TW-J-02: Three Accounts — The White Terror and Its Reckoning: the 19 May 1949 Martial Law Declaration and the World's Longest Martial Law at Its 1987 Lifting, the Punishment of Rebellion Act and the Communist-Spy Statute's Informer-Reward Machinery, the Military Tribunals That Tried Civilians for Four Decades, the 1950s Leftist Eradication and the Green Island Prison World, the Lei Chen and Peng Ming-min Cases, the 1979 Kaohsiung Incident and the Making of the Formosa Generation, the Terminal Murders of the 1980s, and the Contest Among the State-Violence/Transitional-Justice Account, the Cold-War-Necessity Account, and the Complicated-Society/Grey-Zone Account (1949–2026)

Status: DRAFTWords: 12,227

1. Key Takeaways

The White Terror (白色恐怖, Báisè Kǒngbù) was not an event but an era: a four-decade legal-institutional order under which the Republic of China party-state on Taiwan tried civilians before military tribunals for political crimes. Its architecture was erected in a single thirteen-month span: the Temporary Provisions Effective During the Period of National Mobilisation for Suppression of the Communist Rebellion (動員戡亂時期臨時條款, May 1948) suspended constitutional limits; the Taiwan Province martial law declaration of 19 May 1949 (effective 20 May) placed the island under military administration; the Punishment of Rebellion Act (懲治叛亂條例, June 1949) attached the death penalty to broadly drawn sedition offences; and the Statute for the Detection and Eradication of Communist Spies (戡亂時期檢肅匪諜條例, June 1950) built a society-wide denunciation system with material rewards for informers and collective liability for non-reporting. When martial law was lifted on 15 July 1987, it had run thirty-eight years and fifty-six days — at that time the longest continuous martial law in modern history. The repression's legal core outlived martial law itself: the Punishment of Rebellion Act fell only in May 1991, and the speech-crime provision of Criminal Code Article 100 was defanged only in May 1992, the dates this document treats as the era's juridical end.
The repression's scale is documented in ranges, not totals, and every figure must carry its provenance — a discipline this document shares with TW-J-01 on the 228 Incident's casualty contest. The most commonly cited aggregate figures — on the order of 140,000 persons tried by military courts during the martial-law period, with political-case estimates running to 200,000 in movement usage, and roughly 1,200 to over 3,000–4,000 executed, of whom the documented execution case files concentrate overwhelmingly in 1950–1954 — derive from a mix of Judicial Yuan estimates cited in 1980s legislative interpellation, the compensation foundation's certified caseload (around ten thousand cases processed between 1999 and 2014), and the archival case-file research enabled after the 1990s [TBD-VERIFY: the provenance chain of the 140,000/200,000 figures, commonly attributed to a Judicial Yuan estimate cited by former political prisoner and legislator Hsieh Tsung-min, and the execution-count ranges across the compensation foundation, Transitional Justice Commission database, and Academia Sinica scholarship]. The honest statement as of 2026: the certified floor is around ten thousand adjudicated political cases and roughly 1,100–1,500 documented executions in the foundation/TJC record, the demographic ceiling claims remain unverified, and the gap is itself part of the memory politics.
The terror had distinguishable phases, and conflating them is the most common analytical error. The 1950–1954 peak was a genuine counter-espionage eradication campaign turned mass purge: the Chinese Communist Party's Taiwan Province Work Committee — a real underground network that had grown to roughly a thousand-plus members after 228 — was destroyed root and branch after the 1950 capture and defection of its secretary Cai Xiaoqian (蔡孝乾), and with it died not only actual party members (the executed included the deputy chief of the general staff Wu Shi 吳石, June 1950) but a far wider circle of leftist readers, study-group members, rural networks (the 1952–53 Luku Incident), and the merely denounced. The 1960s–1970s pattern was different in kind: with the underground long destroyed, the apparatus turned on liberal constitutionalists (the 1960 Lei Chen/Free China case), Taiwanese-independence advocacy (the 1964 Peng Ming-min case; the 1961–62 Su Tung-chi case; the 1970 Taiyuan prison rising), writers and intellectuals (the 1968 Bo Yang and Chen Ying-zhen cases), and the overseas student diaspora via the campus-informer and blacklist systems. The 1979–1987 terminal phase combined the era's largest political trial — the Kaohsiung Incident tribunal of 1980 — with its most notorious unsolved violence: the Lin family massacre (28 February 1980), the death of Chen Wen-chen (July 1981), and the murder of Henry Liu in California (October 1984), the last of which helped force the system's end.
The 1979 Kaohsiung Incident and the 1980 military trial manufactured the political generation that would end the system and govern the democracy. The 10 December 1979 Human Rights Day rally of Formosa magazine (美麗島, Měilìdǎo) in Kaohsiung, the ensuing clashes, and the island-wide arrests produced a televised military tribunal of eight principal defendants — including Shih Ming-teh, Huang Hsin-chieh, Annette Lu, Chen Chu, and Lin Yi-hsiung — whose defiant courtroom statements, carried in the newspapers under the regime's own publicity decision, converted the defendants into national figures. The defence-lawyer cohort — Chen Shui-bian, Frank Hsieh, Su Tseng-chang among them — supplied the future Democratic Progressive Party its leadership spine: two presidents, multiple premiers, party chairs, and mayors trace their political births to that courtroom (cross-reference TW-H-PRES-02). No other single repression event in the era so directly authored its successor regime's elite.
Account 1 — the state-violence/transitional-justice account — holds that the White Terror was systematic state terror against political opposition real and imagined, that anticommunism was the warrant for one-party consolidation, and that the reckoning remains incomplete on its central question: the perpetrators. Grounded in the political-prisoner memoir canon, the Academia Sinica and museum oral-history corpus, the case-file scholarship enabled by the 2019 Political Archives Act, the compensation foundation's certified record, and the Transitional Justice Commission's 2018–2022 findings (including the rescission of more than 5,800 unjust criminal verdicts [TBD-VERIFY: final count] and the documentation of Chiang Kai-shek's personal interventions in sentencing), this account reads the compensation-without-accountability settlement of the 1990s as a deliberate design choice that paid victims while shielding judges, prosecutors, security officers, and informers from even nominal identification — and treats the naming of perpetrators, the full opening of the archives, and the disposition of KMT party assets as the unfinished obligations.
Account 2 — the Cold-War-necessity/context account — holds that the repression had a real security core, however monstrous its excesses, and must be judged inside its historical situation: a regime that had just lost a civil war, on an island Beijing was visibly preparing to invade, abandoned by Washington until June 1950, facing a genuine and documented communist underground. Its anchors are real: the Work Committee network was not imagined; the 1949–50 invasion threat was not rhetorical; the era's anticommunist security regimes in South Korea, Southeast Asia, and elsewhere were comparably or more lethal; and the same party-state built the land reform, the export economy, and ultimately the institutions through which Chiang Ching-kuo's late liberalisation and Lee Teng-hui's democratisation proceeded. This document states the account's logic without endorsement and states its weakest points with equal clarity: the repression continued for three decades after the underground's destruction; its targets were overwhelmingly non-violent; its scale and its torture record exceed any security rationale; and the "stability built the miracle" claim cannot show that the miracle required the terror.
Account 3 — the complicated-society/grey-zone account — holds that the binary victim/perpetrator frame misses the terror's social texture: this was a denunciation society in which neighbours, colleagues, classmates, and cellmates reported one another, in which coerced prisoners became informers, and in which the largest early victim cohort was mainlander leftists — a fact that complicates the ethnic reading of the era. Grounded in the newer historiography of surveillance files and informer networks — including the post-2019 wave of victims reading their own files and discovering named friends, and the 2021 public confession of DPP legislator Huang Kuo-shu that he had been a student-era informer — this account insists that society was implicated, not only the state; that both DPP foundational-memory uses and KMT context-defences flatten this texture; and that the reckoning's hardest remaining work lies in the grey zone. Its weakest points are equally stated: pushed too far, the account dissolves command responsibility into a fog of complicity and risks the false equivalence of informer and torturer, conscript clerk and sentencing general.
The reckoning ran through three distinct settlements, each more ambitious and each leaving a specific residue. The 1990s settlement (the 1991–92 legal repeals; the 1995 compensation statute for 228; the 1998 Compensation Foundation for Improper Verdicts, which from 1999 to 2014 processed roughly ten thousand martial-law-era cases and paid compensation scaled to years served and lives lost) restored victims financially and reputationally but by design named no perpetrators and annulled no verdicts. The 2017–2022 settlement (the Act on Promoting Transitional Justice; the Transitional Justice Commission's verdict rescissions, perpetrator-and-victim database, surveillance-file disclosures, and Chiang-era findings; the 2019 Political Archives Act) moved from compensation to exoneration and truth — the rescission certificates delivered to families reversed, for the first time, the state's juridical word — but stopped at the perpetrator-naming threshold and left an archives regime in which the intelligence agencies retained broad withholding powers until the 2024 amendments [TBD-VERIFY: the February 2024 Political Archives Act amendment's scope]. The post-2022 settlement (the 2022 rights-restoration act and foundation; property restitution; the National Human Rights Museum's consolidation of the Green Island and Jing-Mei parks) institutionalised memory while the political system divided over it — leaving the perpetrator question and the full archives opening as the explicitly deferred business of the 2030s.
The White Terror and 228 are distinct events whose conflation is itself a memory-politics phenomenon, and this document polices the boundary deliberately. 228 (TW-J-01) was a ten-week massacre in 1947 under the pre-retreat provincial administration, with a casualty debate in the tens of thousands and a predominantly benshengren victim profile; the White Terror was a four-decade judicialised repression under the post-1949 party-state, with documented executions concentrated in 1950–54, a substantial — in the early years arguably predominant — mainlander victim share, and a paper trail of verdicts, files, and last letters that 228's extrajudicial killings mostly lack. Taiwanese public memory, commemorative practice, and partisan rhetoric routinely merge the two into a single "authoritarian past"; the PRC appropriates the 1950s leftist martyrs (its 2013 Xishan memorial to the executed agents; its 2025 television drama on Wu Shi [TBD-VERIFY: title and reception]) in a framing most Taiwanese victims' families reject; and each account treats the relationship between the two events differently. The analytical cost of conflation is real: it ethnicises a terror that was not primarily ethnic, and it flattens a judicial machine into a massacre.
As of 2026 the three accounts coexist in asymmetric strength: the state-violence account holds the official institutions, the Cold-War account holds a shrinking but real partisan and generational base, and the grey-zone account holds the historiographical frontier. The operative contests are concrete: whether the perpetrator database becomes perpetrator naming; whether the remaining classified files — including those on the Lin family and Chen Wen-chen cases — open fully; whether the KMT-TPP legislative majority's budget actions constrain the memory institutions [TBD-VERIFY: 2025–2026 budget effects on the National Human Rights Museum and rights-restoration foundation]; and whether the era's memory, as its last survivors die, hardens into partisan ritual or matures into the more disquieting social history the grey-zone scholarship is writing. The test this document sets itself is the same as TW-J-01's: a reader from each tradition should recognise their account as fairly stated.

