Chinese collective memory on the Internet: Remembering the Great Famine in online encyclopaedias

Abstract

Recent research on how the Great Chinese Famine was debated on Weibo, the Chinese equivalent of Twitter, in 2012 suggests that information and communication technologies can challenge official versions of the past and increase pluralism in collective memory narratives in authoritarian states. This article suggests that analysing change in the treatment of the famine in Chinese online encyclopaedias during and following the debate helps us further explore the debate’s impact. Moreover, it allows us to determine the extent to which Chinese online encyclopaedias function as the type of memory place that previous research on Wikipedia in other contexts might lead us to expect. The article concludes that the changes made to the narratives about the Great Famine in Chinese online encyclopaedias following the debate were rather limited and that the Chinese online encyclopaedias have not yet developed into participatory and pluralistic memory places that challenge official narratives.

Introduction

Research on Chinese online collective memory has largely been dominated by analyses of how specific events are remembered in particular websites such as online museums. In this sense, it has mirrored research on online collective memory more generally. Studies of Chinese online collective memory have, for example, analysed how narratives about past events, such as the Cultural Revolution, are remembered in blogs (Yang, 2007, 2010; Zhang, 2012). One exception to this general preoccupation is Zhao and Liu’s analysis of how the Chinese Great Famine of the late 1950s and early 1960s was debated on Weibo, the Chinese equivalent of Twitter, in April and May 2012, in ways that challenged the official narrative and presented alternative ones (Zhao and Liu, 2015). While Zhao and Liu (2015) suggest that information and communication technologies (ICTs), such as Weibo, have the potential to increase pluralism in collective memory narratives by challenging official versions of the past, the long-term impact of the debate on Chinese online collective memory is not entirely clear. From the perspective of research that conceptualises Wikipedia as a memory place (e.g. Pentzold, 2009), one would arguably expect to see major changes to the entries on the famine in Chinese online encyclopaedias if the debates significantly altered Chinese collective memory of the Great Famine. This article, thus, suggests that analysing changes in the narratives that appear in the longer lasting and influential collective memory repositories constituted by online encyclopaedias is a fruitful way to evaluate the effects of such online debates. Such an analysis can provide indications concerning the extent to which these technologies of memory have altered the opportunities in China for collective memory construction that challenges official narratives. Consequently, the article seeks to address the following question: How has the depiction of the Great Famine in Chinese online encyclopaedias changed following the debate about the famine on Weibo?

In addition to providing insights concerning specific issues related to the construction of the collective memory of the Great Famine, the case study contributes to the more general literature on Internet-based collective memory, and in particular on online encyclopaedias as memory places. This literature argues that Wikipedia can be understood as a global memory place (e.g. Ferron, 2012; Ferron and Massa, 2014; Luyt, 2016; Pentzold, 2009). Several observers have suggested that the Internet has the potential to become a new arena for the promotion of alternative memory narratives and for the contestation of official views (Haskins, 2007; Hess, 2007; Hughes, 2012; Marschall, 2013). Scant attention has been paid to authoritarian contexts such as China, however, where Wikipedia has often been blocked and domestic alternatives have become influential (Liao, 2015). Consequently, we currently lack knowledge about the extent to which Chinese online encyclopaedias function as the kinds of memory places that previous research on Wikipedia in other contexts might lead us to expect. Investigating possible changes to the narrative on the Great Famine in Chinese online encyclopaedias during and following the 2012 Weibo debate allows us to shed light on these issues. This article thus seeks to answer a second question: How do online encyclopaedias operate as memory places in China? The article raises doubt about the extent to which such Chinese participatory memory places really lead to the construction of alternative collective memories that challenge dominant government narratives. In addition, it raises questions about the extent to which Wikipedia can be understood as a truly global memory place.

The article first provides a theoretical discussion of online collective memory and an account of the Chinese context. A brief case study follows the treatment of the Great Famine in Chinese online encyclopaedias. The concluding section discusses Chinese online encyclopaedias as memory places against the background of the results of the analysis and makes suggestions for further research.

