RW-B-02: The 1990 RPF Invasion and the Rwandan Civil War — Mutara to Arusha to the Hundred Days (1 October 1990 – 4 July 1994)

Status: DRAFTWords: 15,903

Section Outline

Key Takeaways — 11 bullets capturing the war's institutional and political shape.
The Long Diaspora: The 1959–1990 Tutsi Refugee Exodus and the Politics of Statelessness — the 50,000–700,000 figure problem; Uganda, Burundi, Tanzania, Zaire camps; the Obote-era persecution; the citizenship-foreclosure inside Rwanda; the conditions of return-as-armed-politics.
The Ugandan Crucible: RANU, the NRA Bush War, and the Founding of the RPF (1979–1990) — Museveni's 1981–1986 campaign; Rwigyema and Kagame inside the NRA; the December 1987 founding of the RPF; the 1989–1990 invasion preparation.
The 1 October 1990 Invasion: Mutara, the Kagitumba Crossing, and the Three-Day Disaster — the cross-border attack; the initial RPF order of battle (~2,500–4,000 fighters); the Habyarimana regime's response and the 4–5 October "phantom attack" in Kigali; the 8,000–10,000 ibyitso arrests.
Rwigyema's Death (2 October 1990), Kagame's Return from Fort Leavenworth, and the Bayingana–Bunyenyezi Question — the contested circumstances of Rwigyema's death; the late-October return of Kagame; the executions of Bayingana and Bunyenyezi; the Virunga retreat and the doctrinal transformation from conventional to insurgency.
Mulindi and the Reconstruction of the RPA (Winter 1990–1991 to mid-1991) — the base in northern Rwanda; the rebuilding of the force; the parallel French Noroît / Belgian / Zairian foreign interventions; the 29 March 1991 N'Sele Ceasefire.
The Northern Offensives, the State-Orchestrated Reprisal Massacres (1990–1992), and the Ibyitso Architecture — the Kibilira massacres (October 1990); the Bagogwe massacres (January–March 1991); the Bugesera massacres (March 1992); the propaganda architecture of ibyitso, Hutu Ten Commandments (December 1990), and Kangura; the foreign documentation by the March 1993 International Commission.
The 8 February 1993 Offensive: From the Ceasefire Lines to 30 km from Kigali — the trigger (January 1993 Gisenyi/Ruhengeri massacres); the operational geography; the military-balance shift; the 7 March 1993 Dar es Salaam ceasefire as a renegotiated post-offensive baseline; the displacement of 800,000–1,000,000 civilians.
From Arusha to the Brink: Military-Integration Negotiations (50:50 Command / 60:40 Force), the 3 August 1993 Protocol, the 4 August Signature — the negotiation of the Armed Forces Integration Protocol; Bagosora's reported "apocalypse" reaction; the implementation timetable and its non-execution; the RPF battalion at the Conseil National de Développement (CND) building in Kigali under UNAMIR escort (December 1993).
The Long Wait: August 1993 – 5 April 1994, UNAMIR, and the Genocide Fax — the 5 October 1993 Security Council Resolution 872 and Dallaire's deployment; the 11 January 1994 cable to DPKO; the Belgian assessments; the 21 October 1993 Burundian shock (the Ndadaye assassination); Habyarimana's procedural obstruction; the 5 January 1994 partial swearing-in.
6 April 1994: The Shoot-Down, the Resumption of Hostilities, and the RPF Decision to Resume the War — the Falcon 50 shoot-down; the RPF battalion's break-out from the CND; the Mulindi-to-Kigali advance; the parallel Tutsi extermination underway in regime-controlled territory; the encounter with UNAMIR (the 7 April killing of the ten Belgian peacekeepers); the regional and international non-response.
The 100-Day Military Campaign: Byumba, Gitarama, Kigali, Butare, Ruhengeri/Gisenyi (April–July 1994) — the operational geography of the RPF advance; the fall of Byumba (early April), Gitarama (13 June), Kigali (4 July), Butare (3 July), Gisenyi (17 July); the regime collapse to Zaire; the parallel French Opération Turquoise (23 June – 21 August 1994); civilian-protection record and contested RPF-conduct allegations.
Three Accounts of the War (RPF Official / Hutu Power / International Observer) — the RPF reading as a justified return-from-exile war culminating in the only stoppage of an active genocide; the Hutu-Power-residual reading as an unprovoked Ugandan-backed invasion that triggered the regime's defensive radicalisation; the Dallaire / Adelman–Suhrke / Carlsson Report reading as a tractable conflict whose mediation was structurally under-resourced and whose implementation phase was diplomatically abandoned.
Forward View: From the 4 July 1994 Capture of Kigali to the Mapping Report and the M23 Question — the institutional inheritance of the war (the RPA-to-RDF transformation; the GoNU; the Arusha framework's partial survival in the post-genocide transitional institutions; the 1996–1997 First Congo War and the 2010 Mapping Report as the long-arc consequence of an army that did not demobilise).

1. Key Takeaways

The 1990–1994 Rwandan civil war was the armed return of a stateless Tutsi diaspora that the post-1959 Hutu-majoritarian settlement had foreclosed from peaceful repatriation for thirty-one years, and the corpus reads it as the institutional carrier of every subsequent feature of modern Rwandan governance — from the 1993 Arusha Accords through the 1994 genocide to the post-1994 RPF state and the 1996–2003 Congo wars. The Rwandan Patriotic Front (RPF), founded in Kampala in December 1987 within the second-generation Ugandan-Tutsi refugee community and built on the older RANU (Rwandese Alliance for National Unity, 1979) framework, launched a conventional military invasion at Kagitumba in the Mutara region of northeastern Rwanda on 1 October 1990 with approximately 2,500–4,000 fighters drawn predominantly from Yoweri Museveni's National Resistance Army (NRA). The invasion is the proximate cause of the 1990–1993 anti-Tutsi reprisal massacres, the 1991 multi-party constitutional reform, the 1992–1993 Arusha negotiations, the 1993 UNAMIR deployment, and the political-existential framing within which the akazu faction prepared the 1994 genocide.
Fred Rwigyema, the charismatic Deputy Commander of the Ugandan NRA and the Deputy Minister of Defence under Museveni, commanded the invasion for less than thirty-six hours before his death on 2 October 1990 — and the contested circumstances of his death are the single most-disputed individual fact of the early war. Rwigyema, the public face of the RPF and by widely-cited accounts (Otunnu 1999; Prunier 1995; Kinzer 2008) the leader the Rwandan Tutsi diaspora most overwhelmingly trusted, was killed near Kagitumba on the second day of the operation. [TBD-VERIFY: the official RPF account attributes Rwigyema's death to a stray bullet during the assault on the Forces Armées Rwandaises (FAR) positions; rival accounts, advanced principally by Ugandan oppositional sources and some former RPF officers cited in Wrong (2021), allege he was killed during an internal command dispute with the senior Majors Peter Bayingana and Chris Bunyenyezi; the subsequent execution of Bayingana and Bunyenyezi at Kagame's order in late 1990 is presented in the second account as confirmation of the internal-dispute thesis and in the first as the disciplining of factionalists whose dissent had contributed to the operational collapse. The corpus records the contestation and does not adjudicate.] With Rwigyema dead, two more senior commanders dead within weeks, and the conventional advance halted, the RPF lost roughly half its initial strength in the first ten days of operations.
Paul Kagame, then Major and head of military intelligence in Museveni's NRA, was undergoing senior officer training at the U.S. Army Command and General Staff College at Fort Leavenworth, Kansas, when the invasion was launched — and his return in late October 1990 to assume command of a near-disintegrating force is the foundational moment of his subsequent thirty-six-year political career. [TBD-VERIFY: precise date of his return, commonly given as the last week of October 1990 with departure from Fort Leavenworth around 15 October and arrival in the Uganda–Rwanda border zone within ten days.] Kagame's first operational decision was the execution of Majors Peter Bayingana and Chris Bunyenyezi [TBD-VERIFY: precise dates of the executions, commonly placed in late October or early November 1990] for the factional dissent alleged to have contributed to Rwigyema's death; his second was the withdrawal of the residual force into the Virunga volcanic mountains along the northern border, where, through the harsh winter of 1990–1991, he reconstituted the RPA from a broken conventional brigade into a disciplined guerrilla force. The Virunga reconstitution — narrated in Rusagara (2009) as the doctrinal founding of the modern RDF and in Kinzer (2008) as Kagame's leadership formation — is the institutional moment from which the army that would halt the 1994 genocide and subsequently invade Zaire in 1996 emerged.
The initial invasion's domestic political effect inside Rwanda was the deliberate ethnic-violence response of a regime that read the RPF military threat as proof of the ibyitso (accomplice) status of every Tutsi citizen — and the resulting 1990–1992 reprisal massacres are the institutional precedent the 1994 genocide would scale. On the night of 4–5 October 1990 the Habyarimana government staged a fake attack on Kigali — the so-called "phantom attack" in which gunfire was reported across the capital but no RPF forces were present, since the RPF was at that moment being driven back from Mutara in the northeast — to justify the mass arrest of approximately 8,000–10,000 suspected RPF accomplices, the overwhelming majority of them Tutsi civilians, held without charge in stadium-and-prison detention centres for weeks; [TBD-VERIFY: dozens to several hundred died of overcrowding, beatings, and medical neglect, with the precise toll variously reported across Africa Watch (1993), African Rights (1995), and Des Forges (1999)]. Within ten days the Kibilira massacres of 11–13 October 1990 in Gisenyi prefecture killed approximately 350–400 Tutsi civilians under the organised direction of communal officials; the January–March 1991 Bagogwe massacres in the northwest killed approximately 300–1,000 Tutsi of the Bagogwe pastoral subgroup; the 4–9 March 1992 Bugesera massacres in southern Rwanda killed approximately 277–300 Tutsi after an RTLM-precursor radio broadcast on Radio Rwanda announced a fabricated Tutsi plot to kill Hutu leaders.
The civil war is conventionally divided into four phases — October 1990 – March 1991 (conventional invasion and FAR restoration); April 1991 – January 1992 (guerrilla insurgency); mid-1992 – February 1993 (Arusha I ceasefire and protocol negotiation); February – August 1993 (the 8 February offensive, the Dar es Salaam ceasefire, and the final protocols) — and the operational arc culminated in an RPF military superiority that the 8 February 1993 offensive made unmistakable. The offensive, triggered by the January–early February 1993 anti-Tutsi massacres in Gisenyi and Ruhengeri prefectures documented contemporaneously by the International Commission of Investigation on Human Rights Violations in Rwanda (the joint Africa Watch / FIDH / ICHRDD / UIDH mission whose March 1993 report was the first internationally-credible documentation of regime-organised ethnic killing), broke the formal ceasefire and brought the RPF to approximately 30 kilometres from Kigali before halting under combined French Operation Noroît reinforcement and OAU diplomatic pressure. The offensive displaced approximately 800,000 to 1,000,000 civilians from the northern prefectures southward — the internally displaced persons (IDPs) whose massed presence in the Kigali periphery and the central south would be a structural feature of the pre-genocide political landscape through April 1994.
The military-integration provisions of the 3 August 1993 Armed Forces Protocol — 60:40 troop-strength ratio favouring the FAR but 50:50 command-position rotation — are the proximate political cause of the akazu-aligned faction's decision that Arusha implementation was unsurvivable, and Bagosora's reported "apocalypse" reaction at Arusha is the most-cited single utterance of the period's hardliner mood. The unified Rwandan national army was to be 19,000 troops at 60% FAR / 40% RPF, with a 6,000-strong gendarmerie at the same ratio; command positions, however, were to rotate 50:50 between the parties, including the most senior posts. For a FAR officer corps of approximately 40,000 personnel at the war's height — including the akazu-aligned senior officers around Colonel Théoneste Bagosora — the protocol meant massive demobilisation and the loss of military instruments of regime survival to a 40% RPF share of force with 50% of command. Bagosora's reported reaction at Arusha — "Je vais préparer l'apocalypse" / "I am going to prepare the apocalypse" — is widely cited (Des Forges 1999; Melvern 2004; Guichaoua 2015) [TBD-VERIFY: precise wording across sources varies between "I am returning to Kigali to prepare the apocalypse" and the shorter form; the sentiment is undisputed].
The 600-strong RPF battalion deployed to the Conseil National de Développement (CND) building in central Kigali in late December 1993 under UNAMIR escort — under the Arusha provisions requiring the RPF political-military leadership to take its place in the transitional institutions — became, during the eight-month non-implementation period and the subsequent 100 days of genocide, the decisive RPF military asset in the capital. The CND deployment, conducted between 28 December 1993 and early January 1994, transferred approximately 600 RPF soldiers with sealed weapons stockpiles from the northern zone to a fortified building one kilometre from Habyarimana's official residence. When the war resumed on the night of 6–7 April 1994, the CND battalion broke out of its quarters within hours of the shoot-down, fought through the Garde Présidentielle and gendarmerie cordons surrounding it, established a defensible perimeter in the city's northern districts, and rescued substantial numbers of Tutsi and moderate Hutu civilians sheltering at the Amahoro Stadium and the Mille Collines hotel through the next two weeks while the broader RPF force advanced from Mulindi. Without the CND deployment the RPF would have begun the 100-day campaign with no foothold in the capital and the rescue operation at the Mille Collines and elsewhere would not have been mounted in the form it took.
The 11 January 1994 cable from UNAMIR Force Commander Lieutenant-General Roméo Dallaire to UN DPKO — the so-called "genocide fax" — is the most-cited documentary instance of the international warning of the prepared genocide that was ignored, and its institutional history is the documentary spine of the Carlsson Report's verdict of UN-Secretariat failure. Dallaire's cable, transmitted from Kigali on 11 January 1994 and addressed to Major-General Maurice Baril at the DPKO Military Adviser's office and forwarded to Under-Secretary-General Kofi Annan, reported the testimony of an Interahamwe trainer informant (codenamed "Jean-Pierre", subsequently identified as Jean-Pierre Abubakar Turatsinze [TBD-VERIFY: precise identification across sources]) — that the Interahamwe were maintaining arms caches in Kigali, that the militia was being trained to kill Tutsi at a rate of "1,000 in 20 minutes", and that lists of Tutsi for extermination were being compiled — and requested authorisation to seize the caches under UNAMIR's mandate. The DPKO reply, drafted by Iqbal Riza for Annan's signature, instructed Dallaire not to seize the caches on the ground that such action would exceed the Chapter VI mandate, and to share the informant's information with Habyarimana and the Belgian, French, and US ambassadors. The Carlsson Report (1999) treats the exchange as the documentary core of the Secretariat's failure.
The 21 October 1993 assassination of the elected Hutu president of Burundi, Melchior Ndadaye, by elements of the Tutsi-dominated Burundian army was the external shock that transformed the prospect of Arusha implementation from fragile to structurally improbable, and the corpus reads it as the single most-consequential exogenous event of the eight-month non-implementation period. Ndadaye — in office only since 10 July 1993, the first elected Hutu president in Burundian history — was killed in a failed coup whose aftermath produced civil-war killings in Burundi estimated at tens of thousands across the subsequent months and a refugee outflow into southern Rwanda of approximately [TBD-VERIFY: 300,000 in late 1993, with estimates running to 600,000 by early 1994]. Inside Rwanda the killing was instantly weaponised by Hutu Power propaganda — and particularly by RTLM (Radio Télévision Libre des Mille Collines), which had begun broadcasting on 8 July 1993 thirteen days before the Arusha signature — as definitive proof that political power-sharing with Tutsi was fatal for Hutu. Within weeks of Ndadaye's killing the MDR, PL, and PSD opposition parties had each split into "moderate" and "Hutu Power" wings whose competing claims to BBTG portfolios became one of the principal procedural obstructions to implementation.
The 100-day military campaign from 7 April 1994 to 4 July 1994 — by which the RPF advanced from its northern bases at Mulindi and Byumba and from the CND battalion's foothold in Kigali to the capture of the capital and the rolling collapse of the génocidaire regime southward through Gitarama (13 June) and to the seizure of Butare (3 July), Cyangugu, Gisenyi (17 July), and the final lines on the Zairian border — is the only operationally successful military halting of an ongoing genocide in twentieth-century history, and it was conducted under conditions of comprehensive international non-intervention. UNAMIR's strength was reduced from 2,548 to 270 by Security Council Resolution 912 of 21 April 1994 after the 7 April killing of ten Belgian peacekeepers at Camp Kigali precipitated the Belgian KIBAT withdrawal; the proposed UNAMIR II reinforcement under Resolution 918 of 17 May was not deployed in time; the French Opération Turquoise of 23 June – 21 August 1994 (Resolution 929) created a "humanitarian protection zone" in the southwestern prefectures that the corpus following Wallis (2007) and Prunier (1995) records as having protected fleeing civilians but also having permitted the orderly withdrawal of génocidaire leadership into Zaire. The RPF's military halting of the genocide was therefore conducted substantially alone, and the contestation of the campaign's own conduct — including allegations of RPF reprisal killings of civilians in zones captured during the advance, documented in part by the 2010 UN Mapping Report and by Des Forges (1999) — is part of the record that the post-1994 corpus must hold alongside the genocide-halting achievement.
The corpus records three accounts of the war which it does not adjudicate. The RPF-official account treats it as the just armed return of a thirty-one-year stateless diaspora, the political-military instrument by which the only operationally successful halting of an ongoing genocide was conducted, and the foundation of the post-1994 reconstruction documented across the C, D, E, and G blocks of the Rwanda corpus. The Hutu Power residual account — surviving in some Francophone polemical and revisionist literature, in elements of the Bruguière hypothesis on the shoot-down, and in the post-1994 génocidaire-diaspora networks reorganised in eastern Zaire from 1994 — treats the war as an unprovoked Ugandan-backed Tutsi invasion that triggered the regime's defensive radicalisation and whose own conduct included civilian-targeting comparable to that of the regime. The international-observer account — Adelman and Suhrke (1999), Dallaire (2003), Jones (2001), the Carlsson Report (1999), and the broader scholarly literature — reads the war as a tractable conflict whose Arusha-process mediation was structurally under-resourced (the CDR exclusion as spoiler; the inadequate UNAMIR mandate; the Belgian-and-American refusal to expand it after the January 1994 warnings; the post-Somalia post-PDD-25 constraints on US engagement), whose collapse was a documented diplomatic failure as much as a sabotage by hardliners, and whose military terminus was determined less by any external intervention than by the operational quality of the RPF and the political bankruptcy of the génocidaire regime by mid-June 1994. The corpus follows the third reading in declining to adjudicate between the first two while documenting all three.