2. The Factual Spine: The Apparatus, the Phases, the Reckoning (1949–2026)

2.1 The legal architecture, 1948–1950

The White Terror was lawful — that is its defining institutional feature, and the feature that distinguishes it from the largely extrajudicial violence of 228. Four instruments composed the architecture. First, the Temporary Provisions Effective During the Period of National Mobilisation for Suppression of the Communist Rebellion (動員戡亂時期臨時條款), adopted by the National Assembly in Nanjing in May 1948, suspended the new ROC constitution's limits on emergency powers and presidential terms; they would remain in force until 1 May 1991, freezing the constitutional order for forty-three years (cross-reference TW-I-01). Second, on 19 May 1949, the Taiwan Provincial Government and Taiwan Garrison General Headquarters under Chen Cheng (陳誠) declared martial law across Taiwan effective 20 May — six months before the central government itself retreated to Taipei in December 1949. Under the martial-law decree and the Martial Law Act, civilians accused of a schedule of offences — sedition foremost — passed from civilian courts to military tribunals, with no meaningful appeal, closed proceedings as the norm, and confirmation of sentences running up the military chain of command to the president. Third, the Punishment of Rebellion Act (懲治叛亂條例, promulgated 21 June 1949) attached capital and life sentences to the Criminal Code's rebellion provisions; its notorious operative clause — Article 2, Paragraph 1, applying the death penalty to the conduct described in Criminal Code Article 100 (overthrowing the state by illegal means) — entered Taiwanese memory as er-tiao-yi (二條一), the clause under which most of the era's executions were pronounced. Because Article 100 as then written criminalised intent manifested in speech and association, the combination made advocacy itself capital. Fourth, the Statute for the Detection and Eradication of Communist Spies during the Period of Mobilisation (戡亂時期檢肅匪諜條例, June 1950) generalised the hunt: it obliged citizens to report suspected communist spies (feidie, 匪諜), criminalised failure to report and harbouring, established guarantor and collective-responsibility systems in workplaces and neighbourhoods, and — the provision that most shaped the social texture Account 3 dwells on — rewarded informers with a share of the convicted person's confiscated property, commonly cited as two-tenths to the informer under the implementing reward measures [TBD-VERIFY: the allocation schedule — figures of 20, 30, and 35 per cent circulate across the statute's reward regulations and their revisions, with shares variously assigned to informers and to the apprehending agencies; the structure created a documented financial incentive to denounce, exploited in fabricated cases]. Property of the convicted was confiscable wholesale; the apparatus thus funded itself from its victims.

The enforcing institution was the Taiwan Garrison Command (台灣警備總司令部, in its successive organisational forms), the martial-law executor that combined censorship, entry-exit control, surveillance, arrest, interrogation, and the military-tribunal system; around it operated the Taiwan Province Security Command's successors, the Investigation Bureau (調查局), the military's intelligence organs, and — coordinating the whole at the political level in the formative 1950s — the Political Action Committee apparatus built under Chiang Ching-kuo (蔣經國), who as head of the secret-police coordination structures and later of the General Political Warfare Department was the security state's chief architect and remains, in consequence, the reckoning's most complicated figure: the man who ran the apparatus in its bloodiest decade is the same man who, as president, permitted its dismantling in 1986–87 (cross-reference TW-B-01). Sentence review ran upward: case files declassified since the 1990s, and systematically examined by the Transitional Justice Commission, document Chiang Kai-shek's personal interventions in tribunal outcomes — annotations in his hand returning verdicts for harsher disposition, converting prison terms to death sentences, in a substantial number of reviewed political cases [TBD-VERIFY: the TJC and Academia Historica counts of Chiang-reviewed cases and the subset altered toward severity; figures on the order of thousands reviewed and hundreds altered circulate in the post-2018 scholarship]. The annotation qiangjue ke ye (槍決可也 — "execution is approved/acceptable") became, in the archival scholarship, the era's signature artefact.

2.2 The 1950s peak: the eradication of the underground and everyone near it

The repression's lethal core was concentrated in roughly five years, 1950–1954. Its proximate object was real: the CCP's Taiwan Province Work Committee (台灣省工作委員會), near-destroyed at 228's time with around seventy members (see TW-J-01 §2.3), had grown rapidly in the post-228 disillusionment — to a commonly cited strength of roughly nine hundred to over a thousand members with a wider penumbra of sympathisers by 1949–50 [TBD-VERIFY: the captured-network documentation; the security archives' own tabulations, published in post-2000 declassification volumes, are the principal source and carry the obvious caveat that a confession-driven apparatus inflates networks]. The committee's secretary Cai Xiaoqian (蔡孝乾) was captured in 1950 and defected, twice — his cooperation unravelling the network comprehensively. The takedown's most consequential case was that of Wu Shi (吳石), deputy chief of the general staff of the ROC armed forces and a CCP agent, executed on 10 June 1950 alongside the courier Zhu Feng (朱楓) and two others — the highest-ranking penetration ever documented, and the case the PRC would later canonise (Section 7.3).

But the eradication did not stop at the network's documented edge. The tribunals of 1950–54 swept in study-group members who had read leftist books, teachers and students of the post-war "red tide" reading culture, aboriginal leaders (the November 1952 cases that took the Tsou leaders Uyongu Yatauyungana (高一生) and Yapasuyongu and the Atayal leader Losin Watan (林瑞昌), executed 1954), and entire villages: the Luku Incident (鹿窟事件, December 1952 – March 1953), the era's largest single case, saw the military cordon the hill country around Shiding outside Taipei and detain on the order of four hundred villagers in pursuit of an armed-base organisation, with thirty-five executed and nearly a hundred imprisoned [TBD-VERIFY: the Luku case counts across the compensation foundation's certified record and the New Taipei municipal investigations]. Executions were carried out principally at the Machangding (馬場町) execution ground by the Xindian River in Taipei; many of the dead, unclaimed by terrified families, were buried in the paupers' section at Liuzhangli (六張犁), where the overgrown grave markers, rediscovered in 1993, became the era's most haunting physical evidence and the principal commemorative site of the pro-unification left's separate memory community (see TW-J-01 §5.3). The condemned's last letters — withheld from families for decades and released only through the post-2008 archives work, anthologised in 2015 as Wúfǎ Sòngdá de Yíshū (無法送達的遺書, "Letters That Could Not Be Delivered") — became the reckoning era's canonical primary text.

