In research, there are documents that shout, and there are others that whisper. The 1994 Genocide against the Tutsi in Rwanda was preceded by both. Initially, Kangura shouted. Radio Télévision Libre des Mille Collines (RTLM) screamed. But long before the shouting reached its crescendo, the whispering had already become policy.
ALSO READ: Kangura's 8th Hutu commandment re-examined
One such soft voice is an intelligence report dated November 21, 1991, written in Kigali, addressed to the G2 (Military Intelligence) of the General Staff of the then Rwandan Armed Forces, and signed by Commander Bujyakera J., an intelligence officer of the État-Major des Forces Armées Rwandaises (EM AR). It is the product of a mission conducted by Commander Bujyakera himself and Second Lieutenant Gasimba, pursuant to orders referenced as Letter No. 038/G3.3.1.0, of August 31, 1991, from the Head of Operations (G3) of the General Staff of the National Gendarmerie.
ALSO READ: How Kangura inspired Ekenge and FARDC
These details matter, because they establish intent, hierarchy, and institutional continuity. This was not a drunken rant, not a pamphlet smuggled in a marketplace. It is not even a rogue military officer’s fantasy. This was predictable intelligence work, filed, referenced, and archived.
And it opens under the shadow of an ideology that had already been normalized.
From Kangura to the army
The intellectual and moral frame of the report is captured in a sentence that appears at its head: "The Rwandan Armed Forces must be exclusively Hutu. The experience of the October 1990 war teaches us this. No soldier must marry a Tutsi woman.”
This decree-like phrase does not originate in the report itself. It comes from Kangura No. 6, published in December 1990, under the infamous "Ten Hutu Commandments.” It is Commandment No. 7.
That background is crucial. Kangura was not a fringe newsletter. It was published by Hassan Ngeze but as an operative of top Hutu-Power ideologues. Kangura circulated widely, read in government offices, cited by officials, and later entered into evidence at the ICTR as a direct instrument of genocidal incitement. The 10 Commandments were not descriptions; they were instructions.
By November 1991, Commandment No. 7 had migrated impeccably from propaganda to practice. The report does not debate it. It operationalizes it.
The army under Juvenal Habyarimana, according to this logic, was no longer a national institution. It was an ethnic sanctuary. Marriage was no longer private. It is a security breach. A woman’s ethnicity became a vector of infiltration. And, that is how ideology becomes administration.
ALSO READ: Habyarimana's role in planning, implementing 1994 Genocide against the Tutsi
The stated purpose of the mission could not be clearer: "to investigate the ethnic affiliation of the officers of the National Gendarmerie cited in the reference letter, who, according to rumors, were considered by the troops to be of the ethnic group TUTSI and would therefore be possible accomplices of the INYENZI (IBYITSO).”
Several things happen in this single phrase. First, rumor is elevated to the status of intelligence. Second, Tutsi identity is treated as prima facie evidence of treason and equated with complicity. Third, the debasing slur "Inyenzi”—cockroach, the term used by Hutu extremists to dehumanize the Tutsi—is incorporated into an official intelligence document without quotation marks, admission of guilt, or distance.
The logic is circular and airtight: if you are Tutsi, you are suspect; if you are suspect, you must be investigated; if you are investigated, evidence will be found. The investigation technique is revealing. These officers did not examine conduct, loyalty, or actions.
In genocide studies, this is known as pre-emptive criminalization: a group is not punished for what it has done, but for what it is imagined to be capable of doing.
The individuals under investigation are named: Commander, Gendarmerie Murangira; Captain, Gendarmerie Ndereyimana; Lieutenant, Gendarmerie Rutayisire; Lieutenant, Gendarmerie Karerangabo; Second Lieutenant, Gendarmerie Mugabo; Second Lieutenant, Gendarmerie Ndagijimana and Second Lieutenant, Gendarmerie Mpakaniye. They are not accused of any insubordination. Not of collaboration. Not leaking sensitive information. They are accused of existing incorrectly and suspiciously.
Investigating blood
The investigative methodology reveals how deeply the state had embraced ethnic policing: "we went to their home hills ... conducted interviews with the local population, particularly elderly people ... consulted communal census records ... parents’ files ... persons sharing family ties (brothers, uncles) ... and observations by the communal authorities (Bourgmestres, communal councillors).”