2. The Long Diaspora: The 1959–1990 Tutsi Refugee Exodus and the Politics of Statelessness

The 1990 invasion is unintelligible without the thirty-one-year refugee politics that produced its officer corps. The cumulative outflow of Tutsi from Rwanda across the 1959–1964 violence — the November 1959 muyaga documented in RW-A-03, the 1960 Belgian-supervised electoral transfer, the 1961 Coup of Gitarama, the 1962 independence under PARMEHUTU, and the cyclical anti-Tutsi pogroms triggered by the 1961–1967 Inyenzi incursions — produced a Tutsi exile population whose principal scholarly estimate by the late 1980s was approximately 600,000 to 700,000 persons, including the children and grandchildren born in exile. [TBD-VERIFY: the figure is variously reported across UNHCR estimates, RPF political claims, and Rwandan government counter-claims; the lower bound is approximately 480,000 (UNHCR 1989) and the upper bound approximately 800,000 (later RPF estimates including children born in exile after 1980); the corpus follows the mid-range Prunier (1995) and Mamdani (2001) estimates.] The dispersion was uneven. The largest single population was in Uganda, principally in the Ankole and Toro regions of the southwest (the Nshungerezi, Oruchinga, and Nakivale settlements), with the most concentrated demographic mass approximately 200,000 to 280,000 persons by the mid-1980s. Burundi held perhaps 150,000 to 200,000; Tanzania approximately 20,000 to 40,000 in the Ngara and Karagwe districts; Zaire (eastern) approximately 80,000 to 120,000 in North and South Kivu. Smaller diaspora communities in Kenya, Belgium, France, Canada, and the United States contributed a few tens of thousands in total.

The political condition of the exiled Tutsi was the central fact: they were stateless. The host countries did not extend citizenship. The 1962 Rwandan independence settlement had constitutionally defined Rwandan political belonging in Hutu-majoritarian terms; the post-1959 confiscation of Tutsi property in Rwanda was treated by successive Kayibanda and Habyarimana governments as definitive; the right of return was denied. In 1986 — at the height of the Ugandan-Rwandan refugee tensions documented below — President Habyarimana, asked at a press conference whether the refugees might be permitted to return, replied that Rwanda was a small densely-populated country that could not accommodate them and that the refugees should accept naturalisation in their countries of asylum. The statement, widely reported in the Rwandan-Tutsi diaspora press, was the political ratification of a foreclosure that had operated since 1962. Mahmood Mamdani's When Victims Become Killers (2001, Chapter 5) frames this foreclosure as the constitutive condition of the RPF's later existence: a refugee population denied citizenship in either the host countries or the homeland is, by structural arithmetic, a politics of armed return.

The Ugandan dimension was the operational pivot. The Tutsi-refugee community in southwestern Uganda — descended principally from the 1959–1964 flight and from a secondary wave in 1972–1973 — was, by the 1970s, demographically substantial, economically integrated in the cattle-and-trade sectors of the Ankole region, and politically vulnerable. Successive Ugandan governments treated the Rwandan refugees as a security category. Idi Amin (1971–1979) tolerated the refugees in part because his own Kakwa-and-Nubian power base shared certain marginal characteristics with the Rwandans, but the late-Amin years produced increasing anti-refugee sentiment. The decisive turning point was the second Obote regime (1980–1985): Milton Obote's Uganda People's Congress government, returning to power in the disputed December 1980 election, scapegoated Rwandan refugees as Amin collaborators and as alleged supporters of the new armed opposition. The October 1982 expulsions — the so-called Nshwerenkye operation in which UPC militants drove approximately 40,000 to 80,000 Rwandan refugees [TBD-VERIFY: figure variously reported across sources, with mid-range scholarly estimates around 40,000–60,000 expelled, and roughly half of those returning to Uganda after several months when the Rwandan government refused permanent re-entry] from the Ankole and Mbarara districts back toward the Rwandan border — produced the radicalising experience that consolidated the Rwandan-refugee youth into armed politics.

The expulsions were the political founding moment of the RPF's officer-cohort. The young Rwandan-refugee men of the Ankole-Tutsi community — including Fred Rwigyema, Paul Kagame, Tito Rutaremara, Patrick Mazimpaka, and Alexis Kanyarengwe (then still inside Rwanda and the Habyarimana government, defecting to the RPF later in the 1980s) — encountered in 1982–1983 a situation in which neither the host country nor the homeland was a politically viable destination. The choice for return became binary: an armed politics, or permanent statelessness. The first vehicle of the armed politics had already been established in 1979: the Rwandese Alliance for National Unity (RANU), founded in Kampala in 1979 by exiled Tutsi intellectuals as a political organisation committed to the principle of return-as-right. RANU was the predecessor of the RPF. Through the 1980s its membership grew alongside the Ugandan-Tutsi cohort's military service in Museveni's National Resistance Army.