Two demographic features of the 1950s cohort matter for everything downstream. First, the victims were both mainlander and Taiwanese: mainlanders — refugees inside the regime's own migration, soldiers, teachers, the suspect cosmopolitans of the civil-war diaspora — were represented among 1950s political victims at multiples of their population share, and by several tabulations constituted the largest single victim category of the early terror [TBD-VERIFY: the proportions; analyses of the compensation foundation caseload commonly attribute roughly 40–46 per cent of 1950s-era political-case victims to waishengren, against a mainlander population share around 13–15 per cent — confirm against the foundation and TJC database tabulations]. Second, the victims were disproportionately leftist — actual party members, fellow travellers, and the merely curious — a fact that sat unusably inside both later memory regimes: awkward for a Taiwanese-nationalist frame built on ethnic victimhood, and awkward for a KMT frame in which all the executed were spies. Both features are load-bearing for Account 3.

2.3 The middle decades: liberals, independence advocates, writers, and the diaspora (1955–1978)

With the underground destroyed by mid-decade, the apparatus did not stand down; it changed targets. The Lei Chen case (雷震案, September 1960) is the hinge: Lei Chen, a mainlander KMT veteran and publisher of Free China (自由中國) — the licensed liberal journal that had moved from anticommunist loyalism to constitutional criticism — was arrested on 4 September 1960, days before the planned founding of the China Democratic Party, the first credible opposition party project, built in alliance with Taiwanese local politicians. Lei was convicted by military tribunal of harbouring a communist spy (his editor) and "spreading propaganda favourable to the rebels", and served ten years; the party died unborn. The case demonstrated the system's actual function with clinical precision: the victim was a mainlander, an anticommunist, and a former regime insider; his offence was organising electoral opposition.

The Taiwanese-nationalist strand followed. The Su Tung-chi case (1961–62) took a Yunlin county assemblyman and his network on independence-conspiracy charges. In September 1964, National Taiwan University political scientist Peng Ming-min (彭明敏) and his students Hsieh Tsung-min (謝聰敏) and Wei Ting-chao (魏廷朝) were arrested while printing the Declaration of Formosan Self-Salvation (台灣自救運動宣言) — a manifesto arguing that the "return to the mainland" was impossible, that Taiwan's future lay in a new constitution and UN membership under self-determination. Peng, internationally known, was sentenced to eight years, released to surveilled house arrest under foreign pressure, and escaped abroad in January 1970 in disguise — re-emerging as the overseas movement's most prominent figure and, in 1996, the DPP's first presidential candidate. The Taiyuan Incident (泰源事件, February 1970) — an attempted prison breakout and rising by independence-movement prisoners at Taiyuan prison in Taitung — ended with five executed; in the martyrology of the independence strand it occupies a place comparable to the 1950s leftists' Machangding.

The writers' cases marked the censorship frontier: Bo Yang (柏楊), arrested in 1968 over a translated Popeye comic strip read as lampooning the Chiangs, served nine years, much of it on Green Island, and became the era's most read memoirist; the novelist Chen Ying-zhen (陳映真), arrested the same year over a leftist reading circle, became after release the pro-unification left's leading literary voice — the two trajectories illustrating how the prison manufactured both of Taiwan's dissident traditions at once. Around the island, meanwhile, the apparatus extended overseas: the campus-informer system (the "professional students" reporting on Taiwanese students abroad, principally in the United States and Japan) and the blacklist (海外黑名單) barred thousands of surveilled students and activists from returning home — some for decades, until the early-1990s reforms — and connected the domestic terror to the diaspora politics from which much of the DPP's international wing later came. Two prison worlds anchored the era's geography: Green Island (綠島), off Taitung, housing the New Life Correction Camp (新生訓導處, 1951–1965, holding roughly two thousand political prisoners at its peak in "re-education" through labour) and later the maximum-security Oasis Villa (綠洲山莊, 1972–1987); and the Jingmei Detention Centre in Taipei, the Garrison Command's interrogation, tribunal, and holding complex.

2.4 The terminal decade: Kaohsiung, the murders, and the end (1979–1992)

The late 1970s dangwai (黨外, "outside the party") electoral movement forced the apparatus into its terminal confrontations. The Kaohsiung Incident (美麗島事件, 10 December 1979): the Formosa magazine organisation — a party in all but name — held a Human Rights Day rally in Kaohsiung that ended in clashes with military police; the regime arrested the movement's leadership island-wide. The military tribunal of March–April 1980 tried eight principals — Huang Hsin-chieh, Shih Ming-teh, Yao Chia-wen, Chang Chun-hung, Lin Yi-hsiung, Annette Lu (呂秀蓮), Chen Chu (陳菊), and Lin Hung-hsuan — for sedition; Shih received life, the others twelve to fourteen years. The regime, under American and domestic scrutiny, permitted unprecedented press coverage; the defendants' testimony — and the closing statements of a defence team that included Chen Shui-bian, Frank Hsieh (謝長廷), and Su Tseng-chang (蘇貞昌) — turned the trial into the opposition's founding broadcast (cross-reference TW-H-PRES-02; TW-A-03). While the trial proceeded came the era's most terrible unsolved case: on 28 February 1980 — the date itself a message — the mother and seven-year-old twin daughters of the imprisoned defendant Lin Yi-hsiung (林義雄) were stabbed to death in the family home, which was under round-the-clock security surveillance; a third daughter survived grave wounds. No perpetrator has ever been identified; the Transitional Justice Commission's 2020 reinvestigation report concluded the security organs' involvement could not be excluded and documented the surveillance records' gaps [TBD-VERIFY: the TJC Lin-case report's precise findings and the case's formal status as of 2026 — reopened, unresolved]. The house became the Gikong Presbyterian Church.

Two further deaths bracketed the system's end. Chen Wen-chen (陳文成), a Carnegie Mellon statistics professor home on a visit, was taken for interrogation by the Garrison Command on 2 July 1981 over his overseas support for Formosa; his body was found the next morning on the NTU campus. The official finding of suicide or accident was never believed; the TJC's 2020 report found the evidence pointed toward homicide while stopping short of attribution [TBD-VERIFY: formulation and case status]. The case cost the regime gravely in Washington, where congressional hearings followed. Then in October 1984, the writer Henry Liu (劉宜良, pen name Chiang Nan 江南), author of a critical Chiang Ching-kuo biography, was murdered in Daly City, California, by Bamboo Union gang figures recruited by the ROC Military Intelligence Bureau under Vice Admiral Wang Hsi-ling (汪希苓). The FBI traced it; Washington's fury was existential for the relationship; Wang and the triggermen were convicted (the principals released within years [TBD-VERIFY: sentences and release dates]); and the scandal — touching the succession politics around Chiang Ching-kuo's son — is conventionally credited with accelerating Chiang Ching-kuo's endgame: his 1985 declaration that no Chiang would succeed him, the tolerated founding of the DPP on 28 September 1986, and the lifting of martial law on 15 July 1987, replaced by a National Security Act that preserved reduced restrictions.

The juridical core outlasted 1987 by five years. The Punishment of Rebellion Act was repealed in May 1991 within days of the public uproar over the Independence Taiwan Association case (the May 1991 arrests of four young people, including a graduate student, for reading-circle association with the historian Su Beng); Criminal Code Article 100 was amended in May 1992, after the 100 Action Alliance campaign, to require violence or threat thereof — ending pure speech-and-association crime; the Taiwan Garrison Command was disbanded in 1992; and the blacklist effectively ended with the entry-control reforms of the same period. The White Terror's legal order was dead; its reckoning had not yet begun.

