plain English. Understandably, some of the behind-the-scenes activity in the formulation and roll out of various government initiatives have been left out of the English translation, though for specialist readers these are amongst the most interesting passages of the book. The footnotes have been reproduced faithfully from the Chinese edition; this includes some archival references, though Yang follows the common though not ideal convention of providing the name and date but not the archive location and fond number for most archival documents cited. While the publication of the abridged English edition should be celebrated, the interest shown to date in Tombstone may yet justify an unabridged English translation.

Mao’s Great Famine

When Frank Dikötter’s Mao’s Great Famine was released in 2010, Tombstone was well established as the standard Chinese-language history of the famine. By the principle of scientific precedent, Dikötter’s contribution to scholarship should be judged against the benchmark established by Yang and other previously published work. Dikötter’s book is a third of the length of Tombstone and published in a single volume. While it repeats the two-fold division of Tombstone, into chapters that explore local perspectives on the famine and others that focus on life at Party Central, the order of the two parts is reversed. (These particular contrasts are not found when comparing Mao’s Great Famine with the abridged English edition of Tombstone, which is similar in length to Dikötter’s book and has several of the chapters on Party Central placed towards the front.) Aside from innovations in form, Dikötter highlights important international dimensions of the famine story, provides a different though not entirely original characterization of the nature of the totalitarian state, and points his readers to at least one new and important archival document.

In the first 16 of 37 bite-sized chapters, Dikötter presents a compelling narrative of how Mao and his colleagues steered the country into and through the famine. Several of these chapters provide fresh perspectives. He foregrounds the political rivalry between Mao and Khrushchev, drawing on recent Cold War histories by Lorenz Lüthi and others. He introduces the important context of international trade within the socialist camp, using new material from the Foreign Ministry archives of the PRC and of the former Democratic Republic of Germany. The documents cited through these early chapters of the book are generally well chosen, though he overplays the significance of his own archival excavations by quoting from archival documents when in most instances he could just as well have cited a standard published edition.

If Dikötter had ended the book at the conclusion of this general narrative he would have produced a worthy sequel to his previous book, which was a forceful rebuttal of the official Chinese Communist Party (CCP) view that all was dark in China before 1949. ³⁴ Instead, he ventures on for a further 21 chapters, each illustrating one aspect of the famine using archival anecdotes stripped of geographic, temporal, and institutional context. Here Dikötter provides his readers with a stream of vivid descriptions of suffering citizens neglected by their leaders, interspersed with numerical lists of terrible things, which are likely to generate discomfort on the part of the reader but not comprehension.

Several reviewers of Mao’s Great Famine, following the lead of the publisher and of the author himself, have made much of the great quantity of official documents cited in the footnotes. In awarding to Dikötter the 2011 BBC Samuel Johnson prize for non-fiction, the judges highlighted the ‘meticulous’ nature of his research in ‘hitherto hidden archives’. ³⁵ The archival documents cited by Dikötter fall into two main categories: those relating to the policy deliberations of senior Party leaders, quoted in the first half of the book; and surveys of conditions in the countryside cited in the second half. The manner in which Dikötter uses both types of documents is problematic, for different reasons.

For the survey materials, the problem concerns how he draws generalizations from fragmentary evidence. Dikötter claims that his book ‘brings together two dimensions of the catastrophe that have so far been studied in isolation’, linking ‘the corridors of Zhongnanhai’ with ‘the everyday experiences of ordinary people’. ³⁶ The main technique he uses to achieve this is one of juxtaposition. The words of Mao and other central Party leaders are juxtaposed against things that happened in a certain village at an uncertain time. The same method is used in reverse, when Dikötter finds universal qualities in the particular. The chapter on cannibalism opens with a description of ‘[t]he Chinese countryside’, once ‘a world full of noise’ where ‘[h]awkers filled the air with their chants’, that had become after a decade of communist rule a place where ‘no birds were left in the trees, which had been stripped of leaves and bark, their bare and bony spines standing stark against an empty sky’. ³⁷ This descriptive flourish is the author’s poetic summary of archival reports in his personal collection. (No witnesses to the noises or the silences are cited.) These reports document desolate social relations (‘a hungry man bartered his shoes at a train station for a parcel of meat that included a human nose and several human ears’) and corrupt practices of a morally bankrupt Party leadership (‘260 cadres spent four days working through 210 kilos of beef, 500 kilos of pork, 680 chickens, 40 kilos of ham, 130 litres of wine, and 79 cartons of cigarettes as well as mountains of sugar and pastries’). The reader is encouraged to draw a causal connection between the two. ³⁸