Online collective memory

Research on Internet-based memory and remembrance has become increasingly common in recent years (e.g. Bernstein, 2015; Garde-Hansen et al., 2009; Kaprāns, 2016; Maj and Riha, 2009; Van Dijck, 2007). When approaching this literature, it is possible to distinguish between research that deals with online remembrance, on one hand, and research focused on the Internet as a technology of memory, on the other. Online remembrance can be said to involve the creation of particular sites for remembrance purposes, for example, online museums (e.g. Foot et al., 2006; Haskins, 2007; Hess, 2007; Marschall, 2013; Walker, 2007; Zhang, 2012). Many such sites take the form of digital archives that are not necessarily collectively constructed and that do not necessarily receive large numbers of visitors. They might thus be described as sites for ‘collected memories’ rather than ‘collective memory’ (Arthur, 2008; Ferron and Massa, 2014: 28; cf. Olick, 1999).

The Internet as a technology of memory, by contrast, arguably has much broader implications. For example, Viktor Mayer-Schönberger (2009) has argued that the Internet has fundamentally altered our capacity to forget. Whereas forgetting used to be the default setting, the Internet has changed this to remembering (pp. 48–49). Nonetheless, despite the fact that much is now stored online, not every event is remembered by societies and all descriptions of past events are not equally influential. There are good reasons for believing that some online representations dominate the virtual memoryscape. This is in part because of how search engines function:

A small percentage of users set preferences to more than 10 results per page; typically they do not look past the first page of results; and they increasingly click the results appearing toward the top. Thus the power of a search engine lies in the combination of its ranking policies (source inclusion toward the top results) together with the users’ apparent ‘respect’ for the orderings (not looking further). (Rogers, 2013: 31; see also Curran, 2012b: 44–45)

This means that the alternative narratives transmitted, for example, by activist groups can get ‘lost on the web … partly because they tend to get a low search engine listing’ (Curran, 2012a: 14). In addition, search engines, such as Google, have tended to privilege dominant news providers at the expense of alternative news sources (Curran, 2012a: 19). In other words, even though much is archived online, only a small proportion is canonised. In this way, much of what is posted online is backgrounded to such an extent that it is practically consigned to online oblivion. In this sense, it might be said that much is ‘stored but not remembered’ (Pentzold, 2009: 260), suggesting that many of the specific memory sites focused on in existing research may in fact not be among the more influential ones. In contrast, online encyclopaedias, such as Wikipedia and its Chinese equivalents Hudong Baike and Baidu Baike, typically appear at the top or near the top of web searches (Jiang, 2014; Lewandowski and Spree, 2011; Rogers, 2013: ch. 8). On the Internet, ‘there are only few sites, among them Wikipedia, that receive most of the attention’ due to links and search algorithms (Pentzold, 2009: 263). Because of the prominence of online encyclopaedias in search returns, many users thus get their information about historical events from such sources.

Wikipedia is known as the ‘free encyclopaedia that anyone can edit’ and has consequently been approached by scholars as an important site for knowledge production in general (e.g. Bilić, 2015; Haider and Sundin, 2014; Luyt, 2015). Nonetheless, as Christian Pentzold has argued, Wikipedia can also be theorised as a global collective memory place (Pentzold, 2009; see also Ferron and Massa, 2014; Luyt, 2016). Pentzold’s theorisation is based on four main building blocks: Maurice Halbwachs’ (1992) conceptualisation of collective memory, Jan Assmann’s (2008) distinction between communicative and cultural memory, Pierre Nora’s (1996) notion of memory places and the ‘floating gap’, as understood by Jan Vansina (1985). For Halbwachs, human memory is fundamentally social – humans remember their pasts as members of particular collectives. Such groups construct a collective memory through a communicative social process, which is selective. Importantly, ‘the past is not preserved but is reconstructed on the basis of the present’ (Halbwachs, 1992: 40). Assmann (2008) distinguishes between communicative and cultural memory, both of which, as in Halbwachs’ theorisation, are linked to a particular collective. Because communicative memory is based on everyday communication between group members, it has a limited time frame of only about 80 years, that is, three interacting generations (Assmann, 2008). Cultural memory, by contrast, is ‘exteriorised, objectified, and stored away in symbolic forms’ (Assmann, 2008: 110), such as ‘monuments, museums, libraries, archives, and other mnemonic institutions’, and is thereby ‘formalised and stabilised’ to a greater extent (Assmann, 2008: 111). Vansina’s work on oral societies has identified a gap between the generational communicative memory of the relatively recent past and the remote past, which is the ancient history and mythical origin of the group. This gap shifts with generations, and Vansina (1985) therefore refers to it as the floating gap. Assmann argues that the floating gap concept is useful for understanding the difference between communicative and cultural memory beyond oral societies. This is because ‘even in literate societies living memory goes no further back than eighty years after which, separated by the floating gap, come, instead of myths of origin, the dates from schoolbooks and monuments’ (Assmann, 2008: 113). Nora (1996) understands memory places as either material or non-material symbolic places that help groups remember the past, in many cases through some kind of archival function. These groups can be abstract collectives that are not necessarily confined to space and time by, for example, national boundaries. These points are important for Pentzold’s conceptualisation of Wikipedia as a global memory place. He argues that the World Wide Web, and especially Wikipedia, links cultural and communicative memory and thereby overcomes the floating gap because it is a symbolic place that makes possible both archiving and communication with multiple dialogue partners (2009: 262–264). Pentzold views Wikipedia as a memory place because it ‘functions as a platform where authors with divergent (national, cultural, religious, etc.) backgrounds can engage in an intense process of discursive knowledge constitution’. He sees it as ‘potentially “global” because it is accessible via the Internet and therefore not bounded by national frontiers’ (2009: 264; cf. Mayer-Schönberger, 2009: 61). Wikipedia’s talk pages also make possible debates over the discursive construction of the past between article editors. In this way, cultural memory is negotiated collectively as ‘different perspectives’ engage in ‘opposition and contestation’ over how to remember the past (Pentzold, 2009: 265).