3. The Ugandan Crucible: RANU, the NRA Bush War, and the Founding of the RPF (1979–1990)

The transformative experience for the future RPF officer cohort was their participation in Yoweri Museveni's National Resistance Army (NRA) bush war (1981–1986). When Museveni — himself partly of Banyarwanda Tutsi descent, with extensive kinship and political ties to the Rwandan-refugee community of Ankole — went into the bush in February 1981 after the disputed December 1980 election to launch armed opposition to the Obote government, the small initial cohort of fighters who began the campaign included a substantial Rwandan-refugee contingent. The often-repeated "27 founders" or "26 founders" figure of Museveni's bush-war origin story includes, by widely cited accounts, Fred Rwigyema as one of the founding cohort [TBD-VERIFY: the "27" or "26" figure is a Museveni-movement narrative trope and its precise composition has been variably reported; what is uncontested is that Rwigyema was among the small initial cohort and that he rose rapidly within the NRA command]. Paul Kagame joined the NRA in 1981 [TBD-VERIFY: precise date of his joining the NRA, commonly given as 1981–1982] and rose, by the time of the NRA's January 1986 capture of Kampala, to head the NRA's military-intelligence apparatus.

The five-year NRA campaign was the operational formation of the future RPF. The fighters who would lead the 1990 invasion were trained, blooded, and disciplined in the NRA's protracted guerrilla war against the Obote and Tito Okello governments. The doctrine — disciplined small-unit operations, deep penetration of the countryside, cultivation of civilian populations through codes of conduct on the "Ten Commandments of the NRA" model, intelligence-led targeting of regime forces, and a long political-military patience — was the doctrine the RPF would import to Rwanda. The institutional architecture — a political wing (the National Resistance Movement, the NRM) and a military wing (the NRA), each with their own command structures but unified at the apex under Museveni — was the architecture the RPF would replicate in the political-RPF / military-RPA distinction.

When the NRA captured Kampala on 25–26 January 1986 and Museveni was sworn in as President of Uganda on 29 January 1986, the Rwandan refugees who had fought for him rose with the movement. Fred Rwigyema was appointed Deputy Commander of the NRA and Deputy Minister of Defence in the Ugandan government — extraordinary positions for a non-Ugandan officer and the most senior posts ever held by a Rwandan refugee in a host-country government. Paul Kagame became Head of Military Intelligence in the NRA [TBD-VERIFY: precise title and dates; commonly reported as Acting Head of Military Intelligence and Counter-Intelligence from 1986 to 1989, with subsequent senior intelligence-officer posts through 1990]. Chris Bunyenyezi became commander of an NRA brigade; Peter Bayingana became the NRA's Director of Medical Services; Sam Kaka rose through the NRA armoured corps; Adam Wasswa, Kayumba Nyamwasa, Patrick Nyamvumba, and others held substantial NRA commands. By 1989 the Rwandan-Tutsi cohort within the Ugandan army was approximately 2,000 to 4,000 personnel [TBD-VERIFY: precise figure variously reported].

The prominence of Rwandan refugees inside the NRA, however, became a domestic political liability for Museveni. The Ugandan political opposition, the Uganda People's Congress remnants, and Banyankole rivals of Museveni began across 1988–1990 to attack the Museveni government as Rwandan-dominated. Pressure mounted on Museveni to demobilise or remove the Rwandan officers. The political-military cohort responded by accelerating its parallel project: the armed return to Rwanda. The founding of the Rwandan Patriotic Front (RPF) was the institutional expression of that decision.

The founding took place in Kampala in December 1987. The 1987 Congress of the RANU — held over three days, with delegates from the Ugandan, Burundian, Tanzanian, Zairean, Kenyan, European, and North American Rwandan-refugee communities — voted to rebrand the organisation as the Rwandan Patriotic Front with a self-consciously non-ethnic, return-oriented political programme. The first political chairman was Alexis Kanyarengwe — a former Habyarimana ally and northern-Hutu MRND politician who had defected to the exiles in the late 1980s, providing the RPF with a Hutu political face that countered the regime's framing of the movement as exclusively Tutsi. The first vice-chairman and political coordinator was Pasteur Bizimungu, another former Habyarimana insider and a Hutu defector. The military command, however, was Tutsi: Fred Rwigyema as the most senior military officer of the movement (still serving simultaneously as Deputy Commander of the NRA); Paul Kagame as the senior intelligence officer; the operational majors (Bayingana, Bunyenyezi, Kaka, Bagire, Banyenzaki, and others) as the officer corps.

Through 1988–1990 the RPF prepared for invasion. The preparation was at once political (the recruitment of Hutu defectors to provide a non-ethnic political face), diplomatic (briefings of Western diplomats in Kampala, Brussels, Washington, and Geneva on the refugee question; the 1988 OAU resolution on the Rwandan refugee problem; the July 1990 Mwanza summit of Rwanda, Uganda, Tanzania, and the OAU on the right-of-return question that produced no operative commitment from Habyarimana), and military (training within the NRA structure, the accumulation of weapons and ammunition, the cultivation of intelligence on the Rwandan border zone). The decision to invade was taken in mid-1990 against a backdrop of three converging pressures: Museveni's increasing domestic political need to disengage from the Rwandan-officer cohort, Habyarimana's continued refusal of the right of return at the July 1990 Mwanza summit, and the regime-internal political crisis in Rwanda (the multi-party reform pressure under donor conditionality, the Bahutu / Banyiginya succession rumours, the deteriorating economic conditions) that the RPF read as an opening.

At the moment the invasion was scheduled, however, Kagame was not in the region. He had been sent for advanced military training at the U.S. Army Command and General Staff College at Fort Leavenworth, Kansas, on a US-military exchange programme [TBD-VERIFY: precise dates of the Fort Leavenworth course, commonly given as mid-1990, with departure from Uganda in approximately June–July 1990 and an expected return in early 1991]. The course was a year-long senior-officer programme covering strategic doctrine, operational planning, and joint-staff procedures — the standard professional development for officers being prepared for general-staff command. That Kagame was at Leavenworth when the RPF chose to invade is the structural reason Rwigyema commanded the operation; whether Kagame's absence was an operational error, a deliberate division of labour, or a coincidence of training schedules and political timing is among the questions the historiography does not definitively resolve.

4. The 1 October 1990 Invasion: Mutara, the Kagitumba Crossing, and the Three-Day Disaster

The invasion was launched on 1 October 1990, the morning of a Monday, with approximately 2,500 to 4,000 fighters [TBD-VERIFY: figure variously reported across sources; Prunier (1995) gives 2,500, Reyntjens (1994) approximately 4,000, Rusagara (2009) approximately 4,000 RPA personnel including support units] crossing the Uganda–Rwanda border at the Kagitumba customs post in the northeastern Mutara region. The operation was conventional in form: armoured-vehicle-supported infantry columns advancing along the principal roads, with the operational objective the seizure of the Mutara hills and a subsequent thrust toward the Byumba prefectural capital. The fighters were uniformed in NRA-pattern dress; the equipment was substantially Ugandan-NRA stocks. Many of the fighters were technically Ugandan army deserters who had taken their weapons with them — a fact that gave the Habyarimana regime, the French government, and the Belgian government their initial framing of the invasion as an act of Ugandan state aggression rather than a Rwandan refugee return. The framing was not without basis: Museveni's government, even if not the proximate operational author, had at minimum tolerated the preparation and the cross-border movement. The Ugandan position throughout October 1990 — that the invasion was the unauthorised action of deserting officers, that Uganda had not sanctioned it, and that Uganda would seal the border — was politically convenient for Kampala but only partly truthful.

The initial RPF order of battle was substantial but operationally thin. The conventional advance into the Mutara plain encountered the Forces Armées Rwandaises (FAR) at the Kagitumba border post and along the Byumba road, where regime resistance was initially modest — the FAR's eastern-zone Gendarmerie units had no preparation for a cross-border conventional attack and fell back in the first twenty-four hours. By the morning of 2 October 1990 the RPF had advanced approximately 30–40 kilometres into Rwanda along the principal northern axes. The operation was, in the first thirty-six hours, an operational success.

Then Rwigyema was killed. The death occurred on 2 October 1990 at a location reported variously as the Kagitumba front, the Nyabwishongwezi area east of Nyagatare, or at a forward command post near the advance line [TBD-VERIFY: precise location varies across sources; Otunnu (1999) gives Nyabwishongwezi; Rusagara (2009) gives the Kagitumba front]. The official RPF account, which has remained substantially consistent since 1990, holds that Rwigyema was killed by a stray bullet — either from FAR positions during the assault or from incidental fire — while leading the forward command from an exposed position. The contrary account, advanced principally by Ugandan oppositional sources, by some former RPF officers cited in Wrong (2021), and by elements of the Habyarimana-government propaganda apparatus at the time, holds that Rwigyema was killed amid an internal command dispute with the senior Majors Peter Bayingana and Chris Bunyenyezi over operational doctrine — specifically, whether the force should continue the conventional advance toward Byumba (Rwigyema's reported preference) or adopt a more dispersed deep-penetration posture (Bayingana–Bunyenyezi's reported preference). The second account holds that Bayingana and Bunyenyezi were subsequently executed at Kagame's order specifically because they had killed Rwigyema; the first account holds that they were executed for the factional dissent that, in the context of military emergency, was treated as mutiny. [TBD-VERIFY: the corpus records the contestation and does not adjudicate; the post-1990 RPF historiography is the principal source for the first account, while Wrong (2021) and elements of Wallis (2007) record the second account as a persistent counter-narrative.]

Within thirty-six hours of Rwigyema's death, the operational situation began to deteriorate. The FAR had absorbed the initial blow; the Belgian government dispatched approximately 350 paratroopers under Operation Greenbeam / Vert-coded mission to evacuate European nationals from Kigali (with arrival approximately 4 October); the French government activated the long-standing 1975 Franco-Rwandan military cooperation agreement and deployed an initial contingent of approximately 300 troops of the 2nd Foreign Parachute Regiment under Operation Noroît, arriving in Kigali on 4 October 1990, with operational coordination at Kanombe air base; the Zairean government of Mobutu Sese Seko, prompted by Habyarimana's personal appeal to Mobutu as the regional patron, dispatched approximately 500 troops of the Special Presidential Division (DSP) to reinforce the FAR. The Zairean force performed poorly — the DSP units engaged in widespread looting and indiscipline, and were quickly withdrawn under Habyarimana's complaint that they were a liability rather than a reinforcement — but the French and Belgian deployments altered the operational balance decisively. By 4–5 October 1990 the FAR, with French and Belgian rear-area cover, was conducting an effective counter-attack along the principal northern axes.

The RPF's domestic political effect inside Rwanda was immediate and devastating. On the night of 4–5 October 1990, the Habyarimana government staged what would be known as the "phantom attack on Kigali" — the so-called attaque de Kigali — in which heavy small-arms and mortar fire was reported across the capital, ostensibly demonstrating that an RPF infiltration had reached the city. No RPF forces were present; the RPF at that moment was being driven back from Mutara in the northeast under combined FAR-French-Belgian pressure. The "phantom attack" was a staged event: FAR units fired into the air across Kigali through the night, and the regime announced in the morning that the capital had been attacked by RPF infiltrators in disguise. The political purpose was the justification of mass arrests. Over the following weeks the government detained approximately 8,000 to 10,000 suspected RPF ibyitso (accomplices) — overwhelmingly Tutsi civilians and a smaller number of opposition Hutu politicians — in the Nyamirambo stadium, the Gikondo prison, the Camp Kigali military installation, and other improvised detention facilities. The detainees were held without charge for periods ranging from several days to several months; [TBD-VERIFY: dozens to several hundred died of overcrowding, beatings, and medical neglect, with the precise toll variously reported across Africa Watch (1993), African Rights (1995), and Des Forges (1999); the lower bound is approximately 40 deaths in detention, the upper bound approximately 500]. The arrests were the regime's first systematic use of the ibyitso propaganda category that would dominate the political-discursive landscape of the next four years and culminate in the 1994 genocide.

