Issues & Studies 37, no. 6 (November/December 2001): 77-106.

Chiang Kai-shek and the 28 February 1947 Incident:

A Reassessment

Gary D. Rawnsley and Ming-Yeh T. Rawnsley*

The objective of this research is to reconsider Chiang Kai-shek’s responsibility in causing and resolving the 28 February 1947 Incident (2-28) using a range of the new archive material now available to historians in Taiwan and to provide our own interpretation of the secondary literature on the Incident.  Moreover, we are concerned with reintroducing the relevance of agency into the early history of the Kuomintang’s (KMT’s) involvement in Taiwan, but to do so in a way that demonstrates the interaction of agents with structures.  To meet these dual objectives we identify areas of principal investigation that drive the study toward a detailed consideration of Chiang Kai-shek’s leadership style and political agenda: the interaction between Chiang and factions in Taiwan in 1947 which has been overshadowed by the simplicity of the Mainlander-Taiwanese dichotomy, and thus ignores the complexity of the crisis.

Keywords: Chiang Kai-shek; 2-28 (the 28 February 1947 Incident); Kuomintang (KMT); factions; Chen Yi

*   *   *

Dr. Gary D. Rawnsley (任格雷) is Lecturer in the School of Politics, and Director of the Institute of Asia-Pacific Studies, University of Nottingham (UK).  He is the author of Taiwan’s Informal Diplomacy and Propaganda (London: Macmillan, 2000). (Note: Dr Gary Rawnsley is now Professor of Asian International Communications, University of Leeds.)

Dr. Ming-Yeh T. Rawnsley (蔡明燁) has been Associate Research Fellow at the Institute of Asia-Pacific Studies, University of Nottingham, since 1999.  She has co-written with Gary D. Rawnsley Critical Security, Democratization and Television in Taiwan (London: Ashgate, 2001). (Note: Dr Ming-Yeh Rawnsley is now Senior Research Fellow, University of Nottingham and Visiting Research Fellow, University of Leeds.)

*The authors would like to express their gratitude to the British Academy which provided the financial support for this project.  The British Academy has also sponsored Ming-Yeh Rawnsley's presentation of this research at the American Association of Chinese Studies, October 2001.  For assistance in the research, the authors would like to acknowledge the help of Academia Historica (國史館) for granting access to the Tahsi Archives (大溪檔案); Professor C.L. Chiou (邱垂亮); and especially Professors Lai Tse-han (賴澤涵) and Huang Zhang-Jian (黃彰健) of Academia Sinica (中央研究院).

Introduction

The role and responsibility of Generalissimo Chiang Kai-shek (蔣介石) in the 28 February 1947 Incident (2-28; 二二八事件) has always been viewed as a sensitive issue.  Several factors have contributed to this: one is the hesitation in ascribing blame for such atrocities to the man still considered by many as the savior of Taiwan, if not China.   In fact, a minor school of hagiography has ignored 2-28 altogether.  The proceedings of the Conference on Chiang Kai-shek, held in Taiwan in 1987, are a serious indictment of the lack of academic objectivity about Chiang in the years following his death in 1975.[1]  In five volumes—and one devoted to Chiang Kai-shek and Taiwan—2-28 is not mentioned once.  Instead, we are treated to purple prose that has few roots in historical fact.  One contribution by Chen Che-san (陳哲三) provides a vivid description of Chiang’s death:

When he left the world, suddenly the wind began to blow in gusts, lightning flashed, and the rain began to pour down.  To most Chinese people, the man passing away with the wind and lightning was not just their chief of state but also their sagacious mentor.  On the next day, “from populous cities to remote villages in the mountains or by the sea, every part of the island had been immersed in deep mourning.”[2]

Without providing any supporting evidence, Professor Chen then states that the “majority of those who knelt down to weep for Chiang’s death are Taiwanese.”[3]  In light of the history of Taiwan since 1947, with divisions between “Mainlander” (大陸人 or 外省人) and “Taiwanese” (台灣人 or 本省人) providing a source of political confrontation left unresolved until the 1990s, difficult is to be this sanguine about the depth of popular affection for Chiang.  Nonetheless, such history reflects how, until the early 1990s, the government of the Republic of China on Taiwan, together with an alarmingly uncritical academic community there, successfully maintained a favorable picture of Chiang Kai-shek.

However, a more critical explanation for the absence of any comprehensive discussion of Chiang’s role in, and responsibility for, 2-28 is that an extraordinary amount of archive material remains closed to scholars, suggesting that Taiwan remains nervous of what researchers may uncover about Chiang.  The authors were promised that the so-called Tahsi (or Chiang Kai-shek) Archives (大溪檔案 or蔣中正總統檔案) would reveal much of what has remained hidden for the past fifty years.  We were, however, disappointed to find that the Tahsi Archives did not include anything that we did not already know or had not already been published in other investigations of the Incident.  Hence, we were both heartened and a little apprehensive to discover that there are now 200,000 more pieces of archival evidence relating to 2-28 stored at Academia Historica (國史館).  Such collections often defy logic; relevant documents are scattered far and wide among a number of libraries, archives, and even private collections, and are not stored in any chronological or even thematic order.  Sometimes, the same material appears in different places with different dates stamped on them to frustrate the most battle-hardened researcher.  A considerable number of Chinese-language studies have been published—bookshops in Taiwan usually devote a whole section to 2-28—but none have tried to access new primary documents, and therefore none provide a comprehensive, objective, and critical analysis of Chiang Kai-shek’s role.[4]

Explanations for 2-28 have avoided both complexity and controversy.  Some hold the governor of Taiwan, Chen Yi (陳儀), singularly responsible.  Such writers from different disciplines, traditions, and backgrounds as John Copper,[5] F.A. Lumley,[6] and the former American diplomat George Kerr[7] have demonized Chen Yi as an unprincipled tyrant who cared little for Taiwan.  Others have analyzed the 2-28 Incident from a structural approach emphasizing its “inevitability” given the deep- seated resentment of the Kuomintang (KMT, 國民黨) in Taiwan and the impending disaster of the civil war in China.[8]  Even Kerr, the most vociferous critic of the regime, chooses to remain silent on the possibility that Chiang Kai-shek bears some responsibility for 2-28.  The 1992 official report, however, concludes that while Chen Yi must be held directly responsible for the Incident, Chiang Kai-shek must bear his own share of blame as President of the Republic.[9]  The idea of Chiang’s ultimate responsibility concurs with the text of the 2-28 Monument (二二八紀念碑碑文) in Taipei:

Governor Chen Yi asked for the dispatch of troops from Nanjing (南京).   The chairman of the Nationalist Government, Chiang Kai-shek, without conducting a thorough investigation, responded by sending troops to Taiwan to crack down [on the protestors].