Significantly, collective memory in Wikipedia differs from that constructed in other sites, such as online museums and archives: ‘Contrary to the individual character of many digital archives expressly devoted to commemoration, Wikipedia is not a fragmentary collection of disconnected pieces of information. Different stories, sources, interpretations, and points of view are selected and organised into coherent narratives’ (Ferron and Massa, 2014: 41).

A number of studies have suggested that the Internet has the potential to become an arena for the promotion of alternative memory narratives that could challenge official accounts (Haskins, 2007; Hess, 2007; Hughes, 2012; Marschall, 2013; for a case study of Cambodian history in Wikipedia, which indicates that this is not always the case, see Luyt, 2013). Whereas modern technologies of memory, such as museums and school curricula, have tended to prioritise official narratives about the past in order to construct national identity in ways that lend legitimacy to the state, the Internet makes it possible for a larger number of voices to participate in the production of collective memory (Haskins, 2007: 402–08). More specifically, understood as a global memory place, Wikipedia moves beyond the scope of traditional memory technologies such as museums, which have typically been tools of national memory construction (Ferron, 2012; Ferron and Massa, 2014; Pentzold, 2009). In contrast to traditional, top-down offline or analogue technologies of memory, Wikipedia has been described as having a bottom-up participatory culture, meaning that ‘memories are not anymore just consumed but also creatively produced in a participatory and decentralised’ and ‘de-territorialised’ way (Ferron and Massa, 2014: 25; cf. Pentzold, 2009: 264):

Consequently, wikis present a collaborative open content system and mark an important step in fulfilling the promise of the Internet to challenge the biased production and distribution structures of the mass media and the asymmetrical relationship between the producer and recipient of media messages. Wikis and Wikipedia … provide alternative patterns of knowledge production through online cooperation. (Pentzold, 2009: 257)

It is, thus, suggested that through online encyclopaedias the emancipatory promise of the Internet to make global, bottom-up cultural practices possible could be fulfilled in the realm of collective memory. Based on the analysis that follows below, this article questions the extent to which this is the case.

Online collective memory with Chinese characteristics

In China, the government largely controls the construction of offline collective memory sites. Through comprehensive programmes of ‘patriotic education’ (aiguozhuyi jiaoyu) and ‘red tourism’ (hongse lüyou), the Chinese Communist Party (CCP) has invested huge resources in its attempt to firmly control the Chinese memory infrastructure and by extension collective memory (Gustafsson, 2011; Wang, 2012; see also Zhongxuanbu, 1998; Zhongguo Hongse Lüyou Wang, 2005). The government views alternative or foreign versions of the past as potential threats to its citizens’ identities and its own legitimacy (Gustafsson, 2014a). Traditional technologies of memory such as physical museums are costly to run and therefore difficult for non-state actors to establish. This is especially true in China, where in 2009 the government decided that patriotic and other education sites, to which many modern history museums belong, would be open to all free of charge. At the same time, government-run museums are being built in large numbers in provincial cities across China. This contributes to making it difficult to operate private history museums (Gustafsson, 2011: 119).