The RPF's first conventional invasion had, by 9–10 October 1990, effectively failed. The FAR had pushed the residual force back toward the Ugandan border; the senior command was dead or wounded; the operational logistics had collapsed; perhaps 1,000 to 1,500 of the original 2,500–4,000 fighters had been killed, wounded, captured, or had deserted; the political objective of forcing the regime to a negotiating table had been achieved only insofar as the invasion had transformed Rwandan domestic politics into a war footing, but not in the form the RPF leadership had anticipated. The conventional invasion ended in defeat within ten days. What followed was a different war.

5. Kagame's Return from Fort Leavenworth, the Bayingana–Bunyenyezi Executions, and the Doctrinal Transformation (October–December 1990)

The news of Rwigyema's death reached Paul Kagame at Fort Leavenworth on 2 or 3 October 1990 [TBD-VERIFY: precise dating; sources variously report that Kagame was informed within twenty-four hours of the death through embassy channels and that he requested immediate departure from the course]. Kagame departed Leavenworth on or about 15 October 1990 — completing some six weeks of the year-long course — and travelled through Washington, Brussels, and Entebbe, arriving in the Uganda–Rwanda border zone in the last week of October 1990. By the time he reached the front, the RPF had been pushed back across the border into Uganda; the residual force was scattered across the Mbarara and Kabale districts; the survivors were demoralised, under-equipped, and operationally ineffective.

Kagame's first operational acts were brutal and decisive. He convened a senior command meeting at a location in southwestern Uganda [TBD-VERIFY: meeting location commonly reported as the Mbarara area or the Lake Mutanda district] and ordered the execution of Majors Peter Bayingana and Chris Bunyenyezi [TBD-VERIFY: precise dates of the executions, commonly reported as late October 1990 or early November 1990; both were executed by firing squad on RPF discipline charges]. The official charges were factional dissent, indiscipline in the field, and contributing through their command failures to the operational collapse of the initial advance. The unofficial accusation — that Bayingana and Bunyenyezi had killed Rwigyema or had facilitated his death — was widely reported in the Ugandan press at the time and has remained a persistent counter-narrative in the historiography. [TBD-VERIFY: the official RPF account denies any direct connection between the Rwigyema death and the executions, while critical accounts including elements of Wrong (2021) record the connection as widely understood within the surviving RPF officer corps; the corpus records the contestation.] Kagame's executions of the two majors served three functions simultaneously: they enforced his command authority over a fractious officer corps, they eliminated potential rivals for the succession to Rwigyema, and they signalled to the surviving RPF cadres that the standard of discipline under the new command would be absolute.

The doctrinal transformation followed. Kagame's strategic assessment, formed across late October and November 1990 and articulated to the RPF political wing under Kanyarengwe and Bizimungu, was that the conventional invasion strategy had failed comprehensively and that a complete reorientation was required. The new strategy — drawing on Kagame's NRA bush-war experience, on the Fort Leavenworth coursework, and on a sober reading of the relative force balances — would be a sustained guerrilla campaign operating from a secure mountain base, building outward through gradual territorial consolidation, designed to wear down the FAR over a multi-year horizon rather than to capture Kigali in a single thrust. The conventional brigade structure was dissolved; the residual force was reorganised into small mobile units; the political-military separation between the RANU/RPF political wing in Kampala and the operational RPA in the field was sharpened.

The secure base chosen for the reconstitution was the Virunga volcanic chain along the northern Rwandan border — the chain of high-altitude volcanoes (including Mount Karisimbi at 4,507 metres, Mount Mikeno, and Mount Visoke) that runs along the Rwanda–Uganda–Zaire frontier and that had served as the operational base of armed groups in the region since the 1960s Inyenzi incursions. The geography was favourable: dense forest at the lower elevations, cold and broken terrain at the upper elevations, multiple crossing points across three international borders, no FAR infrastructure within the mountain interior, and the population of the surrounding kiga highland districts (a Hutu peasant population whose political relationship with the Habyarimana regime was complex — the kiga north was Habyarimana's home region but was also chafing under the akazu faction's monopoly of regional resources). Between November 1990 and February 1991 the RPF reorganised in the Virungas. Conditions were severe: extreme cold, food scarcity (the RPF survived in this period substantially on resources smuggled from Uganda and on what could be acquired from the surrounding population), and continued FAR pursuit. [TBD-VERIFY: RPF strength estimates for the Virunga period vary; Kinzer (2008) reports approximately 2,000 fighters at the nadir, with growth to 5,000 by mid-1991; Rusagara (2009) reports approximately 2,500 at the lowest point.]

The political-military reconstitution in the Virungas is the institutional moment from which the modern Rwandan Patriotic Army — and the post-1994 Rwandan Defence Force — derives its founding mythology. The "Virunga generation" of officers (those who survived the winter of 1990–1991 alongside Kagame) became the senior cohort of the post-1994 RPA-and-RDF, with continuity of command lines visible into the 2020s. The doctrinal discipline established in the Virungas — strict codes of conduct toward civilian populations, harsh penalties for indiscipline, intelligence-led targeting, refusal of forced recruitment, cultivation of long-arc political-military patience — was the operational culture of every subsequent RPA campaign, from the 1991–1993 northern operations through the 1994 100-day campaign and into the 1996 First Congo War. Kagame's personal authority within the movement, in dispute through November 1990, was by February 1991 effectively absolute.

6. Mulindi and the N'Sele Ceasefire (January – March 1991)

The transition from the Virunga reconstitution phase to a sustained operational phase came in early 1991. The RPF descended from the volcano interiors to establish a permanent operational headquarters at Mulindi, a tea-estate complex in the northern Byumba prefecture roughly 60 kilometres north of Kigali that the RPF would occupy through the next three years. Mulindi was the RPF's political-and-military capital from 1991 to mid-1994. It was the location at which Kagame, Kanyarengwe, Bizimungu, and the senior cadre conducted the war's daily command operations; from which the 8 February 1993 offensive was planned; to which Western diplomats, journalists, and humanitarian-organisation representatives travelled when seeking direct engagement with the RPF leadership; and from which the December 1993 deployment of the 600-strong RPF battalion to the Conseil National de Développement (CND) building in Kigali was launched.

The shift from Virunga base to Mulindi political-military capital coincided with the first formal ceasefire of the war. The N'Sele Ceasefire Agreement was concluded at Zairean President Mobutu Sese Seko's N'Sele residential complex outside Kinshasa on 29 March 1991, with mediation conducted under OAU auspices and the participation of Zaire, Burundi, Uganda, Tanzania, and Belgium. The agreement was thin in operative content — a cessation of hostilities, the establishment of a Neutral Military Observer Group (NMOG) under OAU command, and a commitment to subsequent political negotiations — but it was institutionally significant: it confirmed that the war would proceed politically through African-led mediation rather than through Western or UN frameworks, and it set the precedent on which the Arusha process documented in RW-A-04 would build. The text would be successively updated at Gbadolite (16 September 1991), Bujumbura (February 1991 and March 1992 — the date sequence is variably reported across sources), and finally absorbed into the Arusha I cessation-of-hostilities agreement of 12 July 1992, by which point the mediation centre of gravity had shifted from Mobutu's personal Zairean patronage to the institutional OAU framework under Tanzanian leadership.

The 29 March 1991 ceasefire did not hold for long. Within weeks the RPF had resumed low-intensity operations from the Mulindi base — small-unit raids on FAR positions in northern Byumba, ambushes along the main roads, intelligence operations among the Hutu peasant population of the northern districts. The operational character of the second phase of the war — April 1991 to January 1992 — was qualitatively different from the conventional October 1990 invasion: it was an insurgency, conducted at the operational tempo and scale that Kagame's doctrinal transformation had specified, and it gradually built RPF territorial control in the northern Byumba and Ruhengeri prefectures while imposing on the FAR a counter-insurgency burden the regime was structurally ill-prepared to bear. By the end of 1991 the RPF effectively controlled a continuous strip of northern Byumba and parts of northern Ruhengeri — territory denied to the regime's civil administration and from which the RPF could project operations southward.

The foreign military dimension hardened in parallel. Operation Noroît — the French military deployment that had begun on 4 October 1990 with approximately 300 troops — was expanded across 1991 to a sustained presence of approximately 600 French personnel based principally at Kanombe air base in Kigali, with rotation of units from the 8th Marine Infantry Parachute Regiment and the 2nd Foreign Parachute Regiment. The French presence was nominally limited to the protection of European nationals and the training of FAR units; in operational practice, Wallis (2007) and the 2021 Duclert Commission Report documented that French personnel provided intelligence support, equipment supply, doctrinal instruction in counter-insurgency, and on occasion direct participation in FAR operational planning. The Belgian presence was withdrawn earlier — the KIBAT evacuation contingent of October 1990 was withdrawn in November 1990 — but Belgian military advisers continued to operate in smaller numbers through 1992. The Zairean contingent, ineffective in operational terms, was withdrawn in late October 1990. By mid-1991 the Habyarimana regime's external military backing was substantially French; the Franco-Rwandan relationship was the structural condition of the regime's survival through the war's middle phases.

7. The Northern Offensives, the State-Orchestrated Reprisal Massacres (1990–1992), and the Ibyitso Architecture

The most significant political-military development of the war's middle phases (1990–1992) was not in the military dimension but in the domestic-political dimension inside Rwanda: the regime's response to the invasion was a systematic radicalisation of anti-Tutsi politics that produced cyclical state-orchestrated massacres of Tutsi civilians and an unprecedented propaganda apparatus that prepared the political ground for the 1994 genocide. This radicalisation was not a contingent overreaction; it was a deliberate strategy adopted by the akazu-aligned faction of the regime and progressively normalised across the broader political-military establishment.

7.1 The Kibilira Massacres (11–13 October 1990)

The first of the war-period massacres was the Kibilira massacres in the Gisenyi prefecture, conducted in the commune of Kibilira between 11 and 13 October 1990 — nine days after the invasion and three days after the staged "phantom attack" on Kigali. The killings were organised by the local bourgmestre and the prefectural authorities; the operational executors were Hutu peasants mobilised by communal officials under the ibyitso framing — that Tutsi villagers were RPF accomplices and a security threat requiring elimination. Approximately 350 to 400 Tutsi civilians were killed across the three days [TBD-VERIFY: figure variously reported, with Africa Watch (1993) giving approximately 350, Des Forges (1999) approximately 400, and African Rights (1995) approximately 350–500]. The killings were conducted with the participation of the communal réservistes, the gendarmerie, and Hutu civilians armed with traditional weapons. The political significance of the Kibilira massacres was that they established the operational template of the regime's anti-Tutsi violence: prefectural-and-communal organisation, ibyitso propaganda framing, mass civilian-perpetrator participation, and limited national-level disavowal followed by no prosecution. Every element of the template would recur, at progressively larger scales, through the Bagogwe and Bugesera massacres and ultimately in the 1994 genocide.

7.2 The Bagogwe Massacres (January–March 1991)

The Bagogwe massacres in the northwestern prefectures of Gisenyi and Ruhengeri, conducted between January and March 1991, targeted the Bagogwe — a Tutsi pastoral subgroup historically resident in the kiga highlands of the northwest and culturally distinct from the central-and-southern Tutsi communities of the kingdom heartland. The operational trigger was the 23 January 1991 RPF attack on Ruhengeri, in which an RPF assault force temporarily seized the Ruhengeri prison (releasing the imprisoned former regime figure Major Théoneste Lizinde [TBD-VERIFY: name and circumstances of the Ruhengeri prison break]) before withdrawing under FAR counter-attack. The assault triggered a wave of regime-organised retaliatory killings across the surrounding northwestern districts. The Bagogwe massacres ran across some six weeks of intermittent killings in approximately fifteen to twenty separate incidents across multiple communes in Gisenyi and Ruhengeri prefectures. Approximately 300 to 1,000 Bagogwe Tutsi were killed [TBD-VERIFY: precise toll varies substantially across sources; Africa Watch (1993) gives approximately 300, Des Forges (1999) approximately 500, African Rights (1995) approximately 1,000; the wide range reflects the dispersed and clandestine character of the killings, conducted in isolated rural locations and with substantial post-killing concealment of evidence by the local authorities]. The killings were organised at the préfet level by Protais Zigiranyirazo [TBD-VERIFY: Zigiranyirazo's prefectural appointment and direct command role in the Bagogwe killings, as documented at the ICTR; he was prosecuted at the ICTR and acquitted on appeal in 2009 on certain charges relating to the 1994 genocide, with his role in the 1991 Bagogwe events documented in the trial record but not separately prosecuted].