On March 8, the 21st Division of the army (第二十一軍, 即整編二十一師) landed. …  As the troops moved down to the southern part of Taiwan, they began to shoot indiscriminately. …  Within a few months, the number of deaths, injured, and missing persons amounted to tens of thousands. …  It was called the February 28 Incident.[10]

Weeks later, the bodyguard of Chen Yi revealed to a press conference in Taipei that Chen had received a telegram from Chiang ordering him to suppress any and all opposition: “Kill them all, keep it secret.”[11]  We have no way of verifying the historical accuracy of this allegation.  The bodyguard claimed he was instructed to pass the telegram to General Ke Yuan-fen (柯遠芬), the commander-in-chief of the armed forces stationed in Taiwan, i.e., the notorious Taiwan Garrison Command (TGC, 台灣警備總部).  The telegram itself has not turned up, though it may be among General Ke’s private papers.[12]

For the protestors who assembled around the 2-28 Monument in February 1996, such issues of historical detail had little relevance, for they had already decided who was to blame for the Incident.  An effigy of Chiang, draped in a banner that read “Prosecuting the 2-28 Murderer” and already battered by the sticks that many in the congregation had armed themselves with, expired in a plume of smoke.  In Taiwanese folklore at least, Chiang Kai-shek’s legacy was secure: he was little more than a butcher and a barbarian.  Difficult is to accept that these were the people who, according to Chen Che-san, wept profusely when Chiang died.  Many theories abound about why the 2-28 Incident happened and who was to blame.  Even the Chinese Communist Party (中國共產黨) offered an interpretation, as suggested by the following statement issued at a forum to commemorate the event in 1975:

Twenty-eight years ago, in February 1947, the patriotic and anti-imperialist Taiwanese people heroically rose up.  They were inspired by the great leader, Chairman Mao, who had just declared that the people should welcome the high tide of the Chinese Revolution.  Moreover, the Taiwanese people were influenced by the entire nation’s revolutionary victories at that time.  This action of the Taiwanese people shocked the reactionary ruling clique of Chiang Kai-shek and linked up with the revolutionary struggle of the Chinese people.  The 28 February uprising was part of the New Democratic Revolution under the leadership of the Chinese Communist Party.[13]

Clearly, history belongs to its interpreters.

The objective of the present research is to hack away the trees and see the wood: to reconsider Chiang Kai-shek’s responsibility in causing and resolving 2-28 using a range of the new archive material now available to historians in Taiwan and provide our own interpretation of the secondary literature on the Incident.  Moreover, we are concerned with reintroducing the relevance of agency into the early history of the KMT’s involvement in Taiwan, but to do so in a way that demonstrates the interaction of agents with structures.

To meet these dual objectives we identify areas of principal investigation that drive the study toward a detailed consideration of Chiang Kai-shek’s leadership style and political agenda: the interaction between Chiang and factions in Taiwan in 1947 which has been overshadowed by the simplicity of the Mainlander-Taiwanese dichotomy, and thus ignores the complexity of the crisis.  Chen Ming-tong[14] (陳明通) and Chen Cui-lian[15] (陳翠蓮) have done most to recover this important subject and demonstrate that the retrocession of Taiwan introduced to the island’s political culture an unprecedented level of factionalism.  The intricate nature of politics at this time suggests that 2-28 was as much about power struggles among Mainlanders as it reflected seemingly irreconcilable differences between Taiwanese and Mainlander.  A full picture of Taiwan’s political landscape would demonstrate how factions with overlapping ethnic composition lined up against each other.  So, for example, factions within the KMT allied with factions that represented Taiwan’s indigenous elite to oppose the strength of the Chen Yi-led faction (that itself included many prominent islanders; see below).[16]  This complexity is suggested in the memoirs of one Taiwanese factory-owner, Wu Zhuo-liu (吳濁流), reminiscing about Taiwan’s situation in 1947:

Among the Mainlanders, there were some who were satisfied with their positions and others who were very disappointed, and these two groups fought with each other.  Even among the returned Taiwanese, there were some who were satisfied and others still bitter.  These groups quarreled with each other.  The most serious problem was that [Chen Yi’s] government and the KMT were opposed to each other and could never agree on anything.[17]

From the outset of Taiwan’s retrocession to the Republic of China, factional politics endured and dominated Chiang’s style of leadership.  When Chiang appointed Chen Yi as governor of Taiwan, he discovered that Chen intended to appoint members of his own faction to prominent positions within the administration.[18]  Chiang passed the list of names to Chen Guo-fu (陳果夫), the leader of another faction that was close to the Generalissimo.[19]  Chen Guo-fu advised Chiang to ensure that Chen Yi appointed more KMT cadres and members of factions associated with Chiang. [20]  At the very least, Chiang should make sure Taiwan’s administration included personnel committed to the KMT, and that he should closely monitor Chen Yi and the officials he recruited.[21]  Moreover, we will suggest that it is possible to explain the contradiction in written instructions from Chiang in Nanjing to Chen Yi in Taipei—that troops should exercise self-restraint and not engage in “revenge activities”[22]—and how the troops actually behaved, leading to the notorious White Terror (白色恐怖) period, with reference to factions.

The present study also aims to consider the context of the crisis in more detail than previous studies: What else was happening in China at the time of 2-28 that might lead us toward a more rational explanation for Chiang Kai-shek’s behavior?  A cursory glance at the catalogue of Tejiao dang’an  (Special Documents 特交檔案) for the February-April 1947 period, stored in the Tahsi Archives at the Academia Historica, reveals that Chiang Kai-shek was preoccupied with a variety of problems at the time of 2-28.  These included both issues that deepened the civil war with the Chinese Communists, but also a number of political challenges, meetings with Chinese and foreign dignitaries, and related affairs that are otherwise trivial, but nevertheless impose on the time and energy of any president.