At times, Dikötter uses a sampling technique to produce what he claims are conservative estimates of the incidence of suffering amongst the PRC population as a whole. He calculates aggregates and ratios from samples of the documents in his private collection to show that ‘up to 40 per cent of all housing was turned into rubble’ (p. xii); that ‘a prolonged and intense attack on nature claimed up to half of all trees in some provinces’ (p. xii); that ‘the death toll thus stands at a minimum of 45 million excess deaths’ (p. 333); and that ‘at least 6 to 8 per cent of all the famine victims were directly killed or died as a result of injuries inflicted by cadres and the militia’ (p. 298). The latter figure of 6 to 8 per cent is extrapolated from the author’s hand-picked sample of five reports compiled by internal investigators in 1961 that describe the policy errors and resulting hardships endured by the populations of three exemplary counties. Dikötter’s sampling method is at least an improvement on that of Chang and Halliday, the popular biographers of Mao, who induced from one internal report about a single ‘not atypical county’ that near to a quarter of a million people had committed suicide in an unspecified but ‘short’ period of time in the 1950s. ³⁹

The juxtaposition and sampling techniques used by Dikötter fall short of academic best practice. More serious questions of scholarly standards arise out of the passages in Mao’s Great Famine that focus on Mao and his colleagues in Party Central. Dikötter fails to cite the secondary sources that helped him to locate important documents from amongst the many millions stored in Chinese provincial and sub-provincial archives, and to draw meaningful conclusions from those documents. In particular, he has a large unacknowledged debt to Yang Jisheng. Below, the citations given in one chapter, on the people’s communes, are compared with relevant passages and citations in Tombstone. This pedantic exercise in cross-referencing footnotes reveals a striking similarity between the archival documents discovered by Dikötter and those cited by Yang Jisheng in print two and a half years earlier.

The first three citations in this chapter are not found in Tombstone. The first of these is to Mao’s personal physician Li Zhisui, whose dressing gown-and-slippers perspective on politics supports Dikötter’s view that radical programmes such as the establishment of communes emanated out of the private life of the Chairman. (In contrast, Yang documents how the concept of the commune was developed through several months of collaborative work by a large team of senior Party leaders, propaganda agencies, and local Party officials.) Dikötter then cites a document mentioned by Yang in a list of ‘closed’ documents that Yang was unable to access during his visit to the Hebei provincial archives. Dikötter had better luck than Yang in these archives, but could have thanked Yang for the lead. The third footnote cites a report by the Minister of Agriculture that is referred to but not clearly cited by Yang. ⁴⁰

The next three footnotes all concern the mass irrigation policies trialled in Xushui county. In the first of these, the documents cited by Dikötter are cited by Yang and the specific passages quoted by Dikötter are also quoted by Yang; Dikötter’s original contribution lies in removing the economic planning context of the irrigation campaign highlighted by Yang. ⁴¹ In Tombstone, Yang proceeds directly to an overview of an article published in the Renmin ribao on 17 April 1958 that hails the success of Xushui; so does Dikötter, citing the same article and adding that it was written in a martial tone. ⁴² For the next footnote, the main text of Mao’s Great Famine refers to an article in Hongqi (红旗) published on 1 July 1958, but in the footnote a different article published in the subsequent 16 July edition of the bimonthly journal is cited. The source of Dikötter’s error, it would appear, is that Yang Jisheng had referred to both articles in a single paragraph. ⁴³ The error would have been avoided if Dikötter had cited Yang directly, rather than trying to cite the sources behind the actual source of his information.