7.3 The Bugesera Massacres (4–9 March 1992)

The Bugesera massacres in the southern Bugesera region (Kigali-rural prefecture, the region south of Kigali toward the Burundian border), conducted between 4 and 9 March 1992, marked the geographic extension of the regime's anti-Tutsi violence from the northwestern kiga heartland into the southern districts that had been politically central to the First Republic. The trigger was a series of propaganda broadcasts on Radio Rwanda in early March 1992 announcing a fabricated Tutsi plot to assassinate Hutu political leaders — a propaganda template directly precedent to the RTLM broadcasts of 1993–1994. The broadcasts were authored or sponsored by elements of the akazu-aligned propaganda network around Ferdinand Nahimana (then Director of ORINFOR, the Rwandan office of information, subsequently a co-founder of RTLM and convicted at the ICTR in the 2003 Media Trial judgement). The operational killings began on 4 March and ran through 9 March, with approximately 277 to 300 Tutsi civilians killed [TBD-VERIFY: precise toll commonly reported as 277 to 300 across sources, with Africa Watch (1993) giving 277, Des Forges (1999) 300, and African Rights (1995) approximately 300]. The operational executors included the Interahamwe — the youth militia of the MRND, founded in 1990 and given its first large-scale operational deployment in Bugesera. The Bugesera killings are the most operationally direct rehearsal of the 1994 genocide: the propaganda trigger, the Interahamwe deployment, the geographic dispersion, the ibyitso framing, and the absence of national-level prosecution were all present in exactly the form they would assume in April–July 1994.

7.4 The Ibyitso Architecture and the Propaganda Apparatus

Around the massacres, an unprecedented anti-Tutsi propaganda apparatus consolidated across 1990–1993. The most-cited single document is the "Hutu Ten Commandments" (Les Dix Commandements des Bahutu), published in the December 1990 issue (No. 6) of the extremist Kinyarwanda-language periodical Kangura under editor Hassan Ngeze. The Commandments, framed as an explicit racial-ideological programme, declared all Tutsi women to be enemy infiltrators, prohibited Hutu from doing business with Tutsi, declared Hutu women to be the only acceptable wives, declared the army a Hutu institution, and called for the political-military exclusion of Tutsi from public life. The document is widely identified in the historiography (Des Forges 1999; Melvern 2004; Chrétien et al. 1995) as the foundational text of Hutu Power ideology. Its publication ten weeks after the invasion and immediately following the ibyitso arrests of October–November 1990 was not coincidental: the document was the ideological codification of the operational practice the regime had already adopted.

Kangura continued through 1990–1994 as the principal print-propaganda vehicle of the Hutu Power current. From 8 July 1993 — thirteen days before the Arusha signature — it was joined by Radio Télévision Libre des Mille Collines (RTLM), the privately-owned radio station financed by the akazu-aligned business network and broadcasting in Kinyarwanda and French to a broad audience across central and northern Rwanda. RTLM's broadcasts across July 1993 – April 1994 became progressively more incendiary, with explicit incitement to anti-Tutsi violence, the publication of names and license-plate numbers of targeted individuals, and the framing of every political development in ibyitso terms.

7.5 The International Commission and the Bacre Waly Ndiaye Report (January–August 1993)

The cumulative documentation of the 1990–1992 massacres produced a critical institutional response in early 1993. The International Commission of Investigation on Human Rights Violations in Rwanda — a joint mission of Africa Watch (the predecessor of Human Rights Watch Africa Division), the Fédération Internationale des Ligues des Droits de l'Homme (FIDH), the International Centre for Human Rights and Democratic Development (ICHRDD, Montreal), and the Union Interafricaine des Droits de l'Homme (UIDH) — conducted field investigations across Rwanda in January 1993, hearing testimony from survivors, communal officials, and political figures. The Commission's Final Report, published in March 1993, documented the Kibilira, Bagogwe, and Bugesera massacres in detail, attributed organisational responsibility to identified communal and prefectural officials, and concluded that the regime was conducting a systematic policy of ethnic violence under cover of the civil war. The Commission's rapporteur Bacre Waly Ndiaye — the Senegalese jurist appointed UN Special Rapporteur on Extrajudicial, Summary or Arbitrary Executions — subsequently conducted a UN mission to Rwanda from 8 to 17 April 1993 and filed a report (UN document E/CN.4/1994/7/Add.1) on 11 August 1993 that became the first UN document to invoke the word genocide in characterising the Rwandan trajectory. The Ndiaye Report's text — that "the violence to which the Tutsi minority is subjected cannot be characterised as ethnic strife but rather as the result of premeditated and organised attacks ... cases of intentional mass extermination ... might fall within the bounds of the definition of genocide as set forth in Article II of the relevant 1948 international convention" [TBD-VERIFY: verbatim quotation from the Ndiaye report] — was effectively ignored at the time by both the UN Secretariat and the major Western capitals; its rehabilitation came only after 1994, when the Carlsson Report (1999) cited it as one of the principal documentary instances of warning that had been disregarded.

8. The 8 February 1993 Offensive: From the Ceasefire Lines to 30 km from Kigali

The single largest conventional military operation of the war was launched by the RPF on 8 February 1993. The offensive was the operational rejoinder to the January 1993 anti-Tutsi massacres in the northwestern prefectures — a renewed cycle of regime-organised killings in Gisenyi and Ruhengeri in late January and early February that the RPF leadership at Mulindi judged to constitute a comprehensive ceasefire violation justifying the resumption of conventional operations. The offensive's operational planning, conducted by Kagame and the senior RPA staff at Mulindi over December 1992 – early February 1993, was a deliberate departure from the slow guerrilla-attritional doctrine that had governed the previous fifteen months: it was a multi-axis conventional thrust intended to demonstrate to the regime, to the Arusha mediators, and to the Western diplomatic observers that the RPF could, if it chose, capture Kigali.

The offensive's operational geography ran south and southwest from the Mulindi base across the Byumba prefecture, with secondary thrusts through Ruhengeri prefecture toward Gisenyi and through eastern Byumba toward the Mutara plain. The RPF's strength at the moment of the offensive was approximately 15,000 to 20,000 fighters [TBD-VERIFY: estimates vary across sources from 12,000 (Rusagara 2009) to 25,000 (some Western intelligence estimates of the period)], organised in eight to ten mobile brigades with limited artillery and effectively no air capability. The FAR at the same moment had grown to approximately 30,000 to 40,000 personnel but its operational quality was substantially lower; the units defending the northern axes were principally newly-raised conscript formations whose training and leadership had been compromised by the rapid wartime expansion.

The offensive's operational tempo was sharp. Within forty-eight hours of the 8 February launch the RPF had broken FAR lines across the principal northern axes; within a week the RPF had advanced approximately 40 to 60 kilometres into government territory; by mid-February the RPF advance had reached Ruhengeri in the northwest and was approximately 30 kilometres from Kigali along the southern axis. The FAR was in retreat across the northern theatre; the regime called urgently on France for reinforcement. Operation Noroît was expanded in mid-February with the deployment of approximately 300 additional French troops [TBD-VERIFY: exact reinforcement figures, with Wallis (2007) reporting that French strength rose from approximately 600 to approximately 1,100 personnel in mid-February 1993] and with the dispatch of additional French equipment and intelligence support. The French intervention was decisive in halting the RPF advance: by 20 February 1993 the offensive had stalled along a line approximately 30 kilometres north of Kigali, and the operational opportunity to capture the capital had been foreclosed by the combination of French reinforcement, RPF logistical exhaustion, and OAU diplomatic pressure for a renewed ceasefire.

The civilian-displacement effect was vast and structurally consequential. The advance of the RPF across the northern prefectures produced a southward flight of approximately 800,000 to 1,000,000 civilians [TBD-VERIFY: precise figure variously reported, with mid-range scholarly estimates approximately 850,000; some sources report up to 1.5 million] from the Byumba, Ruhengeri, and northern Kigali-rural districts toward the central and southern prefectures and especially toward the Kigali periphery. The internally displaced persons (IDPs) were concentrated in the Nyacyonga camp north of Kigali (with peak population approximately 350,000 in mid-1993), the Kabuga, Rebero, and Runda camps, and a scatter of smaller settlements across Gitarama, Kibuye, and Butare. The IDP population was overwhelmingly Hutu; it was a constituency that the Hutu Power propaganda apparatus would mobilise across mid-1993 to early 1994 with the framing that the RPF advance demonstrated the existential threat of Tutsi political-military power and the consequent necessity of pre-emptive Hutu mobilisation. The Nyacyonga and Kabuga camps were among the principal recruiting grounds for the Interahamwe training programmes of mid-1993, and the IDP population would be a substantial component of the genocidaire mobilisation in April 1994.

The political-diplomatic effect was the renewed activation of the Arusha negotiating track on terms more favourable to the RPF. The Dar es Salaam ceasefire of 7 March 1993 — concluded between the RPF and the Habyarimana government under Tanzanian mediation — restored a cessation of hostilities along the lines approximately reached by the offensive (with limited RPF withdrawals from forward positions) and reactivated the Arusha protocol negotiations on terms that the RPF leadership at Mulindi judged could now produce the substantial military-integration and refugee-return outcomes the movement had sought from the war's beginning. The 7 March ceasefire is therefore the proximate cause of the 9 June 1993 Refugee Protocol and the 3 August 1993 Armed Forces Protocol that completed the Arusha architecture on terms the regime's hardliners would identify as intolerable. The substantive narrative of those protocols and of the 4 August 1993 Arusha Peace Agreement is in RW-A-04 and is not duplicated here; the present document treats the military-and-political dynamic that produced them.

9. The Arusha Endgame: Military-Integration Negotiations, the 3 August Protocol, and the 4 August Signature

The Arusha negotiating round that ran from the March 1993 Dar es Salaam ceasefire to the August signature was the war's diplomatic culmination. The substantive content of the 18 August 1992 Rule-of-Law Protocol, the 30 October 1992 and 9 January 1993 Power-Sharing Protocols, the 9 June 1993 Refugee Protocol, and the 3 August 1993 Armed Forces Integration Protocol is documented in RW-A-04. The present section treats the military-political logic of the Armed Forces Integration Protocol — the most-discussed and most-bitterly-resisted of the protocols, and the proximate cause of the akazu faction's decision that Arusha implementation was unsurvivable.

The Armed Forces Integration negotiations were conducted at Arusha across June, July, and early August 1993 by mixed RPF-FAR military delegations, with Tanzanian and OAU mediation and US, French, Belgian, and Ugandan observation. The principal RPF negotiator was Colonel Frank Mugambage [TBD-VERIFY: lead RPF military delegate at Arusha; sources commonly report Mugambage and Brigadier Sam Kaka as senior RPF military negotiators]; the principal FAR negotiator was Colonel Théoneste Bagosora, the directeur de cabinet of the Ministry of Defence and the most senior akazu-aligned officer in the FAR command structure. The personal-political tone of the negotiations was extraordinarily bitter. Bagosora's reported reaction during the protocol negotiations — that he was "preparing the apocalypse" / "Je vais préparer l'apocalypse" — is the most-cited single utterance of the period's hardliner mood. [TBD-VERIFY: precise wording, context, and audience of the Bagosora remark varies across sources; Des Forges (1999) reports it as having been made to a Tanzanian or French diplomatic interlocutor during a recess or break from the negotiations; Melvern (2004) and Guichaoua (2015) report it in similar terms; the sentiment is undisputed but the verbatim text is variably recorded.]