These questions refocus the research agenda on Chiang’s style of leadership, motivations, and method of making decisions.  Does Chiang’s reaction to 2-28 suggest that the party structure had collapsed and that he was no longer able to make careful decisions based on considered advice?  Alternatively, does Chiang’s style of leadership demonstrate that the bureaucratic style of politics he had engineered, with Chiang playing interests against each other, actually did more harm than good?  Chiang was careful to cultivate loyalty among his subordinates, but never allowed any to enjoy his absolute trust.  After all, many had yet to prove their unconditional loyalty to him, and many had joined with him for pragmatic reasons rather than commitment to his cause.  One such person was Bai Chong-xi (白崇禧) who had begun his career as a warlord (軍閥) in Guangxi Province (廣西省) when he challenged the Nationalist Government in armed conflict.  Like other warlords that confronted Chiang during the Northern Expedition (北伐), Bai “had to be defeated or bought off.”[23]  Chiang knew that in politics, one seldom buys friends; one merely rents them.   The 2-28 Incident clearly demonstrates his hesitation in trusting the small number of advisers that coalesced around him, and his tendency to play one faction off against another.  Is this attributable to Chiang Kai-shek’s style of leadership, or is this characteristic symptomatic of authoritarian government?

Context

The governor of the Taiwan Provincial Administration and head of the Taiwan Garrison Command (a dual position which concentrated both political and military power) was Chen Yi, a trusted general who, like Chiang, graduated from a Japanese military academy with a reputation for ruthless efficiency.  A participant in the Northern Expedition of 1924-25, Chen had acquired his reputation following the capture of Shanghai (上海) in 1927.  Kerr believes that this “placed [Chiang] in great debt to Chen.”[24]  In 1932, Chen suppressed a rebellion of the 19th Route Army (十九路軍) in Fujian (福建) and thereafter was governor of that province.  Chen responded to riots and unrest in Fujian in the same way that he would later deal with 2-28—with brute force.  Kerr goes on to claim that Chiang’s appointment of Chen was “one of the revealing and fateful decisions in Chiang’s career.  In mid-1945, Formosa was a clean slate as far as the Nationalist party was concerned. …  The key to the future of Formosa lay in the choice of personnel to fill the top ranks of the new administration.”[25]  Lumley concurs with this assessment: “So far as the history of Taiwan is concerned the major error the Generalissimo made was to appoint General Chen Yi as the first governor of Taiwan in 1945.”[26]  Thus, so far as Chiang was responsible for Chen’s appointment in the full knowledge of his record for violence in previous positions, Chiang does bear a share of the liability for 2-28.

Lumley believed that “Chen Yi seems to have known little and cared still less about the Taiwanese.”[27]  This is only partially correct.  Although Chen was a key member of the committee convened to restructure the government of the island following the Japanese surrender (the so-called Taiwan Investigation Committee), he was chosen because he was also considered an expert on Japan.  Chen might have cared little about the island, but he did know something about Taiwan from previous visits to the island.  The problem was that he disregarded Taiwanese interests and chose to respond to their requests and grievances in a decidedly provocative manner.   For example, his style of government meant that he became blinkered to Taiwan’s growing misfortunes:

Like most provincial governors, Chen Yi stayed at his desk, conferred with this staff, and did not often attend social gatherings.  He rarely toured the island or associated with the Taiwanese elite. …  Chen Yi seemed oblivious to the rapid economic deterioration, the growing social violence, and the deep social, cultural, and ethnic tensions that increasingly beset the cities.[28]

Chen spoke only in Mandarin, the national language that he believed would unite China, and never learned any of Taiwan’s local dialects, thus widening the distance between the provincial government and the people.[29]  Chen’s attitude toward Taiwan reflected the opinion of many Chinese at the close of World II: the island was a provincial backwater teeming with collaborationists, and contaminated by the half century of Japanese colonialism.[30]  Chen clearly misjudged the Taiwanese, however.   Lumley has, for instance, presented a critical assessment of Chen’s administration: “Chen Yi, who from the start conveyed an impression of pompous arrogance, became hated in a matter of weeks as no Japanese governor had ever been hated. …  Chen Yi put the clock back half a century to the last days of the Manchu rule when the mandarins could still impose their ruthless squeeze on the unfortunate islanders.” [31]  Many wondered why Chiang Kai-shek who had fostered the idea of Taiwan’s retrocession for so long, abandoned the island to a man who cared so little for such a goal.[32]

Perhaps a more useful line of inquiry to pursue is why we should expect Chiang Kai-shek to have been concerned with Taiwan and with the problems in Chen Yi’s administration.  After all, the Japanese had occupied Taiwan since 1895, although the island remained relatively untouched by both World War II and the civil war against the Chinese Communists.  In addition, the situation on the mainland was improving through 1945 and 1946, and any thoughts that the government of the Republic of China might have to retreat to Taiwan were not seriously entertained.  We glimpse a sense of Chiang’s optimism from his diary entries that describe the first meeting of the National Assembly (國代大會), December 1946:

The National Assembly, which met for the first time yesterday, is one of the most difficult undertakings in the history of the revolution, an epoch-making event …  With the convening of the National Assembly, the year-long Communist conspiracy to isolate our party and to subject the Nationalist government to attack from all sides has failed.[33]

The diary likewise records that Chiang was pleased with the constitution adopted by the National Assembly that went into effect on December 25.[34]  The problem for Taiwan was that Chen Yi insisted the constitution did not extend to the island: apparently, Taiwan was not yet ready for the kind of government promised by the National Assembly.  Instead, Chen Yi resolved to treat Taiwan as a separate entity from the rest of China, a decision that merely further antagonized the Taiwanese, many of whom were denied posts in the administration in favor not just of Mainlanders, but of the defeated Japanese.[35]  Christopher Hughes has been precise in his summation of these decisions.  Such “supercilious arrogance,” he notes, was “hardly likely to consolidate feelings of loyalty to the new regime.”[36]  Moreover, the “general style of Chen’s administration was that of victor over the vanquished rather than that of liberator,”[37] a view amplified in the official report on 2-28 that described how “many Taiwanese felt that the government treated them as colonial subjects.”[38]  There are suggestions that Chiang had explicitly instructed Chen Yi to administer Taiwan as a separate political and economic entity, [39] although as of yet there exist no means of verifying the accuracy of this claim.[40]