Dikötter then discusses the Beidaihe meeting of August 1958, for which he cites speeches of Mao recorded in a document found in the Gansu provincial archives; Yang quotes the same speeches from the standard published collection Mao Zedong sixiang wansui (Long live Mao Zedong thought 毛泽东思想万岁). ⁴⁴ Dikötter quotes from Li Rui’s memoir that the Xushui first Party secretary confidently predicted the advent of the historical stage of communism by 1963; Yang cites Li Rui on the same anecdote, highlighting the essential context that the prediction was formulated in conjunction with a new long-term economic development plan that had taken several months to compile with the active involvement of senior central Party officials. ⁴⁵ A pattern emerges at this point, as Dikötter cites in succession: a Renmin ribao article, cited by Yang; a passage in the official biography of Liu Shaoqi, cited by Yang; two Renmin ribao articles, both cited by Yang; a comment by Mao in a standard document collection, cited by Yang; an English-language secondary source, not cited by Yang; and two of Mao’s speeches that are sourced to documents held in the Hunan provincial archives, whereas Yang quotes from the same speeches citing Mao Zedong sixiang wansui. Dikötter then concludes his discussion of the high-level machinations behind the people’s communes with some illustrations of what communization meant in practice. He cites survey materials relating to the militarization of village life, the introduction of work points in place of money, the collectivization of private property, the establishment of urban communes, and so forth. In this impressionistic conclusion to the chapter, the archival sources are all his own.

A similar list of unlikely coincidences could be drawn for several other chapters that describe the genesis of Great Leap policies. In these chapters, Dikötter’s contribution to scholarship is to strip Yang’s archival discoveries and synthesis of published material of the historical context provided by Yang, and to rearrange the resulting fragments into an idiosyncratic vignette of totalitarian folly.

Dikötter mentions Yang Jisheng’s name several times in the main text and once in the notes of his book, but never to acknowledge a substantial scholarly debt. Instead, he launches an extraordinary assault on Yang’s work in the bibliographic essay that concludes the book:

[Yang Jisheng’s] work remains important, in particular in so far as few other historians have been able to research and publish on the famine in Henan province. But … the book is more of a compilation of notes from different sources than a carefully constructed text. At times it looks like a hotchpotch which simply strings together large chunks of text, some lifted from the Web, a few from published sources, and others transcribed from archival material. Invaluable documents are thrown together with irrelevant anecdotes, making it difficult for the reader to see the wood for the trees. In some cases the author spent only a day or two in the archives, missing the most vital, and openly available, documents. This is the case for the chapter on Guangdong, which relies on a single file for the entire famine. ⁴⁶

The allegations Dikötter levels at Yang here are bewildering. They are either sloppily drawn or disingenuous. Henan does feature prominently in Tombstone, but is the subject of only one of eleven provincial case studies. Yang at times redacts published and unpublished secondary sources, and acknowledges these appropriately. The secondary sources Yang draws upon are supplemented with documents from the Party committee section of provincial Party archives, and by interviews with retired Party officials. He does not rely on one single archival file for the entire chapter on Guangdong but four, as well as several substantial secondary sources. In instances noted by this reviewer where Yang has used material from the Web, this material is not ‘lifted’ but appropriately acknowledged, if not always in an orthodox format.

The more substantial charged levied by Dikötter against Yang is that he has no point. Dikötter writes, continuing from the passage quoted above,

But most of all there is no time line: by dispensing with a meaningful historical narrative and focusing heavily on grain shortages, the author misses an important dimension of the disaster. ⁴⁷

It is true that the Great Leap famine was not just about grain shortages, but there are good reasons for a historical account of a famine to linger on problems of food supply. The important dimension Yang missed, one presumes, is the moral dimension of totalitarianism evil, though Yang does make a clear argument about the contribution to the famine of authoritarian governance structures. To this reviewer, Dikötter’s criticism of Tombstone as a poorly organized and poorly researched piece of work reads like a tenuous justification for appropriating its content as his own, and reorganizing its material to conform with his own preferred meta-narrative.

There is one important document cited in Mao’s Great Famine that cannot be found readily in the footnotes of Tombstone or of any other published book or article. This is a document presented by Dikötter as the minutes of a speech delivered by Mao at a Politburo meeting convened in Shanghai in March 1959. Dikötter is proud of his discovery: he quotes from the document no fewer than six times in his book, again when interviewed for a recent television documentary on the famine (the camera scrolls over two yellow pages of archival material), ⁴⁸ and in his response to a negative review published in The China Journal, where he writes,

One fact is that Mao and other leaders knew what was happening in the countryside as a result of the Great Leap Forward. In the case of Mao, the smoking gun is in the minutes of a meeting that took place in the Jinjiang Hotel in Shanghai on 25 March 1959, when Mao ordered that procurements be increased to one-third of all the grain, and made available an extra 16,000 lorries to carry out the task. “When there is not enough to eat,” the Chairman explained, “people starve to death. It is better to let half of the people die so that the other half can eat their fill.” Detailed evidence of this nature clearly undermines the widespread view that Mao, in the many months before the Lushan Plenum in the summer of 1959, “defended the peasants”. ⁴⁹