The CCP’s way of depicting the past has involved constructing its own image in a way that strongly emphasises its patriotic identity. This has made it possible for other actors in Chinese society to exercise power over the CCP to some extent by questioning how patriotic certain policies are, thereby limiting the CCP’s foreign policy options (Gustafsson, 2014b). Nonetheless, the construction of memory narratives through traditional technologies of memory has largely been controlled by the CCP. This fact notwithstanding, the potential offered by the Internet for de-territorialised, bottom-up, participatory collective memory construction, said in previous research to exist in other settings, suggests that similar processes of collective memory creation might be possible in China too. This article explores the possibilities of online encyclopaedias as memory technologies in the Chinese context. Even though some studies have dealt with online memory in China (e.g. Li, 2009; Zhang, 2012; Zhao and Liu, 2015), the topic has received relatively little attention, especially in contrast to the rather large body of work dealing with Chinese offline remembrance (e.g. Denton, 2015; Gustafsson, 2011; Wang, 2012). In contrast, much of the research on online collective memory has focused on Western cases (e.g. Foot et al., 2006; Hess, 2007; Walker, 2007).

In China, Wikipedia has been blocked intermittently due to its refusal to comply with the CCP’s censorship policies. Blocking of Wikipedia has not been monolithic in that the site has at times been accessible in one location while blocked elsewhere. ¹ It has been argued that this sporadic blocking contributed significantly to restricting the growth in the number of Chinese users of and entries in the Chinese-language version of Wikipedia, thereby aiding its Chinese domestic competitors (e.g. Liao, 2015). The Chinese online encyclopaedias Hudong Baike and Baidu Baike, which are similar to Wikipedia in many ways, have both become tremendously influential in China.

Wikipedia is now available in 291 language versions. English Wikipedia contained more than 5 million articles as of late May 2016. It has been argued that Wikipedia’s size makes it ‘an interesting platform for empirical studies of social processes on a large scale’ (Ferron and Massa, 2014: 26). Chinese Wikipedia had more than 880,000 articles as of late May 2016 (Weiji Baike, 2016a). Around the same time, however, Baidu Baike contained more than 13,500,000 (Baidu Baike, 2016) and Hudong Baike more than 14,775,000 articles (Hudong Baike, 2016). Baidu Baike and Hudong Baike are thus not only much larger than Chinese Wikipedia in terms of the number of articles but also considerably larger than English Wikipedia.

Chinese Wikipedia also differs from the other Chinese-language online encyclopaedias in several important respects. For example, unlike Wikipedia, the domestic Chinese online encyclopaedias are for-profit companies that operate in accordance with the Chinese government’s censorship regulations. Wikipedia, by contrast, refuses to do this. In addition, a comparative study of Baidu Baike and Chinese Wikipedia has shown that the former is in effect mainland-centric, while entries in Chinese Wikipedia are created by a considerably more transnational group of users. One important reason for this is technical. Baidu Baike uses a script that only supports simplified Chinese, thereby excluding the possibility of making entries using the type of Chinese characters used beyond the mainland, for example, in Taiwan and Hong Kong. Chinese Wikipedia, by contrast, uses a script that supports all languages, making it possible for Chinese-language users in different societies that use different scripts to contribute entries that are then converted according to user preferences. Users can thus choose which Chinese script they want Wikipedia entries to be displayed in: Mainland simplified, Hong Kong traditional, Macau traditional, Singapore/Malaysia simplified or Taiwan traditional. In addition, Chinese Wikipedia’s policy stipulates that when making entries, contributors should avoid region-centrism and strive for neutrality. This policy in effect treats Mainland China as one region among several Chinese-speaking regions (Liao, 2015).