With the structure of China’s government in place, Chiang could turn his attention to the more urgent matter of destroying the Communist forces.  Again, there were grounds for optimism.  From July to December 1946, Nationalist victories steadily mounted,[41] and Chiang was certain that the enemy would be easily defeated or isolated by the end of the year.[42]  Dick Wilson has provided a lucid account of the civil war by the end of 1946: “Nationalist troops appeared to be on the verge of a successful campaign to recover all of Manchuria.  Their troops far outnumbered Lin Biao’s [林彪].”[43]

Were the Nationalist victories hollow?  There are suggestions that Chiang’s victories in 1946 and early 1947 were part of a grand strategy designed by Mao Zedong (毛澤東) and Zhou Enlai (周恩來) to “not only delude but also to get Chiang Kai-shek—an astute politician but a lamentable war strategist—to mobilise more of his armies, closeting them. …”[44]  Mao is reported to have said: “We shall arrange a great triumph for him. …  We should let Chiang’s brilliant General Hu Zong-nan [胡宗南] capture Yenan [Yan’an延安].  The day Hu’s crack troops enter the emptied city will be the day of his defeat.”[45]  Mao’s grand idea was to pin down Hu and his 300,000 men, thus “removing one of Chiang’s main forces from other war theatres, where the real ‘peaches of victory’ were to be plucked.”[46]  If indeed Mao and Zhou had designed such a strategy to defeat the KMT, there were signs of its success in the first two months of 1947.  With increasing setbacks for the Nationalists in Manchuria, the end of overt American aid to the KMT following the failure of the Marshall mission, and the Communist push southwards along the strategically important Qingdao (青島)-Jinan (濟南) railway, hardly surprising is that events on the mainland overshadowed events on Taiwan.  On March 1, Chiang added to his workload in Nanjing when he became the temporary premier of the ROC following the resignation of Song Zi-wen’s (宋子文, better known as T.V. Soong).[47]  By July 1947, half of the KMT’s territory in Manchuria was lost, two-thirds of the railways had been abandoned, and the KMT’s strength had been cut in half by desertion and battle:

By 1947 the government’s military leadership was in incredible confusion, and the rate of turnover in the important field commands soared to a new high. …  The ministry of defence and the supreme staff were disrupted, there was no continuity in the field commands, and everywhere there was uncertainty, confusion, and stalemate. ...

With no sure hand on the helm, the entire Nationalist Army was rendered incapable of aggressive and coordinated offensives against the communists. …

Against the government forces, the communists launched their divide-and-conquer tactics in the north, dispersing them and then devastating them through successive offensives against [KMT] troops weakened and demoralized by mismanaged allotment of supplies and material.[48]

The context is important for it lends credibility to the claim that Chiang was easily persuaded that 2-28 was part of the civil war and had been instigated by Chinese Communist agents on the island.  While difficult is to accept that Chiang was as preoccupied with Taiwan as he suggested in his communications with Chen Yi,[49] wanting the governor to “report every morning, afternoon, and evening,”[50] the Generalissimo was nevertheless concerned about the possibility of losing the province to Communist bandits.[51]

Agency is thus a fundamental explanation for the government’s response to 2-28.   This is not an original hypothesis: witness the overwhelming demonization of Chen Yi in both the Chinese and English-language literature.  Previous investigations have stopped short of evaluating Chiang Kai-shek’s role and responsibility from this approach, however, since structure avoids individual culpability while Chen Yi provides an easy scapegoat.  As yet, we have no tangible evidence to conclude that the Generalissimo ordered the violent suppression of the demonstrations.  What we have tried to do is suggest that Chiang’s independent behavior was conditioned by the context—in this case the civil war in China.  The war against the Chinese Communists motivated Chiang toward a particular interpretation of events in Taiwan which in turn conditioned his response.  Impossible is to understand and appreciate the decisions that Chiang Kai-shek made without also realizing the extent of his concern with the civil war in mainland China.

However, context is only part of the story.  Another area that must be analyzed is the relevance of factions as an explanation for the government’s response to 2-28.  This approach considers the institutional organization of Taiwan’s administration with relations between factions—relevant because Chiang’s style of leadership encouraged factional-based politics.

The Importance of Factions

In public, Chiang Kai-shek expressed little doubt as to who was responsible for 2-28: '”The trouble was instigated,” he said, “by Formosan Communists who had been defeated by the Japanese to fight in the south seas.”  From the beginning of 1947, Chiang received reports that Communists were inciting trouble on Taiwan: “According to our information there are Communists in Taiwan and [they] are gradually having an effect.  This should be seriously taken care of.  Don’t let one cell become a future source of trouble. ...  The military administration of Taiwan should be able to take care of the situation accordingly.”[52]  Chen Yi easily fed Chiang’s paranoia about Communists: “Taiwan Province has always been against Communists,” wrote the governor in explaining the Incident.  “On 27th [February], however, evil bandits got in touch with gang leaders [角頭流氓] ... to incite a riot and to attack people from the mainland.  On 28th I announced temporary martial law.”[53]  Thus, on March 6, Chen reported to Chiang that the Incident “was a planned and organized rebellious event,” and so “evil parties and rioters must be cleared by military force.  They must not be allowed to exist.”[54]  The Taiwan Garrison Command likewise described the activities of “evil parties and rioters,” intent on spreading “rebellious views … which would betray the country, separate [Taiwan] from the motherland, and [make the island] independent….”[55]  This picture of events is strengthened by Lumley, who has noted that Chen Yi “presented the ‘incident’ to Chiang Kai-shek … not as a peaceful demonstration but as a direct attack on the authority of the Generalissimo himself.”[56]