The substantive provisions of the 3 August 1993 Armed Forces Integration Protocol were:

Total force strength: 19,000 troops in the unified national army, with a 6,000-strong gendarmerie. The combined figure of 25,000 represented an approximate 40% reduction from the wartime FAR strength of approximately 40,000.
Troop ratios: 60% FAR / 40% RPF in both the army and the gendarmerie. The 60:40 was a substantial concession to the FAR's larger numerical strength.
Command position ratios: 50:50 RPF / FAR rotation across senior command positions, with the most senior posts (Chief of Staff of the Army, Chief of Staff of the Gendarmerie, sectoral commands) subject to rotation between the parties.
Specific posts: The Chief of Staff of the Army was to go to the FAR; the Chief of Staff of the Gendarmerie to the RPF; the Deputy positions to the opposite party; the regional and sectoral commands distributed at 50:50.
Implementation timeline: Within 9 months of the BBTG installation, the integration was to be effected. Demobilisation of excess FAR personnel was to be supported by external donors.

For the FAR officer corps — and particularly for the akazu-aligned senior officers around Bagosora — the command-integration provisions represented the loss of the military instrument of regime survival. A 40% RPF share of force with 50% of command meant, in operational practice, that the RPF would control at least half of the post-integration army's effective combat capability — including the most disciplined units, the intelligence apparatus, and the senior staff functions where institutional power resides — while the FAR's much larger pre-integration force would be substantially demobilised. The structural-political consequence was that the génocidaire faction read the protocol as terminal for the regime, and across August 1993 – April 1994 organised, in parallel to the formal Arusha implementation process, the alternative track of comprehensive Tutsi extermination as a strategy of regime preservation.

The 4 August 1993 signature of the Arusha Peace Agreement at the Arusha International Conference Centre — the assembled Protocols of 18 August 1992, 30 October 1992, 9 January 1993, 9 June 1993, and 3 August 1993, together with the N'Sele Ceasefire (as updated) and a framework agreement — was attended by Habyarimana for the regime, Colonel Alexis Kanyarengwe for the RPF, the regional heads of state (Tanzania's Ali Hassan Mwinyi, Uganda's Yoweri Museveni, Burundi's Melchior Ndadaye, Zaire's representative for Mobutu, and Kenya's Daniel arap Moi), the OAU Secretary-General Salim Ahmed Salim, and the observer-state representatives. The signature was the formal terminus of the four-phase war that had begun on 1 October 1990. The next eight months would be the war's most consequential phase — not in the military but in the political and propaganda dimensions, as the regime's hardline faction prepared for what would follow.

10. The Long Wait: August 1993 – April 1994, UNAMIR, the Ndadaye Shock, and the Genocide Fax

The post-signature implementation phase was the diplomatic and operational architecture of the war's catastrophic terminus. The transitional institutions provided for under the Arusha protocols — the Broad-Based Transitional Government (BBTG), the Transitional National Assembly (TNA), the integrated armed forces — were to be installed within thirty-seven days of signature, i.e., by approximately 10 September 1993. They were not. The procedural and obstructions across the next eight months are documented in detail in RW-A-04; the present section treats the elements directly relevant to the military narrative.

UNAMIR — the United Nations Assistance Mission for Rwanda, established by Security Council Resolution 872 of 5 October 1993 under Force Commander Lieutenant-General Roméo Dallaire (Canadian Forces) with Chief of Staff Major Brent Beardsley, with an authorised strength of 2,548 military personnel and a Chapter VI peacekeeping mandate — was the principal external mechanism for guaranteeing the Arusha settlement. The mandate, deliberately constrained in the Security Council negotiation by the United States (which under the post-Somalia Presidential Decision Directive PDD-25 framework was attempting to limit peacekeeping commitments) and the United Kingdom, was confined to ceasefire-monitoring, security-zone supervision in Kigali, and "contribution to the security" of the transitional institutions. UNAMIR lacked the Chapter VII authority and the rules of engagement to seize weapons caches, to interdict the militias whose preparations Dallaire would document, or to use force except in self-defence and the defence of the mandate. The principal troop-contributing nations were Belgium (the 400-strong KIBAT battalion, deployed in November–December 1993), Bangladesh, Ghana, Tunisia, and (in smaller numbers) Canada, Senegal, the Netherlands, and others. The French Operation Noroît, in parallel, was withdrawn under the Arusha provisions in December 1993 — a withdrawal that Bagosora and the FAR command experienced as the loss of the external security guarantee that had sustained the regime through the worst of the war.

The 600-strong RPF battalion deployment to the Conseil National de Développement (CND) building in central Kigali, conducted under UNAMIR escort between 28 December 1993 and early January 1994, was the RPF's military-political insertion into the capital. The CND, a complex of government buildings in the Kimihurura district roughly one kilometre from Habyarimana's official residence and the Garde Présidentielle barracks, was selected as the RPF's transitional-period accommodation under the Arusha provisions requiring the RPF political and military leadership to take its place in the BBTG. The deployment was logistically extraordinary: 600 RPF soldiers under Colonel Charles Kayonga moved by road from the Mulindi northern zone through FAR-controlled territory to Kigali under UNAMIR Belgian-and-Tunisian escort, with sealed weapons stockpiles transported separately and stored under joint RPF-UNAMIR custody at the CND. The political symbolism was sharp: an armed RPF battalion in the heart of the capital, within mortar range of the presidency, was the visible institutional proof that the war had ended in negotiated outcome rather than regime preservation.

The 11 January 1994 cable from Dallaire to UN DPKO — the so-called "genocide fax" — was transmitted from Kigali in the immediate aftermath of an informant approach. The informant, a senior Interahamwe trainer codenamed "Jean-Pierre" (subsequently identified in some accounts as Jean-Pierre Abubakar Turatsinze [TBD-VERIFY: precise identification of Jean-Pierre, with some accounts identifying him as Turatsinze and others as a separate Interahamwe figure]), had approached UNAMIR through an intermediary with information that: (1) the Interahamwe was maintaining multiple arms caches across Kigali, including at the MRND headquarters; (2) the militia was being trained to kill Tutsi at a rate of 1,000 per 20 minutes; (3) lists of Tutsi for extermination were being prepared; and (4) the militia was planning to provoke Belgian peacekeeper casualties as the mechanism for triggering Belgian-and-UN withdrawal. The informant requested UNAMIR protection for himself and his family in exchange for the location of the principal arms caches; Dallaire requested DPKO authorisation to seize the caches under UNAMIR's mandate. The DPKO reply, drafted by Iqbal Riza for Under-Secretary-General Kofi Annan, instructed Dallaire not to seize the caches on the ground that such action would exceed the Chapter VI mandate, and to share the informant's information with President Habyarimana and with the Belgian, French, and US ambassadors. The Carlsson Report (1999) treats the exchange as the documentary core of the Secretariat's failure; subsequent commentary by Annan (1998 reflective remarks at Kigali memorial) acknowledged the institutional error.

The 21 October 1993 assassination of Melchior Ndadaye, the first elected Hutu president of Burundi, was the war's most consequential exogenous shock. Ndadaye, in office only since 10 July 1993 — fifteen days before the Arusha signature — was killed in a failed coup by elements of the Tutsi-dominated Burundian army; the subsequent civil-war violence across Burundi killed tens of thousands across the following months and produced a Burundian-Hutu refugee outflow into southern Rwanda of approximately [TBD-VERIFY: 300,000–600,000 across late 1993 and early 1994]. The political effect inside Rwanda was the immediate hardening of Hutu Power sentiment. RTLM broadcast the Ndadaye assassination across 22–25 October 1993 as definitive proof that political power-sharing with Tutsi was fatal for Hutu; the moderate wings of the MDR, PL, and PSD were progressively isolated from the Hutu Power wings (the Donat Murego and Frodouald Karamira wing of the MDR splitting from Faustin Twagiramungu's moderate wing; Justin Mugenzi's Hutu Power wing of the PL splitting from Landoald Ndasingwa's moderate wing). By 5 April 1994 — the day of the Dar es Salaam regional summit at which Habyarimana committed under regional-presidential pressure to install the BBTG within days — the political-procedural condition of implementation had become structurally impossible.

11. 6 April 1994: The Shoot-Down, the CND Battalion Breakout, and the RPF Decision to Resume the War

On the evening of 6 April 1994 at approximately 20:25 local time, the presidential Falcon 50 aircraft (registration 9XR-NN) carrying President Juvénal Habyarimana, President Cyprien Ntaryamira of Burundi, FAR Chief of Staff Major-General Déogratias Nsabimana, and members of Habyarimana's military cabinet was struck by two surface-to-air missiles on its approach to Kanombe airport in Kigali and crashed into the grounds of the presidential residence. All twelve people aboard were killed. The detailed reconstruction of the shoot-down — the firing point, the missile system, the operational authorship, the three contested accounts (Hutu-extremist coup thesis, RPF-responsibility thesis, Trévidic ballistic finding) — is documented in RW-B-01 and in RW-H-PRES-01; the present section treats the war's resumption.

The genocide began that night. By 22:00 on 6 April 1994 roadblocks had been established across Kigali; by midnight the Garde Présidentielle had begun systematic killings of moderate opposition politicians (the Prime Minister Agathe Uwilingiyimana killed at her residence on the morning of 7 April; the President of the Constitutional Court Joseph Kavaruganda killed in the early hours of 7 April; the PSD founder Félicien Gatabazi killed before the shoot-down on 21 February 1994 and PL founder Landoald Ndasingwa killed on 7 April); by noon on 7 April the ten Belgian peacekeepers of UNAMIR's KIBAT battalion had been killed at Camp Kigali after attempting to escort Uwilingiyimana to the Radio Rwanda studios. The strategy whose preparation the Interahamwe informant had reported in January was operationally underway.

The RPF battalion at the CND was, on the evening of 6 April, the only RPF military force in Kigali and the only force capable of mounting an immediate response. The battalion's commander Colonel Charles Kayonga broke out of the CND complex on the night of 6–7 April under heavy Garde Présidentielle and gendarmerie fire, established a defensible perimeter in the northern districts (Kimihurura, Remera, Kacyiru), and began operations to extract Tutsi and moderate Hutu civilians sheltering at the Amahoro Stadium (which UNAMIR's Bangladeshi contingent had begun protecting from 7 April), at the Mille Collines hotel under Belgian-Tutsi management, at the Sainte Famille church, and at other refugia across central and northern Kigali. The CND battalion's operations across 7–20 April 1994 rescued approximately [TBD-VERIFY: figure variously reported, with mid-range estimates of several thousand civilians extracted from central Kigali across the first three weeks of the genocide].

At Mulindi, Kagame received confirmation of the shoot-down on the night of 6 April through RPF communications and through Dallaire's UNAMIR liaison. The political-military decision to resume the war was taken within hours. Kagame's reasoning, articulated to his senior commanders and subsequently in interviews with Kinzer (2008), was that the Arusha process was dead, that the killings underway in Kigali demonstrated the regime's intent to conduct comprehensive Tutsi extermination, and that the only operational mechanism to stop the killings was the military defeat of the regime. The RPF would not negotiate a ceasefire; it would resume the war and prosecute it to military victory. The political-strategic objective, defined on 6–7 April 1994, was the capture of Kigali and the destruction of the génocidaire regime.

The RPF's order of battle at the resumption was approximately 20,000 fighters organised in eight to ten brigades, concentrated in the northern Mulindi-Byumba zone and at the CND. The principal operational axes were three: a northern axis from Mulindi southward through Byumba; an eastern axis from the Mutara region southward through Kibungo; and the CND foothold in central Kigali. The advance began on 7 April 1994.