Comparative research on Chinese Wikipedia and Baidu Baike analysed the place of residence disclosed by ‘power users’ or administrators who, for example, are tasked with overseeing the contributions to the encyclopaedias. It found that almost all such users on Baidu Baike are located in Mainland China, whereas there is much more diversity in the distribution of such users of Chinese Wikipedia, with many being located in Taiwan and Hong Kong (Liao, 2015: 40). Chinese Wikipedia can thus be seen as enabling the construction of a transnational memory place in the Chinese language where no region is given privileged status (cf. Pentzold, 2009). Because it only supports the simplified Mainland Chinese script, Baidu Baike, by contrast, largely keeps non-Mainland Chinese out, effectively making the online encyclopaedia mainland-centric and national rather than transnational (Liao, 2015). This is likely to result in an online Mainland Chinese national collective memory, rather than one constructed in collaboration and through negotiation with ethnic Chinese beyond the borders of the People’s Republic of China. Given that statements by Chinese leaders, as well as documents published by Chinese government bodies, have long warned of ‘hostile international forces’ that seek to ‘infiltrate’ China, especially its younger generations, ‘ideologically and culturally’, Baidu Baike’s policies are perhaps not coincidental but might instead be part of security policies meant to protect China’s youth and by extension the regime itself (cf. Gustafsson, 2014a: 74–75). This interpretation is supported by an internal document leaked by a Baidu employee, which outlines censorship guidelines, including the specific lists of topics and key words to be monitored, blocked or censored, such as ‘reactionary’ (fandong) content, posts related to ‘state leaders’ (guojia lingdaoren) and posts on ‘ethnic problems’ (minzu wenti) (China Digital Times, 2009).

Remembering the Great Famine in Chinese online encyclopaedias

The Great Famine coincided with and is often regarded as a direct result of policies adopted during the Great Leap Forward. Some accounts describe the famine as having occurred between 1959 and 1961, while others believe it lasted from 1958 until 1962. Estimates of the number of excess deaths also differ. Journalist and author Yang Jisheng (2012: ch. 11), for example, estimates 36 million excess deaths, whereas historian Frank Dikötter (2010) puts the figure at 45 million (pp. 324–334).

The Great Famine has been given scant attention in official Chinese historical narratives and commemoration. In recent years, however, local histories of the famine and memoirs by party officials have been published, and internal party reports have been uploaded to the Internet (e.g. Garnaut, 2013: 225). The publication of popular books, such as Yang Jisheng’s Tombstone (Mubei) in 2008 and Frank Dikötter’s Mao’s Great Famine in 2010, has arguably also had an impact on popular remembrance of the Great Famine. The Folk Memory Project documentary film initiative has also contributed to a greater awareness of the Great Famine by both making and screening documentary films and through the activities of its participants which have included erecting memorial tablets commemorating the victims in rural areas (Pernin, 2014; Wu, 2014).

The case study presented here does not seek to provide an exhaustive account of what has influenced the memory of the Great Famine in China. Instead, its scope is limited to determining how the narrative on the Great Famine in Chinese online encyclopaedias changed during and following the debate on Weibo, which took place between 29 April and 2 May 2012. Such an analysis allows us to determine whether the debate had the effects that might be expected if the Chinese encyclopaedias do in fact function as the kinds of memory places that previous research on Wikipedia in other contexts would lead us to expect (Ferron, 2012; Ferron and Massa, 2014; Luyt, 2016; Pentzold, 2009) or whether such online memory places function differently in China. Limiting the study primarily to the possible impact of the Weibo debate is warranted for several reasons. First, while the books on the Great Famine and the documentary film project have been around for some time, the Weibo debate was of much shorter duration, making it easier to study its possible impact. Second, Zhao and Liu (2015) argued that few Chinese were familiar with Dikötter’s book prior to the debate on Weibo and that the debate led to increased dissemination of Dikötter’s and Yang Jisheng’s books, both of which had been banned in Mainland China. While the case is admittedly limited in scope, it nonetheless produces tentative conclusions that future research can build on.