The established scholarship on PRC elite politics contends that there was a distinct ‘cooling down’ phase of the Great Leap Forward between November 1958 and July 1959, during which Mao championed a series of moderate measures that supported the welfare of the rural population. ⁵⁰ These measures were a response to the central government’s ‘nine fingers and one finger’ evaluation of the first summer of the Great Leap Forward (i.e. the campaign was nine parts successful, with one part marred by radical excess). Dikötter forcefully rejects this view, asserting as ‘fact’ the alternate view that Mao remained a steadfast proponent of radical measures throughout the Leap era. ⁵¹ The evidence he cites is all in the one archival document which, according to Dikötter, proves that Mao pushed for increased procurement targets, mobilized a large fleet of lorries to achieve these targets, and expressed a willingness to sacrifice half the population of China. Before accepting Dikötter’s view as truth, it would be helpful to know more about the context, rhetorical intent and effect of what Mao said in Shanghai.

Some further information can be gleaned from the documentary collection edited and translated by Zhou Xun, Dikötter’s long-term collaborator and colleague at Hong Kong University. ⁵² This book presents abridged translations of 121 archival documents arranged according to eight themes (e.g. ‘Famine in the Communes’, ‘Terror, Repression, and Violence’), each with a short introduction. The documents are mostly internal Party reports of the type used heavily by Yang and Dikötter and discussed in the concluding section of this review, but also include records of speeches at internal Party meetings and letters addressed to senior Party leaders. Scant information is provided on the contexts in which the documents were produced. The translated documents themselves are peppered with ellipses […], indicating passages that have been excised. Though Zhou explains that the excisions were necessary because some of the original documents were too long to be reproduced in full, she does not explain the principles she used in deciding what to include.

A translation of the document cited by Dikötter containing Mao’s ‘speech’ in Shanghai is presented as Document 5 of the collection. ⁵³ Zhou says in her introduction that the full text of this document is reproduced, though the 18 ellipses in the translated document indicate that this is not the case. The title of the document is translated by Zhou as ‘Chairman Mao’s words at the Shanghai Conference, March 25, 1959’. Dikötter had described the document as a record of his ‘speech’ rather than ‘words’, a small difference but one which marks a distinction between impromptu remarks and prepared speech. The next line of the document states that the text which follows is a record of Mao’s ‘[comments] during Comrade Li Xiannian’s report on the current trade and financial situation, as well as further suggestions’, indicating that the ‘words’ were indeed in the form of impromptu remarks.

Dikötter’s claim that Mao ordered ‘procurements be increased to one-third of all the grain’ stems from the following passage of the document, rendered by Zhou in the form of a conversation between Li Xiannian and Mao Zedong,

[Li]: […] It is crucial for all Party members to put their efforts into achieving the [procurement] targets.

[Mao]: By this we mean provincial, regional, and county Party committees. As long as the amount of grain being procured does not go above a third [of grain produced], peasants will not rebel. ⁵⁴

Mao appears to be supporting the fulfilment of ambitious grain procurement targets for the second quarter of 1959 that had been set several months earlier, despite difficulties experienced in achieving similarly ambitious targets set for the first quarter. There is nothing to suggest he advocated raising targets to a new level or increasing the ratio of the harvest procured by the state, as claimed repeatedly by Dikötter. ⁵⁵ The one-third ratio referred to here was in fact nothing new. The state justified increases in grain procurements through the 1950s not by raising the tax rate but by artificially inflating the reported grain production figure, leaving the ratio of procurements to reported production at a relatively stable level of around one-third between 1953 and 1960. ⁵⁶ The reported grain production figure for 1958 demonstrated particular dexterity, serving several different administrative ends. The 1958 harvest figure was first reported as 100 per cent above that of 1957, a full endorsement of the Leap programme; it was revised down in early 1959 to a 50 per cent increase, encouraging some restraint amongst the procurement teams; and was revised down further at the conclusion of the Leap campaign to show a (still spectacular) 6 per cent increase, suggesting that if the Leap programme was not miraculous it was at least seven-tenths correct. We simply do not know the relationship between any of these production figures and how much grain was actually produced.