2.5 The reckoning timeline, 1987–2026

The reckoning proceeded in the three settlements summarised in the Key Takeaways; the spine dates are these. 1995: the 228 compensation statute (TW-J-01) established the template — state payment, no perpetrators. 1998: the Compensation Foundation for Improper Verdicts on Sedition and Communist-Espionage Cases during the Martial Law Period (戒嚴時期不當叛亂暨匪諜審判案件補償基金會) was created by statute; operating 1999–2014, it received over ten thousand applications and certified and compensated the substantial majority — payments scaled by years of imprisonment and capped per life taken [TBD-VERIFY: the foundation's cumulative case, certification, and payment totals — figures around 10,000+ applications and roughly NT$19–20 billion paid circulate]. The design was deliberate: "improper verdicts" (不當審判), not state crime; compensation (補償), not reparation; no verdict annulment; no perpetrator identification. 2000–2009: memorial geography — the Green Island Human Rights Monument (1999, bearing Bo Yang's inscription: "In that era, how many mothers wept through the long night for their children imprisoned on this island"); the Green Island and Jing-Mei sites preserved and opened as human-rights culture parks (2002 and 2007 respectively, consolidated 2009). 2017–2022: the Act on Promoting Transitional Justice (December 2017) and the Transitional Justice Commission (May 2018 – May 2022), which rescinded by administrative announcement the criminal verdicts of the era's political cases — restoring innocence, with rescission certificates issued to survivors and families, in cumulative batches exceeding 5,800 persons [TBD-VERIFY: the TJC's final rescission count — first batch 1,270 announced October 2018; cumulative figures of roughly 5,800–6,000 by 2022 circulate]; built the Taiwan Transitional Justice Database of cases, victims, and the officials who processed them; ran the surveillance-file ("political archives") disclosure programme that let citizens read their files; investigated the Chen Wen-chen and Lin family cases; documented Chiang Kai-shek's sentence interventions; and recommended — without securing — perpetrator-accountability legislation and the Chiang Kai-shek Memorial Hall's transformation. The Political Archives Act (政治檔案條例, 2019) compelled transfer of political case files to the National Archives, including from the KMT as a party — litigated and resisted — while leaving intelligence agencies a national-security withholding channel that kept core files (including in the Lin and Chen cases) closed until the 2024 amendments narrowed it [TBD-VERIFY: the February 2024 amendment's terms and implementation state as of 2026]. 2022 onward: with the TJC sunset, functions passed to an Executive Yuan human-rights and transitional-justice office and line ministries; the 2022 rights-restoration act created a foundation handling remaining compensation, confiscated-property restitution, and victim care [TBD-VERIFY: the act's formal title and the foundation's operations]; and the National Human Rights Museum (established May 2018, administering the Green Island and Jing-Mei parks) became the permanent custodial institution. As of mid-2026 the era is officially adjudicated as injustice, its victims compensated twice over and juridically restored — and its perpetrators unnamed, its archives incompletely open, and its meaning partisan terrain.

3. Account 1: The State-Violence/Transitional-Justice Account

3.1 Provenance and evidentiary base

The first account is the institutionalised reading of Taiwan's post-2016 transitional-justice architecture, but its provenance is older and layered. Its founding stratum is the survivor-memoir canon: Bo Yang's prison writings; the Green Island memoir literature (Tsai Kun-lin 蔡焜霖, the teenage book-club prisoner who became a publishing figure and the reckoning era's most beloved witness; Chen Meng-ho; Hu Tzu-tan; the women's-prison testimony later carried to mass audiences by the 2022 film Untold Herstory); Lei Chen's and Peng Ming-min's accounts (A Taste of Freedom, 1972); and the dangwai-era samizdat that first assembled the era's case lists. The second stratum is the oral-history and case-research corpus built after 1991 by Academia Sinica (the Institute of Taiwan History and Institute of Modern History projects under Hsu Hsueh-chi 許雪姬 and colleagues), the county cultural bureaus, and the memorial museums. The third is the archival stratum: the case files progressively released through Academia Historica and the National Archives from the late 1990s, systematised after the 2019 Political Archives Act into the TJC's database — the documentary base that converted the account from testimony into administrative record [TBD-VERIFY: the access state as of 2026 — the database is public in redacted form; scholarly access to underlying files remains mediated by privacy and the agencies' residual withholding]. The fourth is the jurisprudential stratum: the rescission programme itself, which made the account the legal truth of the Republic of China — every rescinded verdict is a state declaration that the original proceeding was an instrument of injustice.

3.2 The core claims

First: the White Terror was systematic state terror against political opposition, real and imagined — not the excesses of an otherwise defensible security campaign. The account's evidentiary lead exhibit is the system's design rather than its abuses: a legal order in which speech and association were capital, in which civilians faced military judges whose verdicts were reviewed for adequate severity up a command chain ending at the president's brush, in which informers were paid from victims' confiscated property, and in which the security organs held quota-like performance incentives, is terror as architecture; the torture, the fabricated cases, and the execution of teenagers were its outputs, not its aberrations. The phase structure seals the claim: a security rationale that might colour 1950–54 cannot reach Lei Chen (1960), Peng Ming-min (1964), Bo Yang (1968), or the Kaohsiung Eight (1980) — none of whom were communists, all of whom were prosecuted for opposition, and whose cases collectively demonstrate that the apparatus's constant function across four decades was the prevention of organised political alternatives.

Second: anticommunism was the warrant; one-party consolidation was the work. The account does not deny the underground's reality in 1950 — its scholarship documented it — but observes that the emergency powers built against it were retained for thirty-seven years after its destruction, were applied overwhelmingly to non-violent civilians, and structurally entrenched the dang-guo (黨國, party-state): the same instruments that killed the Work Committee froze the Nanjing-elected parliament in place, banned new parties, censored the press, and made the KMT's electoral dominance unchallengeable. On this reading the White Terror is not a chapter of the Cold War that happened to occur under one-party rule; it is the mechanism by which one-party rule reproduced itself, with the Cold War as its permanent justification.

Third: the compensation-without-accountability settlement was incomplete by design, and the rescissions did not complete it. The 1998 foundation's enabling statute compensated "improper verdicts" — a formula, the account notes, chosen precisely to avoid declaring the system criminal or its operators culpable; the KMT majority that passed it was indemnifying its own institutional past. The 2018–2022 rescissions cured the victims' side of the ledger — innocence restored, certificates in families' hands — but the perpetrator side remains empty: no judge, prosecutor, interrogator, or informer has ever been named in an official accountability finding, let alone sanctioned; the TJC's database records the offices through which cases flowed but the perpetrator-identification report it prepared was left unpublished or unactioned at its 2022 sunset [TBD-VERIFY: the disposition of the TJC's perpetrator-research volumes and the Executive Yuan's subsequent handling]. The account's formula: Taiwan has truth about victims and silence about authors.

Fourth: the perpetrator question — judges, prosecutors, commanders, informers, and the man whose brush approved the executions — is the unfinished obligation. The account distinguishes tiers: command responsibility (Chiang Kai-shek's documented interventions; Chiang Ching-kuo's apparatus-building; Peng Meng-chi and the security chiefs); operational responsibility (the tribunal judges and military prosecutors, some of whom lived into the democratic era in honoured retirement); and the informer stratum, where it concedes the grey zone (Section 5) while insisting that coerced informants and salaried case officers are not morally fungible. Its demands as of 2018–2026: publication of perpetrator research with names; statutory perpetrator-accountability provisions (the German Aufarbeitung and lustration vocabularies are its references); the full opening of the agency archives, including the Lin family and Chen Wen-chen files; and the completion of the entangled KMT-assets track — on the theory that the party-state's confiscations, including White Terror victims' property, sit inside the party-assets corpus that the 2016 Ill-gotten Party Assets Settlement Committee pursued through a decade of litigation [TBD-VERIFY: the assets committee's recoveries relevant to political-case confiscations].

Fifth: the White Terror, not only 228, is the foundation of Taiwan's democratic identity — and its lesson is regime-type, not ethnicity. Against the conflation that folds the era into an ethnicised 228 narrative, the account's more careful carriers — the museums, the TJC reports, the scholarly canon — insist on the mainlander victims, the leftist victims, and the era's lesson as they state it: that the danger is unaccountable state power, a lesson with explicit present-tense application in the account's post-2016 official register, where the memory of the terror is deployed as democratic immune-system against contemporary authoritarian threat across the Strait.

3.3 Internal tensions and the account's burdens

The account's burdens are real and should be stated at full strength. The entanglement burden: its institutional carrier since 2016 has been a DPP-governed state, and every instrument it built — the TJC above all, after the September 2018 Dongchang scandal (see TW-J-01 §6) — has been attackable as partisan; the account has never fully answered why perpetrator-naming, which would reach mostly KMT-affiliated dead men and some living elderly officials, is justice rather than electoral warfare, except by the (substantial) reply that the question's deferral was itself partisan in the opposite direction. The Chiang Ching-kuo problem: the account requires the architect of the 1950s security apparatus to be a principal perpetrator, while Taiwanese public opinion — including much green-leaning opinion — retains a favourable memory of his late-era governance and liberalisation; the account handles the contradiction mostly by averting its gaze, and the 2022 controversy over President Tsai's attendance at the Chiang Ching-kuo library opening exposed the unresolved seam inside the DPP itself [TBD-VERIFY: the January 2022 Ching-kuo library episode's particulars]. The selectivity burden: the account's commemorative practice has historically centred the independence-lineage martyrs and, more recently, the liberal cases, while the largest execution cohort — the 1950s leftists, whose surviving comrades commemorate at Liuzhangli under banners the DPP cannot stand beside — fits no green narrative; the museums have in fairness moved substantially to correct this. The maximalism risk: movement usage of ceiling figures (200,000; "tens of thousands executed") beyond what the certified record supports hands Account 2 its cheapest rebuttal, exactly as the "more than 30,000" usage does in the 228 contest.

4. Account 2: The Cold-War-Necessity/Context Account

This account is presented at its strongest, without endorsement, and its weakest points are stated at the section's end. It is the lineal descendant of the party-state's own rationale, but its serious contemporary form is not regime apologetics; it is a contextualist historiography with genuine scholarly carriers, a real evidentiary core, and a live constituency in KMT-aligned intellectual life, parts of the security-studies field, and the waishengren veterans' memory communities.