Online contestation in Mainland China is sometimes depicted in terms of citizens pitted against the censoring authoritarian state, but it has been demonstrated that what happens on the Internet in China is more complex than that. Because of the limitations of censorship, the party-state also seeks to ‘guide’ online opinion. It therefore pays Internet commentators to make pro-regime posts online. However, the use of such comments is not necessarily successful as the posts are often dismissed as propaganda and the commentators derogatively labelled part of the ‘fifty-cent party’ because they reportedly receive 50 cents per post (Han, 2015a). Nonetheless, in addition to those who are paid to do so, there are also those who voluntarily defend the authoritarian regime (Han, 2015b). In other words, the Mainland Chinese online sphere is one where groups of netizens engage in a discursive struggle to promote their preferred versions of the past and visions of China’s future (cf. Gustafsson, 2016). It is in this context that online collective memory is constructed in China. Reports suggest that the ‘fifty-centers’ have been involved in editing entries in both Chinese Wikipedia and Baidu Baike. Entries related to history and politics are the most controversial and cause the most protracted arguments between fifty-centers and their opponents (Global Voices Advox, 2012).

The 2012 debate on Weibo is an example of the kind of contestation described above. Zhao and Liu point out that during the debate, participants increasingly labelled the object of discussion the ‘Great Famine’ (dajihuang) instead of the ‘three years of economic difficulty’ (sannian jingji kunnan) or the ‘three-year natural disaster’ (sannian ziran zaihai), which are the terms typically used in official narratives (Zhao and Liu, 2015). Hudong Baike still labels the events the ‘three-year period of difficulties’ (sannian kunnan shiqi), whereas English Wikipedia calls it the ‘Great Chinese Famine’. Baidu Baike uses the term ‘three-year natural disaster’. There was no change in the labelling before and after the debate. However, Baidu Baike has locked the title of the article, making it impossible for users to edit. Even Chinese Wikipedia still labels the events the ‘three years of economic difficulty’. Of course, since Chinese Wikipedia is edited to a large extent by Chinese-speaking people outside Mainland China, such as in Taiwan, Hong Kong and South East Asia (Liao, 2015), one would not expect any change as a result of a debate on Weibo as the participants in the debate were chiefly Mainland Chinese. A change in the domestic Chinese online encyclopaedias, however, might be expected.

The Hudong Baike article on the ‘three-year period of difficulties’ had been altered 32 times as of 25 May 2016. Hudong Baike censors entries after an article has been updated, which means that the historic pages will provide a list of all the previous versions of an article, but some of these may not be accessible. While this makes comparison of earlier versions impossible and, thus, limits the possibilities for doing research on Hudong Baike, it does reveal the extent to which the topic is controversial. In addition, it indicates the existence of article versions deemed inappropriate and reveals when censorship of an article has taken place. In the case of the article on the ‘three-year period of difficulties’, only one (posted on 16 April 2015) of the 32 article versions was accessible as of 25 May 2016. The first version was posted in March 2007, after which it was not altered until July 2011. A number of edits were then made, beginning in May 2012 shortly after the debate on Weibo. A further 12 edits were made in May and June 2012, suggesting that the debate on Weibo might have had an effect on editing activity (Hudong Baike, 2015). Unfortunately, however, because access to these versions of the article has been blocked, it is not possible to analyse their content.

In the period from April 2006, when it was created, to 25 May 2016, the Baidu Baike article was edited 59 times (Baidu Baike, 2013a). The number of edits of both the Baidu Baike and the Hudong Baike articles is extremely modest compared to that of the equivalent entry on Chinese Wikipedia, which has been edited 897 times (as of 25 May 2016) in the 11 years since its creation on 21 February 2005 (Weiji baike, 2016b). Baidu Baike also exercises censorship, but not after any edits have been made. Instead, users propose edits that are then moderated by Baidu Baike staff before being posted online (Liao, 2015). This obviously influences the content of articles, but it also makes it possible for scholars to conduct comparative research on the development of entries in the encyclopaedia over time. The version of the article on the ‘three-year natural disaster’ that existed when the debate on Weibo started had last been edited on 2 December 2011. It was then altered 11 times after the debate. Six of these edits were made between 2 May – the date on which, according to Zhao and Liu, the Weibo debate ended – and 20 May. Three more edits followed in July and August of the same year, and then two further edits were made in March and October 2013. The relatively high number of edits in May and the summer of 2012 seems to indicate that the debate on Weibo could have influenced the content of the entry. At the same time, it is important to keep in mind that due to the way in which censorship functions on Baidu Baike, these edits would probably have been screened before appearing online. Regardless of this caveat, however, any effects of the debate on China’s online collective memory of the Great Famine should be detectable in this longer lasting and more influential repository. That the current (25 May 2016) version of the article has not been edited since October 2013 might be taken as an indication that consensus has been achieved among the community of editors. At the same time, however, the facts that only one version of the Hudong Baike article is accessible, that edits which cannot be seen due to censorship were made to the Baidu Baike article and that the title of the Baidu Baike article has been locked underscore the controversial nature of the topic.