The decision to deploy a fleet of lorries, excised out of Zhou’s translation, might be understood as acknowledgement that the grain collection agencies had worked the agricultural areas proximate to rail and water transport infrastructure to its limit, requiring them to forage over a wider area to achieve their targets. This also suggests the restoration of a certain baseline of support for industry after a first quarter in which most plan targets were under-fulfilled. The measures introduced in Shanghai indicate neither a pro-peasant bias to Party policy nor the extreme pro-industrial radicalism of the previous summer, but are comfortably within the parameters of the ‘nine fingers and one finger’ programme. ⁵⁷

This leaves the jaw-dropping comment by Mao that suggests he was willing to sacrifice half of China’s population for the Great Leap campaign. The comment brings to mind Mao’s speech to the Moscow Conference of Communist Parties in November 1957, in which he mused on a future nuclear war that might kill off half of the world’s population, destroying capitalism and leaving the surviving half socialist. ⁵⁸ Dikötter claims that Mao in Moscow was just sabre-rattling, but that his words in Shanghai carried a deadly serious intent. On the Moscow Conference, his interpretation is loosely consistent with one careful archive-based study of the episode. ⁵⁹ On the Shanghai meeting, he would like his readers to disregard other scholarship and accept on faith his own interpretation of an archival document to which he has privileged access.

Zhou’s translation clarifies a number of points. It shows that Mao’s comment was made in response to Bo Yibo’s report on the implementation of the industrial plan. This report did not cover questions of food supply, which came under Li Xiannian’s brief. ⁶⁰ The comment is preceded by several remarks by Mao about Party oversight of the industrial sector, none of which touch upon agriculture or rural welfare. However, exactly what Mao meant by ‘It is better to let half the people die so that the other half can eat their fill’, is rendered ambiguous by a significant omission in the translation.While writing this review article, an incomplete copy of the very same document was provided to me by a senior Party historian. ⁶¹ It is clear from reading the original document that it is not a record of Mao’s prepared speech (讲话) made on 25 March, but rather his impromptu interjections (插话) made in response to separate reports by Bo Yibo and Li Xiannian, probably when the two reports were formally discussed in the Politburo meeting on 26 March and 28 March respectively. ⁶² Zhou’s translation omits a passage immediately preceding the comment about letting half the people die, without marking the excision with an ellipsis. The omitted passage introduces the policy initiative of reducing the number of major capital construction projects, so that more resources could be allocated in support of the core aims of the plan. ⁶³ Mao comments,

If we want to fulfil the plan, then we need to greatly reduce the number of projects. We need to be resolute in further cutting the 1,078 major projects down to 500. (要完成计划，就要大減项目。1078个项目中还应該堅決地再多削減，削到500个。)

To distribute resources evenly is a way to sabotage the Great Leap Forward. (平均使用力量是破坏大跃进的办法。)

If all are unable to eat their fill, then all will die. It is better for half to die, so that half of the people can eat their fill. (大家吃不飽，大家死，不如死一半，給一半人吃飽。) ⁶⁴

The omitted passage might relate to a dry detail of the economic plan, but in this instance the detail is crucial to the interpretation of what follows. The ‘people’ whom Mao was willing to let die of starvation turn out to be not people at all, but large-scale industrial projects. Rather than pushing to extract more resources out of the countryside to feed industry, here Mao is calling for the rational use of scarce capital and other industrial inputs. Mao’s comment at the Shanghai Conference about letting half the people die does not tell us anything about Mao’s views on grain procurement, agriculture or the peasantry, and does not take us outside the ‘nine fingers and one finger’ programme. Dikötter’s maverick interpretation of the Shanghai Conference not only blithely ignores the substantial commentary on the conference by other scholars and several of its key participants, including Bo Yibo, but defies the very plain wording of the archival document in his possession on which he hangs his case. ⁶⁵

Though the smoking gun presented by Dikötter was smoking for the wrong reasons, there is evidence that Mao knew of widespread suffering associated with the Great Leap programme and pressed on regardless. Yang Jisheng demonstrates that reports on problems associated with the Leap programme, including severe food-supply problems, were presented to the central Party leadership from the autumn of 1958, and Mao responded in person to many of these. ⁶⁶ The responses were in the form of tactical readjustments, and the ‘three red flags’ (the General Line, the Great Leap Forward, and the commune system) that Yang presents as the fundamental causes of the famine remained beyond reproach. Yang describes the Shanghai Conference as at the tail end of a round of collective self-criticism, where the focus of the Party leadership shifted from past mistakes in agriculture to future achievements in industry, and vocal critics of the Leap programme were given notice that their criticism would not be tolerated for much longer. ⁶⁷

From the village to the nation: A question of statistical method

The works of Yang and Dikötter invite similar questions as those raised by Walker in 1962. How reliable and representative are the available sources concerning the famine, especially the internal investigation reports that mark the main point of departure of the new archival history from the loose consensus of the 1990s?