According to the report, this is not counter-intelligence. It is genealogical surveillance. The hills were not visited to examine conduct, but to interrogate ancestry. Elderly people are not consulted for memory, but for classification. Census records—already a colonial tool of racialization—are resurrected as forensic evidence. This is purely a racist system that transforms political identity into biological destiny. This report shows that transformation was fully underway.
One of the most grotesque aspects of the report is its obsession with paperwork, particularly census forms. On Lieutenant Rutayisire, the investigators write: "on the said census record, the mentions ‘HUTU, TUTSI, TWA’ were all crossed out: ‘HUTU and TWA’ with one pen, ‘TUTSI’ with another pen.”
This is presented as forensic proof. The different pens tell a story. One pen is innocent; the other is criminal. This detail is not incidental. It is treated as probative.
The pen becomes a witness. Ink becomes intent. The investigators conclude: "This leads us to believe that the TUTSI designation was crossed out at the time his identity card was falsified.” The act of crossing out "TUTSI” becomes an attempted escape, a bureaucratic jailbreak.
What is being criminalized here is not fraud in the abstract, but the attempt to escape ethnic persecution. In Rwanda’s earlier pogroms—1959, 1963–64, 1973—many Tutsi altered or concealed their identity to survive. This report retroactively transforms survival into subversion. Individual’s identity becomes like smuggled goods. The report redefined extremism. Commander Murangira’s case offers a masterclass in ideological inversion. His father is racially and ideologically acceptable: "His father is a well-known HUTU in his region.”
But his mother is not: "the mother of Cdt MURANGIRA is TUTSI, from a more or less well-off and extremist family.” What makes the family extremist? "...this family brought a case before the courts around 1964 against people who considered them HUTU.”
In other words, defending one’s identity before a court of law is retroactively classified as extremism. Poverty is innocence; wealth is suspicion. Silence is loyalty; self-assertion is danger. Let us pause. In the aftermath of the 1959 pogroms and the 1963–64 massacres, defending one’s identity in court was an act of courage.
Then comes the marriage: "the wife of Cdt MURANGIRA is TUTSI. All this may influence the behavior and tendencies of this officer.” Just imagine! Marriage is no longer an intimate choice. It is an ideological leak. A wife becomes evidence.
When the body betrays
If documents in the file fail to convict, the body must testify. In the case of Captain Ndereyimana, the report concedes: "All the information gathered about him indicates that he is HUTU.”
But certainty is intolerable. Mistrust and utmost suspicion must be sustained. Morphology. Bone structure had to be presented as counter-evidence. The body betrays the file. And when genealogy grows inconveniently blurry, the language turns evasive.
So, the investigators examine his mother: "his mother presents a morphology tending toward TUTSI, although she is listed as HUTU on her census record.” This is racial science without instruments. Anthropology is reduced to watching. This was colonial racial imagination repackaged by post-colonial bureaucrats. An individual’ sexual history becomes evidence as well: "after the death of NDEREYIMANA’s father, his mother had other children, all born of TUTSI fathers.”
From this, the report draws its conclusion: "his mother displays a TUTSI tendency through her partners.” Ethnicity becomes contagious and a desire. Desire itself acquires an ethnic direction. Desire becomes guilt. Guilt becomes inheritable.
To maintain racial anthropology in the army, courts became ethnic battlefields. The report says Lieutenant Karerangabo sought judicial redress. The report does not deny this. It pathologizes it. According to the investigators, Lt. Karerangabo went to court: "for fear of NOT being dismissed from the FAR for having falsified his identity card.”
Justice is framed not as a right, but as a tactic of evasion. When the court rules in his favor, the explanation is revealing: "the Prosecutor is himself TUTSI, as are his deputies.” The implication is obvious. Justice has an ethnicity. And it is suspect.
The proposed correction is distressing: "if there were a HUTU deputy prosecutor ... the court should be led to review the judgment.” This is not a call to abolish courts. It is a call to ethnically align them. This is when law is bent to serve exclusion.