I present below the results of a comparative analysis of the Baidu Baike article version that existed before the Weibo debate (posted on 2 December 2011) and the versions that then followed. After the title of the entry, the brief introduction to the article is the first text that the reader encounters. This section was not altered during the time period studied. The introduction states that

the ‘three-year period of difficulties’ refers to the national shortage of foodstuffs and famine during the period from 1959 to 1961 in areas on the Chinese Mainland caused by the Great Leap Forward movement and industrial policies that sacrificed agricultural development. In the countryside, the farmers who experienced the period called it the days of hardships, days of shut down grains or years of failed crops. Before the 1980s, PRC government officials called it the ‘three-year natural disaster’ and then changed it to the ‘three-year period of difficulties’. Some overseas scholars, by contrast, call it the three-year great famine. Western scholars also call it the Great Leap Forward Famine. (Baidu Baike, 2013b)

The section on the causes (yuanyin) of the famine follows immediately after the introduction. This section is identical in the versions altered on 2 December 2011 and 2 May 2012. In these versions, the section consisted of 2363 characters. It was then altered significantly on 14 March 2013, ² when it was shortened to 316 characters. The shortened section on causes still states that ‘mistaken strategic policies were a far greater cause than natural disasters. It could be said that it was “30% natural disasters, 70% man-made”’ (Baidu Baike, 2013c). However, the shortened section does not clarify what these mistaken strategic policies were (Baidu Baike, 2013c). This contrasts sharply with the earlier version, which detailed the mistaken policies. The earlier version of the section also discussed in some detail the policies of communising agricultural production and prioritising steel production, both of which were said to have negatively affected the production of foodstuffs. Furthermore, inefficient food distribution and mistaken agricultural methods were said to have contributed to the decrease in agricultural production (Baidu Baike, 2011). In addition, the section stated that

under circumstances of famine occurring all over the country, from 1958 to 1959 the Chinese government still continued the exports of foodstuffs of previous years to the Soviet Union and Eastern Europe and other socialist countries in order to rapidly develop its military industry. The economy experienced difficulties and in rural areas starvation occurred. (Baidu Baike, 2011)

The government was further blamed because of its regulation of information in the media:

under circumstances where the government’s mistaken economic policies every year caused millions of people to starve to death, critical voices were not raised in the National People’s Congress, or in the media. Because of the lack of criticism of the party and a free media, the government’s mistaken economic policies could continue for three years without being adjusted, thereby causing the largest famine in modern times. (Baidu Baike, 2011)

The shortening of the section on the causes of the Great Famine clearly resulted in a considerably less detailed depiction of these causes.

When the article was edited on 14 March 2013, the section on foreign aid (duiwai yuanzhu), which appeared after the discussion of causes in earlier versions, was also removed (Baidu Baike, 2013c). The removed section, which contained 685 characters, referred to declassified documents that provided details of Chinese foreign aid during the famine to various countries, such as Albania, Cambodia, Guinea and Pakistan (Baidu Baike, 2011). Another change made when the article was edited on 14 March 2013 is that a final section listing publications that express ‘other related opinions’, among which was Yang Jisheng’s Tombstone, was removed.

The section that deals specifically with responsibility for the famine states that the Chinese government ascribes responsibility to ‘insufficient familiarity with national conditions’, ‘three years of natural disasters’ and the ‘government of the Soviet Union treacherously tearing up the two countries’ economic and technological cooperation agreement’. ³ It also mentions that prior to 1990, the Chinese government had stated that the reason for the famine was that mistaken decisions added to the difficulties of catching up caused by 3 years of disasters. The section also mentions that the People’s Daily ran articles claiming that people had plenty to eat. These parts were not altered during the period studied.