12. The 100-Day Military Campaign: Byumba, Gitarama, Kigali, Butare, Gisenyi (April – July 1994)

The military campaign across 7 April – 4 July 1994 was the only operationally successful military halting of an ongoing genocide in twentieth-century history. The operational geography ran south from the northern bases through the central highlands to Kigali, then southwestward and westward to Butare and the Zairean border. The campaign was conducted under conditions of comprehensive international non-intervention. Approximately 800,000 to 1,000,000 Tutsi and moderate Hutu civilians were killed across the 100 days by Interahamwe, Impuzamugambi, Garde Présidentielle, FAR, and civilian-perpetrator forces under regime direction; the RPF's military advance was the only operational mechanism by which the killings in any given district were halted.

The campaign proceeded in four phases. The first (7–20 April 1994) was the RPF's break-out from its northern bases and consolidation of the principal advance axes: the CND battalion broke out and established northern-Kigali positions, the main northern force advanced south from Mulindi against fragmenting FAR resistance, and Kibungo and Byumba prefectures were cleared by mid-April. The second (21 April – mid-May 1994) took the bulk of the eastern prefectures (Kibungo, Umutara, eastern Kigali-rural) and coincided with the UN's catastrophic mandate-reduction decision: Security Council Resolution 912 of 21 April 1994, adopted under the political momentum of the Belgian KIBAT withdrawal that followed the 7 April killing of the ten Belgian peacekeepers, reduced UNAMIR's authorised strength from 2,548 to 270 — operationally, the formal abandonment of the peacekeeping mission. Resolution 918 of 17 May 1994, attempting to reverse the abandonment, authorised a 5,500-strong UNAMIR II, but the troop-contribution and equipment-deployment process was so slow that it was not effectively in theatre until after the RPF's military victory; the RPF advanced through this phase under conditions of complete external non-intervention.

The third phase (mid-May – mid-June 1994) was the prolonged, district-by-district battle for Kigali, fought over approximately six weeks between the RPF advancing from the north and east and the Garde Présidentielle, FAR Para-Commando Battalion, and Interahamwe defending the capital's central and southern districts. The political-military collapse of the regime came when the interim government — the Gouvernement Intérimaire under President Théodore Sindikubwabo and Prime Minister Jean Kambanda, formed on 8–9 April — fled Kigali for Gitarama in late April. The fourth phase (mid-June – 17 July 1994) was the southern and western sweep: Gitarama fell on 13 June, Kigali was effectively captured on 4 July with the génocidaire forces in comprehensive retreat, Butare — the last major centre of organised killing — fell a day earlier on 3 July, and Cyangugu and Gisenyi (the regime's final base) fell on 17–18 July, with the interim government and its remaining forces fleeing across the Zairean border into the Goma area, where they would constitute the post-1994 ex-FAR/Interahamwe refugee-cum-militia complex documented in RW-A-01 and the post-1994 documents.

Operation Turquoise — the French military deployment authorised by Security Council Resolution 929 of 22 June 1994 — ran from 23 June to 21 August 1994 with approximately 2,500 French troops in the southwestern prefectures, creating a "humanitarian protection zone" (the Zone Humanitaire Sûre, ZHS) that covered the Cyangugu, Kibuye, and southern Gikongoro districts. The Turquoise zone's evaluation has been the most-contested element of the war's international dimension. The official French position (subsequently confirmed in part by the 2021 Duclert Commission Report) was that Turquoise protected approximately 1.5 to 2 million Hutu refugees and surviving Tutsi from continued killing and from RPF reprisal. The critical position, advanced by Wallis (2007), Prunier (1995), and African Rights (1995), was that the Turquoise zone substantially facilitated the orderly withdrawal of génocidaire leadership and Interahamwe forces into Zaire — and that the French government's reluctance to arrest senior génocidaire figures in the zone, including elements of the akazu and the interim government, was a documented institutional failure. The corpus records both readings.

The RPF's own conduct during the 100-day campaign is part of the contested record. Allegations of RPF reprisal killings of Hutu civilians in zones captured during the advance — at locations including Byumba prefecture (April–May 1994), the Kabgayi area (the killings of approximately 13 Catholic clergy on 5 June 1994, including the Archbishop of Kigali Vincent Nsengiyumva and the Bishop of Kabgayi Thaddée Nsengiyumva and the Bishop of Byumba Joseph Ruzindana [TBD-VERIFY: precise count and identity of the clergy killed at Kabgayi]), and other sites documented by Des Forges (1999), African Rights (1995), and subsequently by the 2010 UN Mapping Report for the post-1994 period — are part of the historiography. The RPF's official position has consistently been that any such killings were the unauthorised acts of individual fighters or units that operated outside command discipline and were prosecuted internally where identified. The critical position holds that systematic reprisal killings were either tolerated or actively organised at command level. The corpus records the contestation without adjudication and notes that the question is a live element of the post-1994 transitional-justice debate documented in RW-J-01 (when written).

13. Three Accounts of the War (RPF Official / Hutu Power Residual / International Observer)

The corpus records three accounts of the 1990–1994 war without adjudicating between them, in the discipline applied throughout the Rwanda documents to genuinely contested questions.

The RPF-official account — articulated in Rusagara (2009), in the speeches and writings of Kagame and the RPF political leadership across the post-1994 decades, and in the institutional historiography of the post-genocide Rwandan state — treats the war as the just armed return of a thirty-one-year stateless Tutsi diaspora that the 1962 PARMEHUTU foreclosure of the right of return had rendered politically necessary. The invasion of 1 October 1990 is read as the political-military instrument of return-as-right; the four-phase war as a disciplined campaign that progressively forced the regime to a negotiated settlement; the Arusha Accords as a serious good-faith negotiated outcome that the akazu faction subsequently sabotaged through a deliberate strategy of obstruction and parallel preparation of the genocide; the shoot-down as the operational trigger of the prepared killing; and the 100-day campaign as the only operationally successful military halting of an ongoing genocide in twentieth-century history. Allegations of RPF civilian-targeting during the war and during the 100-day campaign are treated as either fabrications, as the unauthorised acts of individuals subsequently prosecuted, or as the contested allegations of a génocidaire-diaspora propaganda apparatus that has continued to operate from Zairean and European exile.

The Hutu Power residual account — surviving in elements of the post-1994 génocidaire-diaspora propaganda apparatus reorganised in eastern Zaire from 1994, in some Francophone polemical literature including the writings of regime apologists in Paris and Brussels, in elements of the November 2006 Bruguière indictments on the shoot-down, and in some elements of the contemporary FDLR ideological framework — treats the war as an unprovoked Ugandan-backed Tutsi invasion that triggered the Habyarimana regime's defensive radicalisation. The 1 October 1990 invasion is read as an act of Ugandan state aggression conducted by Rwandan-refugee proxies under Museveni's strategic direction; the regime's ibyitso arrests and the subsequent anti-Tutsi massacres are framed as security responses to a real and existential military threat; the Arusha negotiations are read as a coerced settlement imposed on the regime by external pressure under threat of further RPF military advance; the shoot-down is attributed to the RPF as a deliberate act to remove Habyarimana as the political obstacle to RPF military victory; and the 1994 genocide is framed either as a spontaneous popular reaction to the assassination of Habyarimana or as a regrettable but understandable collective self-defence against an existential Tutsi threat. The corpus records this account as a historiographical fact of the field while noting that its empirical claims — particularly on the shoot-down authorship, the systematic character of the 1990–1992 reprisal massacres, and the documented preparation of the genocide before the shoot-down — are extensively contradicted by the documentary record assembled by Africa Rights, Des Forges, Melvern, Guichaoua, the ICTR judgements, and the Duclert Commission Report.

The international-observer account — Adelman and Suhrke (1999), Dallaire (2003), Jones (2001), the Carlsson Report (1999), Wallis (2007), and the broader scholarly literature — reads the war as a tractable conflict whose Arusha-process mediation was structurally under-resourced. The principal structural weaknesses identified across this literature are: (1) the CDR exclusion as spoiler — the protocol decision to exclude the Coalition pour la Défense de la République from the BBTG and TNA, on the ground of the CDR's racially exclusivist programme, left a significant Hutu Power constituency outside the transitional institutions with no stake in their preservation; (2) the inadequate UNAMIR mandate — the Chapter VI peacekeeping mandate under the post-Somalia post-PDD-25 constraints lacked the Chapter VII authority and rules of engagement that the situation required; (3) the Belgian-and-American refusal to expand the mandate after the January 1994 warnings, and the subsequent Resolution 912 reduction of the force in the genocide's first days; (4) the post-Somalia post-PDD-25 constraints on US engagement, which structurally foreclosed the most consequential potential intervention; and (5) the French military-political relationship with the Habyarimana regime, which the Duclert Commission characterised as involving "heavy and overwhelming responsibilities" while declining to find established complicity in the genocide itself. The international-observer account reads the war's collapse as a documented diplomatic failure as much as a sabotage by hardliners, and its military terminus as determined by the operational quality of the RPF and the political bankruptcy of the génocidaire regime by mid-June 1994 — not by any external intervention. The corpus follows this account in declining to adjudicate between the first two while documenting all three.

14. Forward View: From the 4 July 1994 Capture of Kigali to the Mapping Report and the M23 Question

The 4 July 1994 capture of Kigali and the 17–18 July fall of Gisenyi terminated the war but did not terminate the war's institutional consequences. Three forward arcs are recorded here as the bridges to subsequent corpus documents.

The RPA-to-RDF institutional transformation is documented in RW-I-01. The wartime RPA — approximately 25,000 fighters at the moment of victory — became the institutional core of the post-1994 Rwandan armed forces, renamed the Rwandan Defence Force (RDF) in 2002. The "Virunga generation" of officers who had reconstituted the force in the winter of 1990–1991 became the senior RDF cohort across the next three decades. Kagame retained personal command authority over the force through the Vice Presidency and Defence Ministry from 1994 to 2000 and through the Presidency from 2000 onward. The doctrinal discipline established during the war — disciplined small-unit operations, civilian-conduct codes, intelligence-led targeting, refusal of forced recruitment — became the institutional culture of the RDF and was the operational foundation of the 1996–1997 First Congo War and the 1998–2003 Second Congo War.

The Arusha framework's partial survival in the post-1994 transitional arrangements is documented in RW-C-05 (Government of National Unity, 1994–2003). The post-genocide transitional government, constructed under a modified Arusha framework, retained the BBTG portfolio allocations as a starting point — but with the CDR comprehensively excluded as a génocidaire organisation, the MRND dissolved, and the moderate wings of the MDR, PL, and PSD recognised as the legitimate continuators of those parties. The Hutu-Tutsi political balance was preserved formally (with Pasteur Bizimungu as Hutu President and Faustin Twagiramungu as Hutu Prime Minister) while the RPF retained operational control through Kagame's Vice Presidency and Defence Ministry. The Arusha refugee protocol's principle of "right of return without restitutionary maximalism" was inherited and substantially expanded under the conditions of the 1996–1997 mass return from Goma and Tanzania documented in RW-C-07.

The 1996–1997 First Congo War, the 2010 UN Mapping Report, and the contemporary M23 question are the long-arc consequences of an army that did not demobilise after the 1994 victory. The 1996–1997 AFDL campaign that toppled Mobutu Sese Seko was conducted substantially by the RPA in alliance with Ugandan and Angolan forces; the 1998–2003 Second Congo War followed Kabila's break with his Rwandan backers; the 2012–2013 M23 episode and the 2021–2024 M23 resurgence have been the subjects of continuous UN Group of Experts documentation of Rwandan support (denied by Rwanda). The 2010 UN Mapping Report (documented in RW-J-02 when written) catalogued the alleged killings of Hutu refugees and Congolese civilians during the 1996–1997 campaign as potentially genocidal in character. The DRC engagement is therefore the principal contemporary consequence of the 1990–1994 war: the army that ended the genocide became the army that has been continuously engaged in the eastern DRC for thirty years.