4.1 The account's anchors

First anchor: the underground was real, large, and aimed at the regime's destruction in coordination with an imminent invasion. The post-2000 declassifications — exploited as readily by this account as by Account 1 — document a CCP Taiwan Province Work Committee that grew from roughly seventy members in early 1947 to a network of a thousand or more by 1949–50, organised in provincial, regional, and cell structures, with armed-base ambitions (Luku's referent), penetration of the military to the level of the deputy chief of the general staff (Wu Shi), of the constabulary, the schools, and the press, and an explicit mission: preparation for the People's Liberation Army's landing [TBD-VERIFY: the captured-network documentation — the security organs' consolidated figures, republished in the declassification volumes and PRC martyrology alike, with the standing caveat that confession-driven counts inflate]. The PRC's own subsequent canonisation of the executed agents — the 2013 Beijing Xishan memorial names hundreds sent to Taiwan and killed [TBD-VERIFY: the memorial's stated count, commonly given as ~846 named plus unnamed] — is cited by this account as adverse-party confirmation that the espionage was no fabrication.

Second anchor: the existential window of 1949–50. The regime that declared martial law in May 1949 had lost the mainland within the year; Shanghai fell days after the declaration; the PLA took Hainan in April 1950 and massed for the Taiwan operation; the United States had walked away — Acheson's January 1950 defensive-perimeter speech and Truman's 5 January 1950 statement explicitly declining to defend Taiwan — and reversed only with the Korean War's outbreak in June 1950. Within that window, the account holds, Taiwan was a refugee state under genuine invasion threat with a documented fifth column; no state in that situation, in that era, ran peacetime due process. The comparative exhibit follows: South Korea's contemporaneous anticommunist violence (the Bodo League massacres of 1950 killed at minimum tens of thousands in months — a scale the entire four-decade White Terror never approached), the Malayan Emergency's detentions and deportations, Singapore's Internal Security Act detentions without trial into the 1980s, and the era's Latin American and Southeast Asian anticommunist regimes frame the KMT's record, in this account's telling, as severe but unexceptional for its cohort — and notably more judicialised, paper-trailed, and ultimately self-correcting than most.

Third anchor: the developmental-state defence. The same party-state, the account continues, executed the land reform (1949–53) that broke rural radicalism's base without Maoist bloodshed, built the export economy through the 1960s–70s, schooled the population, and maintained the social stability within which Taiwan's "miracle" compounded — and then, under Chiang Ching-kuo, initiated from strength the liberalisation that became democratisation: the dangwai tolerated, the DPP's 1986 founding unprosecuted, martial law lifted by the regime's own hand in 1987, the transition completed without civil violence under the KMT's own constitution (cross-reference TW-B-01; TW-I-01). On this reading the era should be judged as a package: the repression was the dark face of a state-building project that, uniquely among its authoritarian cohort, delivered both prosperity and an internally driven democratic landing — and the retrospective application of 2020s liberal-democratic norms to a 1950s siege state is anachronism, not justice.

Fourth anchor: the reckoning's own excesses. The account's contemporary edge is procedural: the 2017 transitional-justice statute defined the "authoritarian period" as 1945–1992 in a single criminalising sweep; the TJC's Dongchang scandal exposed partisan instrumentalisation in the commission's own leadership; the rescission programme annulled verdicts en masse without case-by-case review, thereby — in this account's telling — exonerating along with the innocent some number of actual armed-network members and spies whose convictions, whatever the process defects, had factual bases [TBD-VERIFY: the rescission programme's treatment of documented-espionage cases; the TJC excluded certain categories, and the contours are contested]; and the perpetrator-naming project would attaint elderly clerks and dead judges under retroactive standards while the same political coalition celebrates, in the PRC-spy canon, convictions obtained by the identical tribunals. The KMT-context position concedes — in its post-Ma mainstream — that grave injustices occurred, that the compensation and apologies were owed (Ma Ying-jeou as president apologised repeatedly at White Terror commemorations and presided over expansions of the memorial architecture [TBD-VERIFY: Ma-era apology occasions and the 2008–2016 memorial-park decisions]), and that the era's files should open; what it refuses is the criminal-regime characterisation, the Chiang attainder, and the treatment of the party's continuity as itself a wrong to be settled.

4.2 The account's weakest points, stated

The weaknesses are structural, and the serious carriers of the account know them. The duration problem is unanswerable on the account's own terms: the security core it invokes was destroyed by 1954–55, the invasion threat receded behind the Seventh Fleet from June 1950, and the apparatus nonetheless ran at full legal capacity for another three decades against liberals, writers, independence advocates, and electoral organisers — none of them within any reading of the security rationale. A necessity defence that covers 1950 cannot cover 1960, 1968, or 1980, and the account's characteristic move — sliding the 1950 context forward across the whole era — is its central evasion. The targeting problem: the apparatus's documented victims were overwhelmingly non-violent; its signature cases (Lei Chen, Peng, the Kaohsiung Eight) were opposition-prevention, prosecuted as such in the files; and the Taiwanese-elite and aboriginal-leadership cases sit in a recognisable continuity with 228's elite destruction (TW-J-01) that no espionage record explains. The torture and fabrication record — documented in the case files, the medical evidence, and thousands of concordant testimonies — breaks the "judicialised and proportionate" comparative claim from inside: a tribunal system fed by tortured confessions and informer bounties is not a legal process with excesses; it is an extraction machine with legal costume. The developmental claim proves less than it asserts: the account cannot show the miracle required the terror — the land reform, the export turn, and the schooling were not built by the Garrison Command — and the era's own counterfactual is supplied by the account's comparative exhibit: polities in the same Cold War with far less internal repression developed comparably. And the self-correction claim inverts the agency: martial law was lifted under siege — dangwai electoral pressure, the Kaohsiung generation's sacrifice, the Henry Liu scandal's American consequences, Washington's post-1979 leverage — and crediting the regime's endgame to its own virtue requires erasing the people it imprisoned for forcing it.

5. Account 3: The Complicated-Society/Grey-Zone Account

5.1 Provenance: the historiographical frontier

The third account is the youngest and the least partisan-aligned: it is the reading emerging since the 2010s from the surveillance-file scholarship, the informer-network research, and the oral-history projects that asked harder questions than "what was done to you?" Its carriers are academic (the Academia Sinica and university historians working the political archives; the sociology of the xianmin 線民 — "line people", the registered informants), institutional (the National Human Rights Museum's later exhibitions, which moved from martyrology toward social texture), and literary-documentary (the file-reading memoirs and essays of the post-2019 disclosure wave, in which victims and their children narrate discovering the names of friends in the margins of their own files). Its founding empirical fact: the White Terror's apparatus was thin, and its reach was social. The Garrison Command did not have agents in every tea house; it had informers — paid by statute from confiscated property, conscripted through guarantor systems and workplace cells, recruited under threat from among the arrested, embedded in student associations, congregations, and publishing houses — and a population schooled by example into self-censorship and pre-emptive denunciation. The terror, on this account, was a social system the state seeded and society ran.

5.2 The core claims

First: the victim/perpetrator binary misses the era's actual texture, which was a continuum of collaboration under duress. The files document the categories the binary cannot hold: the cellmate who reported conversations for sentence reduction; the arrested student turned and sent back to his campus association; the colleague who embellished under interrogation to end his own; the village guarantor who reported a neighbour to protect his own registered household; the prisoner-trustee administering Green Island's "re-education" of fellow prisoners. The post-2019 file disclosures made the claim concrete and personal: among the era's surveilled democrats, file after file revealed informants inside the circle — classmates, comrades, in documented cases close friends of decades — and in 2021 the sitting DPP legislator Huang Kuo-shu (黃國書) confirmed, when the files surfaced, that he had reported on fellow activists as a student informer in the 1980s, apologised, and withdrew from his party and from re-election — the single case that carried the grey zone from the seminar room into national politics [TBD-VERIFY: the Huang Kuo-shu case particulars and the parallel 2021–2022 revelations touching other politicians]. The account's point is not that the informers were guilty instead of the state; it is that several hundred thousand acts of reporting, guaranteeing, and looking away were the medium in which the state's terror propagated — and that a reckoning which names only generals will have explained nothing about how the thing actually worked.

Second: the victim-perpetrator overlap is real, and the reckoning's categories cannot process it. The compensation and rescission programmes required clean victim status; the files deliver people who were both — coerced informants who were themselves later tried; turned prisoners; the security-organ clerks of mainlander refugee background who processed cases under the same statutes that could consume them (and sometimes did: the apparatus purged inward, and security personnel themselves appear in the victim databases [TBD-VERIFY: the proportion of security-apparatus personnel among certified victims]). Account 3 holds that these cases are not marginal anomalies but the era's representative figures, and that both rival accounts need them not to exist: Account 1 because they blur the perpetrator ledger it wants completed, Account 2 because they document the coercion that voids the "society consented to necessary measures" implication.