However, crucial is to understand that many of the reports that reached the Generalissimo were contradictory, a fact which reveals much about the nature of Chiang’s bureaucracy at that time.  Chiang’s style of leadership was based on cultivating friendship and, above all, loyalty through playing one layer of advisers and associates off against another.  Chiang had nurtured this talent since forming the Whampoa Military Academy (黃埔軍校), as suggested by the late Barbara Tuchman.  She described how Chiang formed a “band of adherents” at Whampoa, and that he “attracted loyalty and respect not through political inspiration as Sun Yat-sen [孫逸仙] did, but by the magnetism of an impressive personality. …  His great talent was not military but political, exercised through a mastery of balance among factions and plots. …”[57]  The documentary evidence reinforces such assessments.  As we have seen, Chen Yi described the influence of “evil bandits” and “gang leaders”—interchangeable terms of reference for Communists.  There was no reason to doubt Chen who had promised Chiang that he had everything under control.[58]  This helps to explain why Chen felt it appropriate to bypass Chiang, sensing perhaps that unacceptable was to seek the Generalissimo’s approval for every course of action that he, as governor, should have the authority to take.  When the self-styled 2-28 Management Committee (二二八事件處理委員會) called on Chen on March 7 to present a number of proposals, the Committee was immediately dismissed.  Chen told Chiang that the Committee’s proposals were “unreasonable.”[59]  The proposals were not itemized for Chiang’s consideration.[60]  Instead, Chen detailed the plans for suppressing the “devious individuals and traitors” once the military arrived in Taipei, and urged Chiang not to worry.[61]

On the other hand, the archival evidence clearly shows that Chiang received contradictory reports from others involved in the Incident.  For example, naval commander Gui Yong-qing (桂永清) visited Taiwan, and on February 15 met with local elites.  In his reports on these meetings, Gui assured the Generalissimo that the Taiwanese did not want independence, but instead expressed genuine interest in improving their relationship with Mainlanders.[62]  The Taiwanese wanted government troops to be disciplined, and asked that any disobedience be severely punished.  Gui also described how the Taiwanese wanted the Nanjing government to resolve the problems of unemployment and poverty, which led the commander toward a more rational explanation for the Incident than reports of incitement by Communists.  Gui told Chiang that the riots were due to unemployment, food shortages, and the requisition of houses by government civil servants.[63]  These views were reinforced by Huang Chao-qin (黃朝琴), speaker and head of the Taiwan Provincial Council (台灣省參議會), who explained to Chiang Kai-shek in a telegram of March 6, 1947 that the riots were due to a profound “loss of confidence” in Chen’s administration:

These disturbances have occurred because of a widening gap between the provincial government and the people.  Officials are corrupt, the administrative system has been in chaos, and some officials and police have not obeyed the law, even refusing to use Taiwanese of great talent in the administration.  There was also a different wage scale for Taiwanese and Mainlanders.  Of the Japanese property that had been built up through the sweat and tears of the Taiwanese people, more than half was confiscated by the officials and police.[64]

Huang Chao-qin undermined Chiang’s faith in Chen Yi by asking that Chiang send someone “important” to Taiwan to deal with the problems before they could escalate further.[65]  Chiang also received telegrams from the military.  For example, one from the commander of the military police, Zhang Mu-tao (張慕陶), declared the local governments on Taiwan could not control the situation, and that Chen Yi “seemed unable to appreciate the seriousness [of the situation] and still pretended everything was under control.”[66]

In other words, Chen Yi blamed agents of insurrection; Gui Yong-qing and Huang Chao-qin held structural forces responsible; Zhang Mu-tao believed Chen Yi was responsible.  With so many contradictory reports and explanations flooding into Chiang’s Nanjing office, why did the Generalissimo believe Chen above all others?  The reason is that if Chiang agreed that unemployment, food shortages, poverty, and housing were the key variables in explaining unrest, his responsibility would be undeniable.  After all, Chiang was president of the Republic of China and he alone had appointed Chen Yi as governor of Taiwan.  Accepting failure in Taiwan risked undermining the legitimacy of his government, thereby giving Communist propaganda an unfavorable advantage.  Moreover, we know that the Americans in Taiwan had, in early 1946, already suggested to the Chinese ambassador that “Taiwan is on the brink of rebellion,” while the Mintai News Agency (閩台通訊社) advised the government to send an “investigation committee to Taiwan instead of allowing the bureaucracy there turn the situation into a revolution.”[67]  The question then becomes: Why did Chiang Kai-shek decide not to pay any attention to these alarms of unrest in Taiwan?  In January 1947, Chiang had told Chen that he would send to Taiwan a military officer to assume responsibility for security affairs.  On January 11, Chen advised Chiang that this was not necessary, and reiterated that the provincial governor should deal with both civil and military issues.[68]  Believing Chen when he said he could handle matters, Chiang decided not to send any officer to take over Taiwan’s military, but urged the governor to “take strict precautions” against the mobilization of Communist forces.[69]  From such documentary evidence, possible is to construct a convincing argument that Chen did not in fact deliberately mislead the central government in Nanjing.  After all, Chiang was already worried by reports of the popularity of communism in Taiwan before the unrest of February 1947.  The outbreak of violence, the turning fortunes of the civil war in China, and the contradictory reports he received all convinced him that he was right to be so concerned.  Moreover, Chen had already assured Chiang that he was capable of handling the situation: Why would the governor risk undermining his position by asking for troops to crush the riots, and thus indicate his failure to control the situation?  Hence, Chen’s attempt to deal with the uprising in own way first and find a peaceful resolution to the crisis.  Military reinforcements from the mainland were a last resort.