Most of the official reports cited by Yang and Dikötter are examples of the ‘typical survey’ (典型调查). This is a specific type of numerically laden report that was a key tool of mass mobilization used by the CCP from the 1920s. ⁶⁸ From 1956, and especially in 1958 and 1959, the typical survey was promoted as an alternative to the full enumeration censuses that had been the lifeblood of the state statistical agencies. ‘Typical surveys are an important road by which cadres engaged in statistical work can become red experts,’ reads the title of an article published in Zhongguo tongji (China statistics) in March 1958. ⁶⁹ Another article published in the March 1960 edition of the Sichuan-based journal Caijing kexue (Financial science) explains that ‘Typical surveys are an innovative survey method utilized in our nation under particular historical and social conditions; it is the concrete manifestation in statistical survey methodology of Chairman Mao’s synthesis of Marxist–Leninist theory with the practice of revolution in China.’ ⁷⁰ The important point here is that typical surveys were to serve the political needs of the Party, generating information in a timely manner to support the practical tasks of the political leadership.

Mao’s surveys compiled in the late 1920s on the revolutionary movement in Hunan and Jiangxi were promoted in the late 1950s as exemplary models of the typical survey; the prefaces written by Mao for a collection of agricultural surveys published in Yan’an in 1941 defined their ideological purpose and orthodox methodology. In the latter Mao wrote, as quoted in the same article in Caijing kexue, ‘The method of gathering materials is to assemble a group of experienced people for a survey meeting, with from 3 to 5 up to 7 or 8 people. The survey is a typical survey of one problem in one administrative village, one district, one county, one city, one township, one army, one battalion, one factory, one shop or one school.’ ⁷¹ That is, a good typical survey had a sample size of one, filtered all information through the local cadres in charge (the ‘experienced people’), and directed that information towards the resolution of one specific political problem. The typical survey of the Great Leap era came with a forceful rejection of the notion that social surveys should be constrained by questions of representativeness, for the task of the red expert was not to describe society as it was but to affirm a constructed vision of what society should be. ⁷²

While the typical survey was praised in the late 1950s as a Chinese invention – one of the many new practices through which the CCP distanced itself from Stalinism in the wake of Khrushchev’s secret speech of February 1956 – it actually bore a close resemblance to the investigatory reports produced by the Soviet security and propaganda agencies in support of the mass mobilization campaigns of the high Stalin era (1928–37). The many mood surveys, collections of letters to Stalin and reports recording the words and actions of ‘typical’ workers, peasants and folk inhabiting the grass roots of Soviet society that became available from the 1970s, inspired a whole genre of writing on everyday life in the Soviet Union. While for a few years in the 1990s it seemed that Soviet society under Stalin was largely constituted by reactionaries and counter-revolutionaries, it has been acknowledged by at least one of the leading historians of Soviet everyday life that much of the classified archive, especially the reports produced by the security agencies, tended to overplay certain forms of anti-Soviet behaviour. ⁷³ The Chinese statistical agencies during the Leap era appear to have gone a step further in embracing the typical survey than their Soviet counterparts did a generation earlier. For the years 1958 and 1959, it is doubtful whether any reliable statistical data for items at the heart of Great Leap mobilization effort such as grain production and mortality were compiled above the county level. This further confounds the efforts of historians to discover a reliable core to the Party archives.

To question the reliability of the evidence of an atrocity is not, as Dikötter alleged, to replicate the view of the perpetrators. ⁷⁴ The famine was a catastrophic affair, a fact not disputed by any credible historian. What is at stake here is whether we can understand the famine independently of how the CCP came to terms with it during the two brief moments, in the early 1960s and again in the early 1980s, when the Party had an interest in reflecting critically on its own past.