The report concludes with bureaucratic neatness, attaching letters, census forms, and annexes: "I attach to this note the letter No. 112/04.05/1 from the Prefect of KIBUNGO ... as well as copies of the census records established in 1982 and in 1988.”
Everything is in order. Everything is documented. And this is perhaps the most terrifying aspect of the document: its normalcy. There is no rage here. No hysteria. No bloodlust. Only procedure.
Raul Hilberg, in his study of the Holocaust, warned that genocide is not primarily the work of fanatics, but of clerks. This report is a clerk’s masterpiece. Bujyakera’s report was not a warning sign. It was a progress report, because genocide does not begin with killing. It begins when a society agrees that some people must be permanently explained, classified, suspected, and corrected—and that this work is administrative, rational, even patriotic.
Practicing Genocide
The document does not shout. It does not incite with drums or slogans. It measures. It inspects. It crosses out boxes with different pens. It peers at cheekbones, interrogates mothers, dissects marriages, and delves through family trees like a forensic accountant of blood. And in doing so, it reveals a society already policed by genocidal tendencies, long before genocide needed to be named. This is not an opinion scribbled in a margin. It is an operational principle. The army is no longer a national institution; it is an ethnic enclosure. Marriage itself becomes a security threat. Love and marriage reclassified as infiltration.
This report was submitted nearly two and a half years before the Genocide against the Tutsi. Yet the architecture of genocide is already complete: ethnicity as destiny, courts as ethnic fronts and survival as crime.
No machetes appear here. No militias. No death orders. Only files, hills, elders, and "tendencies.” But genocide does not begin with killing. It begins when a state decides that some citizens must be permanently explained, monitored, corrected, and purified—and that this work is patriotic.
This report is not a warning sign. It is a rehearsal. And like all rehearsals, it tells us exactly what the actors intend to do when the curtain finally rises.
Suspects forever
On February 20, 1993, as Rwanda edged closer to catastrophe, Colonel Déogratias Nsabimana, the Chief of Staff of the Rwandan Armed Forces, wrote a seemingly routine letter to the Minister of National Defence. Its subject was inoffensive enough: "Access to military installations.” Its implications were anything but.
The letter opens with bureaucratic courtesy: "Mr. Minister, I have the honour to bring to your attention that the individuals named MUSHINZIMANA Théogène, KANYAMAHANGA KABANDANA, and NSENGIMANA Jean de Dieu were designated by MINIFIN to carry out inspections in our military camps.”
There is no single allegation of wrongdoing. No reference to espionage. No claim of breach of protocol. These men were civilian financial inspectors, officially mandated by the Ministry of Finance to perform audits in military camps—an entirely normal function in a state that still pretended to have civilian oversight.
Then comes the turn. "These individuals, whose identity card photocopies are attached, have become suspect and NO LONGER inspire confidence, especially in these times of war against the Inkotanyi.”
The emphasis is not accidental. "Have become suspect” begs the obvious question: on what grounds? The letter offers none. Instead, it quietly gestures to the annex: photocopies of identity cards. Those photocopies revealed one thing, and one thing only: they were Tutsi. From that fact alone flows the request: "For this reason, I request your intervention so that these three individuals may be replaced by others who are less suspect.”
"Less suspect” does not mean more qualified. It does not mean more discreet—or more trustworthy by conduct. It means of a different ethnicity. Suspicion is no longer a conclusion drawn from evidence; it is a preloaded condition, activated by identity. The letter then codifies this logic into policy: "Furthermore, for reasons of security of our installations, any person called upon to access them should first be approved by the duly authorized military authority.”
This is not merely about access control. It is about ethnic clearance. Civilian oversight is redefined as infiltration. Auditing becomes espionage. The state’s own paperwork becomes a filter to exclude citizens deemed inherently untrustworthy. What makes this letter unsettling is its serenity. There is no rage or propaganda flourish. Just administrative prose performing quiet exclusion. The war "against the Inkotanyi” becomes the all-purpose justification for transforming Tutsi identity into a security liability.
By February 1993, Rwanda was no longer asking what people did. It was asking what they were—and treating the answer as enough. This is how genocidal systems mature: not always through shouting, but through memos; not always through violence, but through denial of trust. Long before people were denied life, they were denied access, credibility, and belonging.