15. Conclusion and Spiral Index

The 1990–1994 Rwandan civil war was, in its constitutive features, three wars at once. It was the military campaign of a stateless diaspora seeking armed return after thirty-one years of foreclosure — a campaign that began as a failed conventional invasion at Kagitumba on 1 October 1990, was reconstituted as a disciplined guerrilla insurgency through the Virunga winter of 1990–1991, conducted a sustained operational campaign from the Mulindi base across 1991–1992, demonstrated decisive military superiority in the 8 February 1993 offensive, and culminated in the 100-day campaign that captured Kigali on 4 July 1994 and halted the only operationally successful military halting of an ongoing genocide in twentieth-century history. It was the diplomatic process of the Arusha negotiations across 1992–1993 — the four-phase mediation under Tanzanian and OAU leadership that produced the most institutionally elaborate African civil-war peace settlement of its decade and that collapsed eight months and four days after signature in the most catastrophic failure of post-Cold-War international peacekeeping. And it was the propaganda-and-mobilisation campaign of the akazu faction inside Rwanda — the systematic radicalisation of anti-Tutsi politics through the ibyitso arrests of October 1990, the Kibilira, Bagogwe, and Bugesera reprisal massacres, the December 1990 Hutu Ten Commandments, the July 1993 launch of RTLM, and the parallel preparation of the genocidal architecture that, when the shoot-down of 6 April 1994 provided the operational trigger, produced the 100-day catastrophe documented in RW-B-01.

The corpus reads the war's structural lesson as the foreclosure-of-return-produces-armed-return arithmetic that Mamdani (2001) made canonical: a refugee population denied citizenship in either the host countries or the homeland is, over a sufficient time horizon, a politics of armed return. The 1959 Hutu Revolution's choice to define Rwandan political belonging in ethnic-majoritarian terms produced, three decades later, the armed return of the population that had been ethnically excluded. The Habyarimana regime's response to the armed return — the ibyitso radicalisation, the parallel preparation of genocide as an alternative to negotiated settlement, the catastrophic political wager that comprehensive Tutsi extermination could preserve regime continuity — was the political-strategic failure that produced the genocide. The RPF's military victory was simultaneously the military halting of the genocide and the institutional foundation of the post-1994 Rwandan state whose own contested governance trajectory is documented across the C, D, E, F, G, and J blocks of the Rwanda corpus.

The forward narrative resumes in RW-B-01 (the 1994 genocide as the political-military terminus of the eight-month Arusha non-implementation phase that this document narrates as the war's diplomatic culmination), in RW-A-04 (the content of the Arusha protocols and the post-signature implementation failure), in RW-H-PRES-03 (Kagame's biographical trajectory through the war and into the post-1994 state), and in RW-I-01 (the RPA-to-RDF institutional history and the post-1994 military-political architecture). The war's longest-arc consequence — the continuous Rwandan engagement in the eastern Democratic Republic of the Congo from 1996 to the present — is documented in the F-block foreign-policy documents and in the J-02 Mapping Report document. The war is the institutional pivot of the Rwandan twentieth century: every feature of the Habyarimana regime's collapse, of the 1994 genocide, of the post-1994 reconstruction, and of the contemporary Rwandan developmental-authoritarian state runs through it.

Sources

Prunier, Gérard. The Rwanda Crisis: History of a Genocide (London: Hurst, 1995; 2nd edn. 1997). The principal Anglophone scholarly narrative of the 1990–1994 period; chapters on the invasion, the war, Arusha, and the genocide trigger.
Mamdani, Mahmood. When Victims Become Killers: Colonialism, Nativism, and the Genocide in Rwanda (Princeton University Press, 2001). The analytical reading of the war as the armed return-from-exile of a stateless diaspora and the regime's response as nativist mobilisation.
Des Forges, Alison. Leave None to Tell the Story: Genocide in Rwanda (Human Rights Watch / FIDH, 1999). The 800-page documentary report; the principal source on the 1990–1994 massacres of Tutsi civilians inside Rwanda.
Melvern, Linda. A People Betrayed: The Role of the West in Rwanda's Genocide (London: Zed Books, 2000; rev. edn. 2009). The diplomatic-and-Western-policy reconstruction; substantial Arusha and UNAMIR coverage.
Melvern, Linda. Conspiracy to Murder: The Rwandan Genocide (London: Verso, 2004; rev. edn. 2006). The forensic reconstruction of the genocide's preparation under the cover of the war.
Reyntjens, Filip. L'Afrique des Grands Lacs en crise: Rwanda, Burundi: 1988–1994 (Paris: Karthala, 1994). The contemporary Francophone monograph; detailed civil-war chronology.
Reyntjens, Filip. Rwanda, Trois jours qui ont fait basculer l'histoire (Paris: L'Harmattan, 1995). The focused reconstruction of the 6–9 April 1994 transition.
Dallaire, Roméo (with Brent Beardsley). Shake Hands with the Devil: The Failure of Humanity in Rwanda (Toronto: Random House Canada, 2003). The UNAMIR Force Commander's first-hand record of the deployment, the 11 January 1994 cable, and the 100-day campaign as witnessed.
Adelman, Howard, and Astri Suhrke (eds.). The Path of a Genocide: The Rwanda Crisis from Uganda to Zaire (New Brunswick, NJ: Transaction, 1999). The principal multi-author scholarly assessment; chapters by Bruce Jones, Adelman, Suhrke, and others on the war's diplomatic management.
Jones, Bruce D. Peacemaking in Rwanda: The Dynamics of Failure (Boulder: Lynne Rienner, 2001). The detailed Arusha-process monograph; the source for the four-phase periodisation and the implementation-failure analysis.
Wallis, Andrew. Silent Accomplice: The Untold Story of France's Role in the Rwandan Genocide (London: I.B. Tauris, 2007). The principal Anglophone investigation of Operation Noroît and the French military-political relationship with the Habyarimana regime.
Carlsson, Ingvar, Han Sung-Joo, and Rufus Kupolati. Report of the Independent Inquiry into the Actions of the United Nations during the 1994 Genocide in Rwanda (UN Security Council document S/1999/1257, 15 December 1999). The Carlsson Report; the UN's own institutional self-assessment.
Africa Rights. Rwanda: Death, Despair and Defiance (London: African Rights, 1995). The contemporaneous documentation of civilian massacres 1990–1994 and the genocide; the most extensive single survivor-testimony archive of the period.
International Commission of Investigation on Human Rights Violations in Rwanda (Africa Watch / FIDH / ICHRDD / UIDH). Final Report (March 1993). The field-investigation report documenting the Kibilira, Bagogwe, and Bugesera massacres; the first internationally credible documentation that the regime was organising ethnic killing.
Ndiaye, Bacre Waly. Report by the Special Rapporteur on Extrajudicial, Summary or Arbitrary Executions on his Mission to Rwanda from 8 to 17 April 1993 (UN document E/CN.4/1994/7/Add.1, 11 August 1993). The first UN report to invoke the word genocide in characterising the Rwandan trajectory.
Kinzer, Stephen. A Thousand Hills: Rwanda's Rebirth and the Man Who Dreamed It (Hoboken: Wiley, 2008). The sympathetic Kagame biography; the principal Anglophone source on Kagame's Mulindi-period reorganisation of the RPA and his command across 1990–1994.
Rusagara, Frank K. Resilience of a Nation: A History of the Military in Rwanda (Kigali: Fountain Publishers, 2009). The RDF institutional history; the official RPA account of the 1990–1994 campaign.
Guichaoua, André. From War to Genocide: Criminal Politics in Rwanda 1990–1994 (Madison: University of Wisconsin Press, 2015). The principal Francophone forensic reconstruction; substantial Arusha non-implementation material and ICTR-derived archival work.
UN Security Council. Resolution 872 (5 October 1993), Resolution 912 (21 April 1994), Resolution 918 (17 May 1994), Resolution 925 (8 June 1994), Resolution 929 (22 June 1994). The UNAMIR mandate sequence.
Otunnu, Ogenga. "An Historical Analysis of the Invasion by the Rwanda Patriotic Army (RPA)" in Adelman and Suhrke (eds.) (1999). The principal Ugandan-perspective scholarly account of the refugee-diaspora origins of the invasion.
Mamdani, Mahmood, and Joel Rocamora. Conversations with senior RPF cadres archived in the Mamdani fieldwork supporting When Victims Become Killers (2001). [TBD-VERIFY: specific archival placement.]
International Criminal Tribunal for Rwanda (ICTR). Judgements in Prosecutor v. Bagosora et al. (Case ICTR-98-41, Trial Chamber I, 18 December 2008), Prosecutor v. Bizimungu et al. ("Government II" / Military II, Case ICTR-99-50), and Prosecutor v. Nyiramasuhuko et al. ("Butare", Case ICTR-98-42). The judicial reconstruction of the regime's planning across 1990–1994.

RW-A-01: 1959 Hutu Revolution and First Republic (the refugee-diaspora origin context)
RW-A-02: Habyarimana Era 1973–1994 (the regime that the invasion confronted)
RW-A-03: Hutu Revolution, First Republic, and the Roots of Genocide (the deep-historical anchor; the 1959–1964 exodus from which the RPF emerged)
RW-A-04: Arusha Accords 1992–1994 (the negotiated track that this military campaign produced and constrained)
RW-B-01: The 1994 Genocide Against the Tutsi (the 100-day catastrophe whose military terminus this document narrates)
RW-D-01: 2003 Constitution and First Direct-Presidential Election (the post-1994 institutional sequel)
RW-G-01: Mutuelle de Santé (the post-1994 reconstruction policy domain referenced in conclusion)
RW-H-PRES-01: Juvénal Habyarimana Biography (the principal counter-actor of the war)
RW-H-PRES-02: Pasteur Bizimungu Biography (the Hutu RPF political face during and after the war)
RW-H-PRES-03: Paul Kagame Biography (the RPF commander and the principal architect of the war's military prosecution)
RW-R-01: Governance Books Canon (the source-canon document)
RW-D-06: back-reference added by symmetry sweep
RW-F-01: back-reference added by symmetry sweep
RW-F-02: Rwanda-Uganda Relations 1990-2026 — From the Uganda-Trained RPF Origins to the 2019-2022 Border Closure and the Doha-Era Reset
RW-G-03: Umuganda — The Rwandan Community Service Programme
RW-D-02: The RPF Party Architecture — Inkotanyi Origins, the Cell-Sector-District Cadre System, and the Developmental-State Model
RW-D-03: Rwandan Decentralisation and the Imihigo Performance-Contract System — Districts, Sectors, Cells, Imidugudu and the Subnational Delivery Architecture
RW-G-04: Girinka — The One-Cow-Per-Family Programme
RW-D-04: Rwandan Local Elections and the Mayor-Pyramid Architecture — District Mayors, Sector Executives, Cell and Village Selection, and the Imihigo Accountability Cycle
RW-F-03: Rwanda's Cabo Delgado Deployment in Mozambique (July 2021 – 2026) — The RDF Intervention Against the ISIS-Mozambique Insurgency, the EU Funding Architecture, and the Africa Peacekeeping-Export Model
RW-D-05: Rwandan National Electoral Architecture and the RPF Dominance — Presidential and Parliamentary Cycles, the Candidate-Registration Practice, the 30 Per Cent Women's Quota, and the Contested-Record on Electoral Integrity
RW-E-03: The Rwandan Mining Sector and the DRC Conflict-Mineral Question — Tantalum, Tin, Tungsten, and Gold from the iTSCi Era Through the EU Critical Raw Materials Pact and the 2025 Sanctions Reset
RW-E-04: Rwandan Agribusiness, Specialty Coffee, and the Tea Sector Strategy — NAEB Architecture, the Specialty Shift, Tea Privatisation, Horticulture Diversification, and the 2024–2026 Export Targets
RW-J-01: The Failure to Prevent — Three Accounts of Why the 1994 Genocide Was Not Stopped, and the Contested International Response
RW-I-01: The Rwanda Defence Force — The Army-State Relationship from Rebel Victory to Security Exporter