Nevertheless, there is clear evidence that Chiang did not blindly accept Chen Yi’s version of events, sending instead a series of trusted individuals to Taiwan with the task of checking out Chen Yi’s reports.  One such emissary was Yu Qiao-feng (俞樵峰), most likely a KMT intelligence agent.  Via a cable to Chen Yi, Chiang told Yu that he wanted to hear Yu’s report on the situation in person.[70]  Other trusted envoys included Yang Liang-gong (楊亮功), the censor of Fujian and Taiwan, and Bai Chong-xi, minister of defense.  Like Chen Yi, Bai had also been involved in the purge of Communists from Shanghai in 1927.  Bai was sent by Chiang to investigate the Incident, arriving on March 17 and presenting his report to the Generalissimo on April 7, 1947.  Bai suggested to Chiang on March 12, 1947 that political and military power should not be concentrated in the hands of one man, and thus advised Chiang to separate control of the Taiwan Provincial Government and the Taiwan Garrison Command, advice that Chiang accepted in the post-2-28 changes in the structure of Taiwan’s provincial administration.  The most significant detail of this correspondence is not the content, but the fact that Bai sent his reports directly to the Generalissimo, thus bypassing Chen Yi.

Nevertheless, Bai remained outside the small circle of friends and advisers who were close to Chiang.  Bai was regarded by the Chinese and by foreign military experts as a military genius—an astute strategist and brilliant field commander.  Edgar Snow, no friend of the Nationalists, described Bai as “one of the most intelligent and efficient commanders boasted by any army in the world.”[71]  However, Bai was passed over as Chief of the Supreme Staff in 1946, a position that most who know him thought was his by right, because he was not part of Chiang’s close circle and had even countermanded orders from above when he felt that such orders were misguided.[72]  In public, Bai concurred with Chen Yi’s view that the Communists were responsible for the Incident, but ascribed some of the blame to the Japanese occupation: “Background causes: the Taiwanese people had received a sordid, evil education from the Japanese, and had been misled by depraved propagandists.   Proximate causes: the Communist Party and mad, ambitious schemers had used the case of an arrested smuggler to launch their uprising.”[73]  In private, however, Bai presented a more complex picture of events that gave less weight to the machinations of Communists,[74] as did Yang Liang-gong who found no evidence that the Communists were responsible for the Incident.[75]

The conclusions are clear.  As C.L. Chiou writes in his 1993 analysis of the official report on 2-28, Chiang decided to order the deliberate suppression of Taiwan’s political elite “on the basis of the misleading information brought to him.”[76]   The accumulation of evidence implies that Chiang was guilty of placing far too much trust in Chen Yi, and of having insufficient confidence in other individuals that investigated the Incident:

According to the available documents, Chiang, engaged in the civil war with the Chinese Communists, trusted Chen Yi too much.  His ready acceptance of Chen’s request for reinforcements was a misjudgment,… mistakes were committed, and his own people acted unlawfully.  In the aftermath of the Incident, when prominent Taiwanese officials … asked Chiang to reprimand those of his officers who had misconducted themselves in the crisis, Chiang refused.  Instead, he promoted [them].[77]

Archival evidence suggests, however, that from the outset Chiang did not want any violent response to the Incident.  At 6 p.m. on February 28, Chiang sent by special airplane detailed instructions to Chen on how to handle the crisis.  These included the proclamation of martial law, the necessity of compromise when possible, and noninterference by the military.[78]  Contrary to the unverified account given by Chen Yi’s former bodyguard that Chiang did order the massacre, the Tahsi Archives reveal that Chiang was sympathetic to reports reaching him that the Taiwanese were terrified of how the Taiwan Garrison Command would respond to the demonstrations: “Tell Chen Yi to stop revenge activities,” he instructed the Intelligence Bureau (Zhongtongju, 中統局) on March 12.[79]  In a personal (undated) letter to Chen, Chiang said, “Please take responsibility for revenge activities, or we will say they [Nationalist troops and the Taiwan Garrison Command] are committing a crime.”[80]

Chen Yi offered his resignation in March, claiming responsibility for the Incident.   If Chen thought Chiang would refuse, he was mistaken.  Chiang only asked Chen to stay in post until a successor could be found.[81]  Throughout March and April, Chiang received petitions from several representative committees urging him to keep Chen as governor.  Only one, dated March 28, 1947, asked Chiang to investigate and arrest Chen Yi as an immediate solution to all of Taiwan’s problems.  On April 5, Chiang responded that he was dealing with the request, but had already decided that he would not launch an investigation into either Chen’s administration of Taiwan or his management of 2-28.[82]  Chiang could not sack Chen.  After all, Chen had been appointed by the president against the advice of his bureaucracy.  To suggest now that Chiang had made a mistake would undermine his authority and legitimacy.  Far preferable was to allow Chen to offer to resign.  Chen Yi advised Chiang to appoint Chiang Ching-kuo (蔣經國) as his successor as provincial governor.  While Chiang Ching-kuo was in Taipei, Chen asked him twice to consider the appointment, “but he rejected [the suggestion] adamantly.”[83]  Jay Taylor’s biography of Chiang Ching-kuo[84] sheds no light on his activities in Taiwan at this time, and unless the dairies that Chiang Ching-kuo is alleged to have kept are discovered, we will likely never know precisely the role that Chiang Ching-kuo played in the aftermath of 2-28.[85]

The factionalism that pervaded the organizational structure of the KMT has been well documented, most perceptively by Chen Ming-tong (1995) and Chen Cui-lian (1995).  Dick Wilson details how the various “cliques” within the party and the military resented each other and jealously guarded both their interests and their access to the leadership.[86]  This factionalism reflected, however, the style of government that Chiang Kai-shek nurtured.  For example, Short has noted that “the incurable root” of the KMT’s problems “lay in the nature of the system of rule Chiang Kai-shek had created.   It was too weak and faction-ridden to impose its will by force, too corrupt and careless of public welfare to command broad-based support.”[87]

In the long term, factionalism damaged the new government’s legitimacy.  Only individuals with strong factional backgrounds, regardless of whether they were Mainlander or Taiwanese, were able to occupy important political positions.  However, still noticeable is that far fewer Taiwanese than Mainlanders occupied high-level posts in the administration.[88]  This bias added to the frustration among Taiwanese that led to 2-28, and helps to explain how the suppression of opposition was designed.  Factionalism, or patron-client networks, also combined with the central control of the economy required by the civil war to add to problems of corruption and venality that characterized Taiwan’s political landscape at this time, again fueling the resentment that manifested in 2-28.

