Narratives, Newspapers and The Tibetan-China Dispute - SOAS China Institute

//Narratives, Newspapers and The Tibetan-China Dispute

Narratives, Newspapers and The Tibetan-China Dispute: The Divergent Discourses Project

Lhasa, Tibet. – Photo credit: Insignifica (Flickr) / CC BY-NC 2.0

By Robert Barnett and Franz Xaver Erhard | 27 June 2024

Since the interpretive turn in the social sciences some fifty years ago, the study of conflict, and of long-running disputes in particular, has given prominence to the role of narrative and discourse in the perpetuation of antagonisms. As Peter Coleman wrote in a much-cited essay in the 1990s, conflicts of all kinds are driven and sustained by stories: once “contradictory narratives emerge for each of the disputing groups and become promoted to unquestioned fact or truth,” he wrote, those disputes “often cross a threshold into intractability”. But how do such narratives emerge? How do they relate to the original events that triggered the dispute? And how much change do these narratives undergo in their early stages?


Divergent Discourses is a joint SOAS-Leipzig University project, funded by the UK and German research bodies (the AHRC and DFG), that aims to explore these questions by studying the earliest accounts of the Sino-Tibetan conflict. That conflict began with the entry of China’s People’s Liberation Army into Tibet 74 years ago. At that time, the two parties to the dispute immediately turned to public media – primarily to newspapers – to convey their interpretations of events. By collecting and studying newspapers from the late 1950s and early 1960s, the project aims to trace the early formation of these accounts, which evolved into the deeply divergent narratives that have sustained the conflict till today.


Since very few Tibetans in the 1950s or 1960s would have been able to read Chinese, many of the newspapers produced by the new Chinese rulers in Tibet were published in Tibetan, among them the Bod ljongs nyin re’i gsar ’gyur, the Tibetan version of the official organ of the Chinese Communist Party in Tibet, Xizang ribao (“Tibet Daily”). Such publications were then, as now, primarily vehicles for praising the Party and its policies, often couched in difficult jargon and polemic. Unsurprisingly, historians of Tibet have not treated newspapers as important sources; their focus has been instead on the rare archival documents released by the Chinese government or on testimonies of survivors and refugees. Tibetan newspapers, however, served in effect as daily instructions to local cadres as well as providing contemporaneous accounts of events and political thinking at the time. They remain significant historical documents and a primary source for narrative analysis.


Tracking Parallel Discourses


In rereading these papers as historical sources, we have two advantages. One is that a significant number of newspapers were published at this time by Tibetans outside Tibet, both by those who fled in 1959 to form the exile community and its government, and by Tibetans whose ancestral homes were within India or Nepal. In all, 17 Tibetan-language newspapers were published in Tibet or India during the years 1955-62, amounting to some 10,000 pages. Reading these two types of newsprint in parallel will allow us to compare the ways in which the two communities, one newly under Communist rule, the other in the South Asia-based diaspora, shaped their separate discourses to explain and frame events and concerns of the moment.


Scholars like Melvyn Goldstein and John Powers have written about recent manifestations in the China-Tibetan context of what Todorov called the discours level of narrative construction – the use of certain forms of argument or rhetoric to promote a given view of past events. The Divergent Discourses project, by contrast, will look at the earliest versions of those discourses, formed in response to the epochal events following the arrival of Chinese Communists and the PLA in Tibet, the spread of uprisings in the late 1950s, the flight to India, the “elimination of the rebellion” and the mass introduction of “democratic reform”. This means we can study, through the newspapers of the time, what Todorov referred to as the “histoire level of narrative”—the primary selection of events out of which meaningful “stories” are initially constructed.


Histoire-level narratives are not static: even in the early weeks and months, their content, argument and emphasis shift. In the Tibet case, for example, in the 1950s the PRC described its troops as having entered Tibet to liberate Tibetans from imperialists, meaning the British and Americans (although there were reportedly only six westerners in Tibet at that time). It was only after the outbreak of uprisings in 1956 in eastern Tibetan areas, and after 1959 in central Tibet, that Chinese officials redefined “liberation” as freeing the working masses of Tibet from oppression by the upper classes and the monasteries. Early Tibetan exile accounts also changed over time, with certain events, notably episodes of resistance, being highlighted, along with a gradual shift from a largely religious sense of common purpose to one of ethnonational unity and survival. Mapping such changes in initial historical argumentation will be a major objective for the project.


Big Data and its Challenges


Our second advantage is technological: recent developments in computational methodologies will allow us to scan the newspapers we have collected, compile them into a corpus, and identify keywords, topics, sentiments, named entities, and other analytic features in the texts. Few computational tools, however, have so far been trained to work on texts in modern Tibetan. Although previous projects at SOAS and elsewhere have more or less solved the problems of word segmentation (there are no spaces or other markers distinguishing words in Tibetan) and part-of-speech recognition for classical Tibetan, the first 18 months of our project have involved developing such tools for use with the modern language. We have also had to develop a Tibetan-language model for SpaCy, the underlying software package used by the Integrated Leipzig Corpus Miner, the text-mining tool that we will use for guided topic modelling and other tasks. We have developed a model for automated character recognition of printed books in Tibetan, but a more challenging task is to create a model that will enable an automated recognition tool such as Transkribus to recognise the lay-out of a typical 1950s newspaper page, with its varying arrangement of headings, columns, titles, paragraphs and page numbers.


Aided by our team of computational linguists – Nathan Hill, James Engels, Yuki Kyogoku, Christina Sabbagh, Rachel Griffiths, Xiao Ying and Tenzin Yangzom – the project expects to have searchable versions of the newspapers available online by the autumn, the first step in the effort to understand the initial formation of narratives in the Tibetan-Chinese conflict.

Robert Barnett is a Professorial Research Associate and Senior Research Fellow in the Department of East Asian Languages and Culture at SOAS.


Franz Xaver Erhard is a Research Fellow in Zentralasienwissenschaft (Tibetologie/Mongolistik) at the University of Leipzig.


The two authors are the joint principal investigators of the Divergent Discourses project.

The views expressed on this blog are those of the author(s) and are not necessarily those of the SOAS China Institute.