Third: the mainlander-victim dimension breaks the ethnic reading of the era — and its suppression in public memory is itself a finding. The 1950s victim cohort's heavy mainlander share (Section 2.2 [TBD-VERIFY: proportions]) means the White Terror cannot be narrated, as 228 plausibly can, as an ethnic regime's violence against a native society: the waishengren refugee schoolteacher denounced by a fellow refugee and shot at Machangding is as representative a victim as the Taiwanese book-club teenager shipped to Green Island. That Taiwanese public memory nonetheless absorbed the White Terror into the 228-anchored ethnic frame — green commemoration centring Taiwanese martyrs, the mainlander dead claimed principally by the small pro-unification left at Liuzhangli, and the largest victim cohort thereby orphaned between the memory regimes — is, for this account, the clearest demonstration that the memory politics serve present identity coalitions rather than the record.

Fourth: both major instrumentalisations flatten the texture. Against Account 1's institutional form, this account observes that foundational-memory politics requires heroes, martyrs, and villains, and that the social history delivers mostly frightened people in impossible positions; the DPP-era memory regime, it holds, has preferred the usable past. Against Account 2, it observes that the context defence requires the society's complicity to have been consent, when the files show it was fear organised by statute. The reckoning's hardest work, on this account, is neither compensation nor perpetrator-naming but the one Taiwan has barely begun: a public language for the grey zone — for the informer who was also a victim, the family that survived by silence, the society that is the heir of both the imprisoned and the reporting neighbours — of the kind German memory culture built around the Stasi files and the Mitläufer, and which the account's carriers explicitly invoke as the comparative model [TBD-VERIFY: the state of the file-access regime for naming informers — the Taiwanese disclosures have generally redacted informer identities, a design choice the German model rejected].

5.3 The account's weakest points, stated

The false-equivalence risk is the account's standing hazard: "everyone was implicated" is one careless sentence away from "no one was responsible", and the account's vocabulary — continuum, grey zone, texture — can be and has been borrowed by Account 2's carriers to dilute command responsibility into social tragedy. The serious grey-zone scholarship guards the distinction (the coerced informant and the sentencing general are on the same continuum only in the sense that a conscript and a field marshal share an army), but the account's popularised forms do not always. The responsibility-dilution critique from Account 1 has force: the perpetrator ledger's emptiness is a documented fact about Taiwan's settlement, and an account whose effect — whatever its intent — is to make naming seem simplistic supplies the deferral with intellectual cover; victims' organisations have said so plainly. The evidentiary asymmetry: the informer files the account reads were produced by the apparatus itself, and treating the security organs' registries as transparent windows onto society's complicity repeats, at one remove, the era's own epistemology — some registered xianmin were fabrications of case officers inventing networks for bounty and quota, a possibility the account's own method documents. And the account has no politics: it prescribes a memory culture, not a programme, and in a contest where the other two accounts run states and parties, texture without institutional carriage tends to function as commentary on a settlement others write.

6. The Memory Infrastructure

6.1 The parks and the museum

The White Terror's memorial geography is carceral: where 228's monuments are civic (parks, plazas, the renamed public square), the White Terror is commemorated in its own prisons. The Green Island White Terror Memorial Park preserves the New Life Correction Camp's grounds and the Oasis Villa cellblocks, with the 1999 Human Rights Monument's curved wall carrying Bo Yang's mothers-weeping inscription and the engraved roll of victims' names; the Jing-Mei White Terror Memorial Park in Taipei preserves the Garrison Command's detention and tribunal complex, including the courtroom in which the Kaohsiung Eight were tried. Both passed in 2018 to the National Human Rights Museum (國家人權博物館) — making Taiwan, by its own framing, the first state in Asia with a national museum dedicated to its own human-rights violations. The museum's curatorial arc has tracked the historiography: early-period exhibitions centred victim martyrology; the post-2019 programme increasingly stages the file world, the informer system, and the grey zone, and the museum runs survivor-docent programmes whose actuarial end is now imminent — the same communicative-to-cultural memory transition TW-J-01 §6.7 records for 228, arriving a few years later because the White Terror's survivors were younger. Machangding is a riverside memorial park with a bare earthen mound; Liuzhangli's rediscovered grave terraces remain the pro-unification left's commemorative ground, state-designated but uneasily so, since the families and comrades who gather there reject the democratic-memory frame the designation carries.

6.2 The rescission certificates and what they mean to families

The TJC-era rescission of criminal verdicts — continued after 2022 through the rights-restoration machinery — is the reckoning's most under-analysed instrument. Formally administrative, it is experientially profound: families who had carried "rebel" household registries, employment bars, and inherited stigma across two generations received state documents declaring the convictions void ab initio — not pardoned, but erased as injustice. Survivor and family testimony collected at the certificate ceremonies recurs on two notes: that the paper came two or three decades after it could have been believed by the person it named, and that it nonetheless changed what could be said at family tables — the de-criminalisation of the dead licensing, in many families, the first full telling. The certificates' limit is the settlement's limit: the document voids the verdict without saying who procured it; several families have publicly declined or criticised certificates on exactly that ground [TBD-VERIFY: documented refusals and the representative statements]. Property consequences followed more slowly: the 2022 act's restitution track addresses confiscated estates — including high-profile cases where confiscated land sits under present-day public and party-linked holdings [TBD-VERIFY: the restoration foundation's property-case docket].

6.3 Curriculum and the generational transmission

The era entered the school curriculum late and contentiously: martial-law-era textbooks did not name it; the 1990s–2000s reforms introduced it within the Taiwan-history reframing (the Renshi Taiwan lineage); and the 2019 curriculum guidelines embed the White Terror, 228, and transitional justice in the compulsory sequence, with human-rights education a mandated cross-cutting theme. The KMT-aligned critique tracks Account 2 (context stripped, the era reduced to perpetrator-regime narrative); the activist critique tracks Account 1 (hours minimal, treatment sanitised, perpetrators absent); survey evidence suggests the median younger Taiwanese now reliably identifies the era as state injustice while conflating it with 228 — the conflation problem institutionalised [TBD-VERIFY: the survey literature on youth knowledge of 228/White Terror distinction].

6.4 The cultural production: Detention and the memory market

The reckoning's most consequential transmission vector since 2017 has been commercial culture. The horror video game Detention (返校, Red Candle Games, 2017) — set in a 1960s martial-law high school, built on the informer system and a guilt-bearing student protagonist who is the grey zone personified — became an international hit, a 2019 feature film that led the Taiwanese box office and swept Golden Horse awards, and a 2020 streaming series [TBD-VERIFY: the box-office figures and awards record]; its protagonist's arc — the informer as tragic centre rather than villain — arguably did more to mainstream Account 3 than the entire scholarly literature. The cinematic lineage runs from Hou Hsiao-hsien's A City of Sadness (1989; 228, with the White Terror's onset in its final movement) and Good Men, Good Women (1995; the 1950s leftist cases via Chiang Bi-yu's story) through Super Citizen Ko (1995; the informer's guilt, the account's cinematic prototype) to Untold Herstory (流麻溝十五號, 2022; the Green Island women's compound). The Green Island Human Rights Arts Festival [TBD-VERIFY: founding year and continuity] stages contemporary art in the cellblocks annually; the publishing wave — the last-letters anthology, the Tsai Kun-lin graphic-biography series, the file-reading memoirs — has made the White Terror, half a century after it ended, a living genre. The pattern matters analytically: as with A City of Sadness and 228, the culture has repeatedly moved ahead of the state, and the memory's present carriers are increasingly creators born after 1987 working from archives, not memory.

6.5 The archives-access state

As of 2026 the archives regime is three-tiered. The open tier: the political case files transferred to the National Archives under the 2019 act, searchable through the Transitional Justice Database — the largest disclosed political-case corpus in Asia. The mediated tier: personal-file access for victims and families (the post-2019 reading programme), with informer identities generally redacted — the design choice that distinguishes Taiwan's regime from the German Stasi-records model and keeps Account 3's deepest layer officially sealed. The withheld tier: intelligence-agency files retained under national-security and source-protection exemptions — centrally including materials bearing on the Lin family and Chen Wen-chen cases — narrowed by the 2024 amendments compelling declassification on timetables [TBD-VERIFY: the 2024 amendment's implementation and what has actually opened by mid-2026]. The KMT's own party archives remain a contested annex: the party-assets and political-archives litigation compelled transfers the party resisted as confiscatory [TBD-VERIFY: the state of KMT archive transfers]. The 2030s question is whether the third tier opens fully while anyone it names still lives.