Factionalism also came to inhabit a central position within Taiwan’s political culture and remains as important today as in 1947.[89]  Although far too naive is to claim that the KMT introduced factionalism to Taiwan in 1945, there is little doubt that their own style of politics was accented following the retrocession.  The configuration of power within Taiwan, together with the distance of the preoccupied central authority in Nanjing, is an elementary explanation for the government’s approach to 2-28.

Thus, we are presented with a confusing picture of a complex network of elites and factions based on personal loyalty and friendship, but above all mistrust and competing interests.  This is consistent with Chiang’s style of leadership that used personal allegiance as the cornerstone of his government.  This allegiance was never allowed to evolve into trust, however, since many in Chiang’s circle had been former enemies and had had to be pacified—usually by the promise of power but sometimes by military force.  While factional-based politics can serve an authoritarian leader by exploiting divisions within his government, 2-28 demonstrates the dangers of divide-and-rule tactics.  Factionalism can make an already complex situation even more confusing, as demonstrated by the number of contradictory reports that reached Chiang.  Moreover, the problem can also be exacerbated by the competing demands and interests of each network that had an investment in how the crisis was managed.  By surrounding himself with subordinates who represented different factions, Chiang gradually lost touch with political reality.  Factionalism was so extensive within the KMT because its political survival at the center depended on informal patron-client networks with regional commanders, local elites, and even secret societies.  This very informality, however, in contrast to strong and capable party mechanisms, was a source of institutional weakness.  One can trace this factionalism and the prevalence of informal patron-client networks to the failure of the Northern Expedition that merely pacified, but did not extinguish, the power bases of local warlords.[90]  Factionalism came to dominate the administration and the military both in China and in Taiwan.  We know, for example, that the central government was adamant that there should be no reprisals for the Incident, but orders to this effect were not obeyed.  Lai, Myers, and Wei present data and a convincing argument that “show much of the killing was against the orders of the Generalissimo and Chen Yi.”[91]  That orders were not obeyed was suggested by Yang Liang-gong, the censor for Fujian and Taiwan who arrived on the island in early March:

Taiwan martial law has been suspended and mopping-up operations have ceased.  These orders have been publicly announced by the Taiwan Provincial Government. …  Normalcy is supposed to have returned.   However, I have recently received many reports that military authority has used the February 28 Incident as a pretext to continue arresting people in all districts and cities. …  These actions obviously contradict the Taiwan Provincial Government’s orders and are against the central government’s principles for handling the Taiwan incident. …   Hereafter, no one should use the Incident as a pretext to arrest people or execute them willfully.  Such evil trends should cease. …[92]

If factions are ignored altogether, difficult is to agree with Lai, Myers, and Wei that neither Chen in Taiwan nor Chiang in Nanjing could have been “expected to control those division and regimental commanders and officers who rounded up and shot citizens, secretly disposed of their bodies, and strafed residences and shops.”[93]  If neither the governor of the province nor the president of the Republic could control China’s military, who could?  However, once we understand that the military was itself divided in loyalty to a number of different factional leaders and local commanders, and that neither Chiang nor Chen could exercise ultimate control over each of these, then the contradiction is more easily explained.  Although their historical value is questionable, the memoirs of Peng Meng-qi (彭孟緝), commander of the Kaohsiung Fortress Headquarters (高雄要塞司令), suggest such division.  Peng maintained that the only solution to unrest in the south of Taiwan was military force, even though Chen Yi was pursuing conciliation.[94]  Peng also reveals that he found the “rioters” who came to see him to discuss a peaceful settlement to be extremely “rude” and he arrested them.[95]  Conveniently for Peng, telephone lines between Kaohsiung and Taipei were reported to be down from the beginning of the uprising through early March, making his autonomy all the more justifiable.  Peng states that on March 6, he received the following telegram from Chen Yi: “The Incident should be resolved through political channels. ...  Withdraw your troops as soon as you receive this telegram and return order.  Xie Dong-min (謝東閔) will come to discuss solutions with you, otherwise you must take full responsibility for intensifying the Incident.”[96]  Peng comments: “The telegram made me extremely confused and worried.  After careful consideration, however, I realized that I could not back down.  This telegram became top secret between me and two other colleagues in order not to disturb morale.”[97]  It should be noted that neither Peng’s original telegram to Chen Yi requesting guidance nor Chen’s reply has been found.  Peng’s memoirs later allege that his actions enjoyed the full support of the Taiwan Garrison Command Headquarters, again suggesting division and factionalism within the island’s administration and security apparatus.[98]

That the violent response to 2-28 went against direct orders issued by both Chiang Kai-shek and Chen Yi has been contested.  Chen Yi-shen (陳儀深), for example, has advanced three key points that together make a compelling argument.  First, why did Chiang not punish those who disobeyed him?  In fact, Chiang rewarded Zhang Mu-tao, Chen Yi, and Peng Meng-qi for their military actions.  Second, according to former members of the Taiwan Garrison Command, many individuals were singled out for execution on the orders of Chen Yi.  Finally, “temporary martial law” was proclaimed in Taiwan from February 28 to March 1, 1947, and then from March 9 to May 15, 1947.  According to the law of Nationalist China, “temporary martial law” could only be formally recognized when the highest local commander (Chen Yi) asked for official consent from the central government (Chiang Kai-shek).[99]

However, our inquiries have suggested that Chiang Kai-shek pursued an obstinate style of leadership that made it difficult, if not impossible, for him to admit he might have made mistakes, was simply wrong, or was losing control.  Hence Chiang did not “reprimand those of his officers who had misconducted themselves in the crisis.”[100]  Chiang later accepted Chen Yi’s resignation and, after sufficient time had elapsed, eventually ordered his execution.  As for the immediate promotion of Zhang Mu-tao, Chen Yi, and Peng Meng-qi after the Incident, this can be explained by Chiang playing his game of factional politics since they belonged to different factions.  Second, although Chen Yi did secretly endorse the murder of identified individuals, this does not support the hypothesis that Chiang Kai-shek ordered an indiscriminate massacre.  Finally, Chen’s proclamation of “temporary martial law” with approval from Chiang Kai-shek cannot be deemed concrete evidence that Chiang ordered the massacre.  We remain confident that while we cannot absolve Chiang Kai-shek from ultimate responsibility for 2-28, difficult is to accuse him of giving direct orders to massacre Taiwanese until more substantial evidence is discovered.