The analysis also shows how small edits can make a big difference. In the version from 2 December 2011, it was mentioned that Mao Yushi, described as a ‘Chinese economist’, or ‘scholar of economics’ (jingji xuezhe), considered China’s import and export policies to be a key cause of the famine (Baidu Baike, 2011). On 2 May 2012, the word ‘economist’ was replaced with ‘traitor’ (to the Han ethnic group, hanjian) (Baidu Baike, 2012a). This minor edit not only strongly questions Mao Yushi’s credibility as a scholar but also ascribes him the epithet of traitor, further adding to the loss of his authority. Further down the page, it was also stated that

According to Zhang Rong (in Mao: The Unknown Story, chapter 40), ⁴ during the Great Leap Forward, 30 million people died unnatural deaths not due to the unintentional, ‘mistaken policies’ of the Chinese government, but as a result of premeditated policy.

In front of Zhang Rong’s name, one small change was made on 2 May 2012: the label ‘inventor of history’ (lishi famingjia) was inserted (Baidu Baike, 2012a). On 5 May 2012, another user changed the description of Mao Yushi from traitor back to economist (Baidu Baike, 2012b). These descriptions have not been edited since. Zhang Rong is still described as a fabricator of history. Following these changes back and forth, the editors seem to have agreed on these issues as no further edits have been made to the ways in which the two figures are described.

Conclusion

This article set out to investigate how the depiction of the Great Famine in Chinese online encyclopaedias, conceptualised as collective memory places, has changed following the debate about the famine on Weibo. If these debates really had a significant impact on Chinese collective memory of the Great Famine, one would have expected to see major changes to the entries on the famine. Yet, while increased editing activity took place around the time of the debate, no major changes were made in the Baidu Baike article dealing with the Great Famine. Instead, major changes were made in 2013 that downplay the Chinese government’s responsibility for the famine by replacing the previous, detailed accounts of the causes with descriptions that merely mention mistaken government policies without describing them in any detail. That is not to say that the debate on Weibo was necessarily unimportant – it is quite possible that it was highly significant for those who participated in it and for those who accessed and read the discussion posts. Nonetheless, the results of the analysis suggest that the impact was rather limited.

These results also raise important questions concerning how Chinese online encyclopaedias function as memory places and how such sites differ from Wikipedia. As mentioned above, the editing activity on the articles on the Great Famine in the Chinese online encyclopaedias Baidu Baike and Hudong Baike (59 and 32 edits, respectively, since the entries were created) is very modest compared with Chinese Wikipedia (897 edits). This is despite the fact that Baidu Baike and Hudong Baike have come to dominate the market for online encyclopaedias in Mainland China, which appears to be at least in part due to the fact that Chinese Wikipedia has been blocked numerous times in China. This raises an important question for future research: Why is it that, despite the fact that these online encyclopaedias contain many more articles than Wikipedia, editing activity in Baidu Baike and Hudong Baike is so limited? That so much of the content in the Chinese online encyclopaedias is copy-pasted from other sources (e.g. Liao, 2015) in part explains the large number of entries. Nonetheless, as memory places where users can negotiate the content of collective memory, Baidu Baike and Hudong Baike – at least when it comes to the entry on the Great Famine – appear considerably less dynamic than Wikipedia. Put another way, it seems that these sites have not developed into participatory and pluralistic memory places that challenge official narratives to the extent that Wikipedia has. It could perhaps be that users find it less meaningful to edit entries on Baidu Baike and Hudong Baike because they know that entries, at least on controversial issues such as those related to modern history, are likely to be censored. Future research should explore these issues in greater detail, for example, by analysing whether articles on Baidu Baike and Hudong Baike are generally edited less often than those on Wikipedia, or if the entry on the Great Famine is an exception. In addition, interviews and/or surveys could be used to learn more about the attitudes and practices of the editors and users of Chinese online encyclopaedias.

Finally, the results of the analysis also demonstrate the limits of Wikipedia as a global memory place. While Wikipedia, especially its English-language version, is indeed edited by users in a large number of countries and thus might be said to involve the construction of a collective memory that traverses national borders, the fact that domestic alternatives clearly dominate in China, with its large number of Internet users, suggests that there are limits to the extent to which Wikipedia understood as a memory place can be said to be truly global. It also suggests that authoritarian states can, at least to some extent, ‘protect’ themselves from what they consider potentially subversive transnational memory narratives.