6.6 The comparative design question: truth without trials

Taiwan's settlement is a recognisable type in the comparative transitional-justice literature: compensation-led, exoneration-completing, perpetrator-avoiding — truth about victims without trials, lustration, or naming. The South African model traded amnesty for individualised perpetrator truth under subpoena; Taiwan never put perpetrators in the room. The South Korean sequence prosecuted its living ex-presidents and ran iterated truth commissions with naming powers; Taiwan rescinded verdicts administratively and let its officials die honoured. The Chilean trajectory — closest in structure, with a negotiated transition, a long compensation-first phase, and late-arriving prosecutions as courts outflanked the amnesty (cross-reference CL-J-01) — shows the path Taiwan's design forecloses: Taiwanese law created no prosecutorial track at all, and the principals are now dead, making the Chilean late-justice swerve unavailable even in principle. The standard explanations for the Taiwanese design are structural: the perpetrating party survived as a democratic — and repeatedly governing — party, so every settlement was negotiated with the heirs of the apparatus; the transition was pacted and gradual, generating no rupture moment; and the victims' own movement prioritised recognition and the political contest over retribution. The standard critique is Account 1's: design is choice, and the choice protected the choosers. The standard defence is that the design purchased what Korea and Chile paid heavily for — a transition without backlash coups, prosecution wars, or memory-law cycles — at the price of an empty perpetrator ledger. The three accounts each read the comparison as vindication.

7. The Three Accounts in Dialogue

7.1 The partisan topology

The accounts map onto Taiwan's party system asymmetrically and incompletely. The DPP's foundational-memory use is structural: the party's origin story is the White Terror's terminal decade — its founders the Kaohsiung defendants and their lawyers, its first commemorative acts the 228 and martial-law-era taboo-breakings — and the transitional-justice programme of 2016–2022 was simultaneously a governance project and an institutionalisation of the party's own genealogy. The use is real and the account's carriers concede it; the question Account 1 cannot dissolve is whether justice administered by the heirs of the victims is thereby invalidated — its answer, as in TW-J-01, being that the alternative administrators spent seventy years declining the work. The KMT's context-defence has softened generationally: the party's post-2016 cohort mostly declines to defend the era and contests instead the reckoning's procedures, budgets, and uses (the 2024–2026 KMT-TPP legislative majority cut and conditioned transitional-justice and museum allocations within its broader budget offensive [TBD-VERIFY: the specific 2025–2026 line items], while party rhetoric centres the Dongchang precedent and the "settling scores, not history" charge). Its harder contextualist wing persists in the veterans' and waishengren memory associations and parts of the commentariat. The TPP occupies the topology's revealing third position: substantively uninvested in either memory regime, its electorate the youngest and least 228/White-Terror-motivated, its founder Ko Wen-je himself a 228 victim's grandson who built his early political identity on that fact and later campaigned on memory-politics fatigue — the party treats transitional justice as a budget line and a DPP mobilisation device, and its rise is the strongest evidence that the memory cleavage's electoral yield is declining [TBD-VERIFY: TPP platform statements on transitional justice, 2023–2026]. The structural finding: the White Terror, unlike 228, never became a single party's possession — its victims included the KMT's own people, its martyrology includes communists no DPP commemoration can canonise, and its grey zone implicates the society every party courts — which is precisely why its memory politics are more diffuse, and analytically richer, than 228's.

7.2 The 228/White-Terror relationship in memory

The two events' relationship is the Taiwanese memory field's central confusion, and each account manages it instrumentally (the full 228 treatment is TW-J-01; this section maps only the junction). The events are historically continuous — 228's suppression flowed into the 1949 emergency architecture; the post-228 disillusionment fed the underground the 1950s terror destroyed — and categorically distinct: a ten-week, largely extrajudicial massacre under the provincial administration versus a forty-year judicialised repression under the party-state; an overwhelmingly benshengren victimhood versus a mixed and initially mainlander-heavy one; a memory regime anchored on a date (28 February) versus one anchored on places (Green Island, Jing-Mei, Machangding). Public commemoration persistently merges them — "228 and the White Terror" functioning as a single compound noun in political speech, the 1980 Lin family murders on 228's anniversary fusing the two in the most literal way — and the merger does identifiable work for each account: for Account 1's nationalist register, it extends 228's ethnic-foundational reading across four more decades; for Account 2, it usefully buries the White Terror's distinct profile (where the communist-threat defence is strongest for 1950–54 and weakest for everything 228); for Account 3, the conflation is itself the primary exhibit of memory politics overwriting history, since the White Terror's mainlander and leftist victims vanish precisely when the era is read as 228 continued. This corpus separates the documents for that reason.

7.3 The PRC dimension

Beijing's relationship to the White Terror inverts its 228 posture (TW-J-01 §5). Where it claims 228 as a popular uprising, the White Terror offers it something more concrete: its own dead — the Work Committee members and military agents executed in 1950–54 are PRC martyrs by any internal definition, and the post-2013 canonisation has been systematic: the Xishan Unsung Heroes Memorial in Beijing (2013) names the executed agents with Wu Shi's cohort at its centre; anniversary commemorations and an expanding film-and-television canon (most prominently the 2025 drama on the Wu Shi network [TBD-VERIFY: title — commonly reported as The Silent Honor / 沉默的榮耀 — broadcast and reception]) present the Taiwan underground as the unfinished revolution's vanguard, with the explicit contemporary subtext that the unfinished business remains. The appropriation generates a triangular friction with no 228 parallel: Taiwanese victims' families and survivor organisations have publicly rejected the martyr framing's enlistment of their dead into a unification narrative — including descendants who accepted rescission certificates from the Taiwanese state while declining Beijing's honours [TBD-VERIFY: the documented family responses to PRC commemorations] — while the small Taiwanese pro-unification left, custodian of Liuzhangli, accepts the lineage and rejects the DPP-era democratic-memory frame instead, and Account 1's official carriers face the genuine awkwardness that the era's largest execution cohort is claimable, with documentary basis, by the adversary state. The result, visible each commemorative season, is a four-cornered contest over the 1950s dead — democratic-memory state, KMT, pro-unification left, Beijing — that the three-accounts structure can map but not simplify.

7.4 The 2030s question

Two forks define the horizon. The perpetrator-naming fork: the TJC's database and unpublished research make naming technically trivial and politically deferred; the choices are publication with names (the German path, demanded by the victims' organisations, requiring a legislative majority no recent configuration has supplied), indefinite institutional silence (the present default, ratified each budget cycle), or the actuarial resolution — the question dissolving as the last implicated officials die, leaving naming as pure historiography. The archives fork: the 2024 amendments set the trajectory toward full opening; whether the intelligence agencies' residual withholding survives the decade — and whether informer identities ever unredact — determines whether Account 3's social history can be completed and whether the Lin and Chen cases ever close evidentially. Crossing both forks is the generational transition: the survivor generation ends in the 2020s; the memory's carriers thereafter are institutions, archives, and the culture market; and the era's political salience — already visibly declining in the TPP electorate — will either be renewed by the cross-Strait present (Account 1's explicit strategy: the terror as vaccine against the authoritarianism across the Strait) or settle into the curated past. The honest forecast mirrors TW-J-01's: no convergence, but a migration of the contest from the public square into the budget, the curriculum, and the archive reading room.

8. Conclusion

The White Terror is the longer, quieter, and analytically harder half of Taiwan's authoritarian inheritance. 228 (TW-J-01) is a massacre with a date; the White Terror is a legal order with a duration — thirty-eight years of martial law, forty-three of constitutional suspension, a tribunal system that processed political opposition as crime with paperwork, signatures, and a presidential brush at the chain's top. That bureaucratic character determines everything about its reckoning: because the terror was judicialised, it could be un-judicialised — verdicts rescinded, certificates issued, files transferred — and Taiwan has done that more completely than almost any state in Asia; and because it was judicialised, its perpetrators were officials doing documented jobs, which is precisely why no Taiwanese settlement has yet borne to name them. The three accounts hold their ground because each is anchored to something real: the state-violence account to the case files, the testimony canon, and the rescission jurisprudence that made it the Republic of China's own legal truth; the Cold-War account to a genuine underground, a genuine invasion window, and a developmental record that the same party-state genuinely built; the grey-zone account to the informer registries and the file-reading wave that put friends' names in victims' margins. And each carries a burden it cannot shed: the first its partisan carriage and its Chiang Ching-kuo silence, the second its unanswerable duration problem, the third its drift toward responsibility's dissolution. What is settled, as of 2026, is the factual spine and the victims' juridical innocence; what is deferred is the perpetrator ledger and the archives' last tier; what is dying is the living memory; and what is undecided is whether the era's lesson hardens into one party's ritual, dissolves into the TPP generation's fatigue, or matures into the more implicating social history that would make the White Terror what its best scholarship already treats it as — not a story about what one party did to Taiwan, but a story about what a state under licence of emergency did through a society, and what that society, two settlements and one museum later, has still not finished saying about itself. The reader who needs one sentence: the terror ran four decades on law, the reckoning has run four decades on truth-without-perpetrators, and the unfinished half of each is the same half.