Conclusions

This paper has tried to demonstrate the complexity of 2-28, arguing against any reduction of the intricacies of the Incident to a single simple explanation.  Thus, the uprising was not merely due to political incompetence, the brutality of any one individual, or the animosity between Mainlander and Taiwanese.  Rather, one must also consider the deepening civil war on the Chinese mainland in order to understand how Chiang Kai-shek responded to news reaching him from Taiwan, and why the decision to divert troops from the front line of the war against the Communists was not taken lightly.  Moreover, one can only appreciate this decision in light of the competing contemporaneous interpretations of the Incident, and thus the patron-client relationships or factions that structured Nationalist politics.

Our conclusion is that, according to archival evidence, Chiang Kai-shek must bear a share of the responsibility for 2-28 in the following ways.  First, Chiang was responsible for Chen Yi’s appointment in the full knowledge of his record for violence in previously held administrative positions.  Second, Chiang was preoccupied with the civil war in mainland China at the time, and was therefore vulnerable to misleading information, especially information that Communists were responsible for 2-28.  In other words, Chiang Kai-shek was guilty of placing far too much trust in Chen Yi, and of having too little confidence in other sources of information.  Third, the style of politics that Chiang Kai-shek had engineered allowed factionalism to dominate the administration and the military both in China and in Taiwan.  During the crisis of 2-28, factional struggles were so intense and complicated that even Chiang could not exercise ultimate control over different local commanders.  Nevertheless, we are unable to hold Chiang Kai-shek directly responsible for the massacre unless further archival evidence reveals otherwise.

One hypothesis that we feel able to develop following our access to the partial documents available in the Tahsi Archives is that Chiang preferred to issue instructions verbally rather than through written communication.  Only a few telegrams from the Generalissimo survive in the archives, and those that arrived in Nanjing from Taiwan provide little indication of his response to their content.  Several documents have been annotated with “received” or “read”—nothing more substantial is included.  For the researcher seeking the critical historical evidence required to piece together the remaining pieces of the jigsaw, this is the most frustrating part of the inquiry.  However, can one theorize that Chiang did not like the idea that difficult, unpleasant, and objectionable decisions might be traced back to him?  Some individuals have declared they received orders from Chiang Kai-shek to crush the uprising, contrary to published documents that record how he conveyed explicit instructions to the Taiwan Provincial Government to abstain from “revenge activities.”  Clearly, our judgment of 2-28 would completely change if the telegram that Chen Yi’s bodyguard referred to in his 1995 press conference (“Kill them all, keep it secret”) was ever found.  However, such evidence is still missing.  We know that many documents were lost or destroyed in the move to Taiwan.  Perhaps a more significant question, however, is why we should expect Chiang to be so concerned at such a critical time about his legacy.  There may instead be a simple explanation: the absence of documentary evidence may not necessarily suggest a grand conspiracy to deceive history; such evidence cannot be found simply because there is none to be found.

Nevertheless, much primary research does remain to be done before the full truth of 2-28 can be revealed.  The excitement over the opening of the Tahsi Archives, referred to in the English-language press as the Chiang Kai-shek Archives, was regrettably short-lived.  The authors discovered that these archives do indeed contain mountains of information about Chiang Kai-shek and 2-28, but contain little that connects the individual with the event.  Furthermore, many of the documents in this collection have already been published in Ererba shijian yanjiu baogao (二二八事件研究報告2-28 research report) by the Executive Yuan 2-28 Research Committee (行政院研究二二八事件小組), and Ererba shijian ziliao xuanji (二二八事件資料選輯 Selection of 2-28 material) by the Institute of Modern History, Academia Sinica (中央研究院近代史研究所), and have provided a rich source for historical interpretations of the Incident.  However, Professor Huang Zhang-Jian of Academia Sinica drew the authors’ attention to serious errors in the documents, especially inconsistencies in recording the dates of their transmission and reception.[101]  While one can reasonably dismiss these errors as trivial—after all, they do not guide us toward discovering who was responsible for the Incident, such mistakes can confuse the reader and thus revise his understanding and interpretation of the material.  Researchers must beware of such inaccuracies, deliberately conceived or not.  Moreover, possible is that many documents were written after the event to construct a particular version of reality.  In particular, memoirs such as those written by Peng Meng-qi in 1953 should be approached with caution.

We hope that others may seek out and assess the assortment of primary materials that have so far eluded historians.  In particular, we await full access to the 200,000 additional documents relating to 2-28 that have recently been deposited in the Academia Historica.  However, as most of these documents are local in origin, detailing events in individual areas of Taiwan, they have limited value if we continue to seek culpability.[102]  Other sources of possibly relevant material include: the KMT Party History Committee (黨史會); the Institute of Modern History, Academia Sinica; the Document Bureau (檔案局); and the recently-discovered private collection of documents maintained by one Ms. Ruan Mei-shu (阮美姝).  Still missing is Chiang Ching-kuo’s diary, without which we shall never know precisely why Chiang Ching-kuo was sent to Taiwan in 1947 and what he did there.

For fifty years, the government of the ROC worked hard to create an infallible image of Chiang Kai-shek, most spectacularly in relation to the 2-28 Incident.  How the government did this was very simple: they ignored Chiang’s role completely.  Most objectionable about this process was the tacit acquiescence of Taiwan’s academic community who, for political reasons, felt obliged to maintain this noncritical attitude (represented most disturbingly in the 1987 Conference on Chiang Kai-shek).  However, such efforts dispense with an undeniable fact: that as the president of the Republic of China, Chiang has to bear his share of accountability and responsibility for the management of that dreadful incident.  The published memoirs, interviews with participants, and official investigations have not yet been able to establish a direct link between Chiang and the massacres that happened on Taiwan in 1947, but there is sufficient evidence in the public domain to suggest that his role and responsibility requires further investigation.  The image of infallibility has been weakened by Taiwan’s democratization, but it will only be destroyed when all the evidence related to the Incident are finally presented for analysis.