Sources

Legislative and statutory record (ROC): Martial law declaration of 19 May 1949 (Taiwan Province); Temporary Provisions Effective During the Period of National Mobilisation for Suppression of the Communist Rebellion (1948, abolished 1991); Punishment of Rebellion Act (懲治叛亂條例, 1949, repealed May 1991); Statute for the Detection and Eradication of Communist Spies (戡亂時期檢肅匪諜條例, 1950, repealed 1991) and its reward regulations; Criminal Code Article 100 and the May 1992 amendment; National Security Act (1987).
Compensation Foundation for Improper Verdicts on Sedition and Communist-Espionage Cases during the Martial Law Period (戒嚴時期不當叛亂暨匪諜審判案件補償基金會), case statistics, certified-victim tabulations, and final reports, 1999–2014 — the certified caseload floor used throughout this document [TBD-VERIFY: cumulative totals].
Transitional Justice Commission (促進轉型正義委員會), interim and final mission reports, 2018–2022: the verdict-rescission announcements (October 2018 onward), the Taiwan Transitional Justice Database, the political-archives disclosure programme, the Chen Wen-chen and Lin family reinvestigation reports (2020), and the documentation of presidential sentence interventions [TBD-VERIFY: document titles and final counts].
Act on Promoting Transitional Justice (促進轉型正義條例, 2017); Political Archives Act (政治檔案條例, 2019) and the 2024 amendments; the 2022 rights-restoration legislation and the Restoration of Victims' Rights foundation materials [TBD-VERIFY: formal titles].
Academia Historica (國史館) and National Archives Administration, declassified political case-file compilations, including the Taiwan Province Work Committee case volumes, the Lei Chen case files, and the Kaohsiung Incident records, 1990s–2020s.
Academia Sinica, Institute of Taiwan History and Institute of Modern History, White Terror oral-history series (Hsu Hsueh-chi 許雪姬 et al., eds.), 1990s–2020s — the core survivor-testimony corpus, including the Luku, Green Island, and 1950s-case volumes.
National Human Rights Museum (國家人權博物館), permanent-exhibition catalogues, survivor-docent oral histories, and research publications for the Green Island and Jing-Mei White Terror Memorial Parks, 2018–2026.
Wúfǎ Sòngdá de Yíshū (無法送達的遺書, "Letters That Could Not Be Delivered: Last Letters of White Terror Victims"), ed. Lü Cang-shang et al., 2015 (expanded editions) — the executed prisoners' last letters [TBD-VERIFY: editor attribution].
Bo Yang (柏楊), prison memoirs and the Green Island Human Rights Monument inscription; Tsai Kun-lin (蔡焜霖) oral-history and graphic-biography series (The Boy from Clearwater); Chen Ying-zhen (陳映真), autobiographical essays — the memoir canon's representative texts.
Peng Ming-min, A Taste of Freedom: Memoirs of a Formosan Independence Leader (Holt, Rinehart and Winston, 1972).
Lei Chen (雷震), memoirs and the Free China (自由中國) corpus; Hsueh Hua-yuan (薛化元) and colleagues' scholarship on the Free China movement and the 1960 China Democratic Party project.
Roy, Denny, Taiwan: A Political History (Cornell University Press, 2003), chapters on the martial-law state, the dangwai, and democratisation — the corpus's standard one-volume political history.
Greene, J. Megan, and academic literature on the KMT developmental state; Dickson, Bruce, and Taylor, Jay, The Generalissimo's Son: Chiang Ching-kuo and the Revolutions in China and Taiwan (Harvard University Press, 2000) — the standard treatment of Chiang Ching-kuo's security-state architecture and late liberalisation.
Kaohsiung Incident record: the March–April 1980 military-tribunal transcripts as published in the contemporaneous press; Formosa magazine (美麗島) corpus; defendant and defence-counsel memoirs (Shih Ming-teh, Annette Lu, Chen Chu); Jing-Mei park courtroom documentation (cross-reference TW-H-PRES-02 for the defence-cohort trajectory).
Transitional Justice Commission and Control Yuan materials on the Chen Wen-chen case (1981) and the Lin family murders (1980); the Chen Wen-chen Memorial Foundation and NTU memorial documentation [TBD-VERIFY: the 2020 report findings and case statuses].
United States congressional record: hearings on the Chen Wen-chen death and campus surveillance (1981) and on the Henry Liu murder (1985); Kaplan, David E., Fires of the Dragon: Politics, Murder, and the Kuomintang (Atheneum, 1992) — the standard account of the Henry Liu case.
The grey-zone and informer historiography: post-2019 political-archives scholarship on the xianmin (線民) system and surveillance networks; the 2021 file-disclosure controversies including the Huang Kuo-shu (黃國書) case; National Human Rights Museum exhibitions on the informer system [TBD-VERIFY: principal monographs — the literature is recent and substantially Chinese-language].
Comparative transitional-justice literature: Wu Nai-teh (吳乃德), "Transition without Justice, or Justice without History" (Taiwan Journal of Democracy, 2005); Stolojan, Vladimir, China Perspectives 2017/2; Caldwell, Ernest, on Taiwan's transitional-justice legislation; Hayner, Priscilla, Unspeakable Truths (on the commission-design typology); cross-reference CL-J-01 for the Chilean comparator.
PRC commemorative corpus: the Beijing Xishan Unsung Heroes Memorial (无名英雄广场, 2013) materials; PRC state-media treatments of the Wu Shi network including the 2025 television drama [TBD-VERIFY: title and reception]; Taiwanese press documentation of victims'-family responses to PRC martyr framings.
Cultural corpus: A City of Sadness (Hou Hsiao-hsien, 1989); Good Men, Good Women (Hou, 1995); Super Citizen Ko (Wan Jen, 1995); Detention (Red Candle Games, 2017; film, John Hsu, 2019; series, 2020); Untold Herstory (流麻溝十五號, 2022); Green Island Human Rights Arts Festival programmes; Lin, Sylvia Li-chun, Representing Atrocity in Taiwan: The 2/28 Incident and White Terror in Fiction and Film (Columbia University Press, 2007).
Curriculum record: Ministry of Education, the 2019 (108) curriculum guidelines' history and human-rights education provisions, and the associated public controversy record.
Liberty Times, United Daily News, China Times, Taipei Times, and Focus Taiwan (CNA) coverage of the TJC's operations and the Dongchang scandal (September 2018), the rescission-certificate ceremonies, the 2021 informer-file controversies, the 2024 Political Archives Act amendments, the KMT party-archives litigation, and the 2025–2026 budget actions affecting the memory institutions.

TW-J-01: Three Accounts — The 28 February 1947 Incident and Its Memory Politics — the upstream companion document: 228 as the terror's prologue and the conflation problem treated at §7.2 here; the two documents share the three-accounts discipline and the casualty-attribution rules.
TW-A-02: Cross-Strait Relations 1949–2024 — the macro frame for the 1949–50 invasion window (Account 2's anchor) and the PRC's martyr-commemoration dimension (§7.3).
TW-A-03: Chen Shui-bian Presidency 2000–2008 — the first DPP-era handling of the authoritarian legacy, between the 1990s compensation settlement and the 2017 statutory phase.
TW-B-01: 1987 Lifting of Martial Law and Democratisation — the terminal arc (1986–1992) treated here at §2.4 in repression perspective and there in transition perspective.
TW-C-01: Ma Ying-jeou Presidency 2008–2016 — the KMT reform-era commemorative position's governing phase and the pre-TJC memorial expansions.
TW-H-PRES-02: Chen Shui-bian Biography — the Kaohsiung defence-lawyer cohort's representative trajectory from the 1980 tribunal to the presidency.
TW-I-01: The ROC Constitutional Architecture — the Temporary Provisions, the frozen parliament, and the 1991–92 constitutional normalisation that ended the era's legal order.
TW-K-01: Tsai 2019 Five Musts Rejection of Xi Address — the contemporary cross-Strait frame within which Account 1's terror-as-democratic-vaccine register and Beijing's martyr appropriation compete.
TW-N-01: Taiwan in International Perceptions — the external-lens treatment of Taiwan's authoritarian-to-model arc on which the reckoning's international standing rests.
TW-O-01: Taiwan Megatrends — The 2030s Questions — the forward horizon for the perpetrator-naming and archives forks (§7.4).
CL-J-01: Three Accounts — The Pinochet Legacy and Chilean Memory Politics — the principal comparative document for the truth-without-trials design question (§6.6): the Chilean late-prosecution path Taiwan's settlement forecloses.
TW-M-01: Taiwanese Identity and the Status-Quo Doctrine — The Island's Organising Ideas
TW-H-PRES-03: back-reference added by symmetry sweep