South America's military dictatorships ran a joint campaign of cross-border assassination and terror

Three tons of the regimes' own paperwork

The strongest evidence for Condor did not come from a journalist, a defector or a foreign intelligence service. It came from the dictatorships' own filing cabinets. On 22 December 1992, Paraguayan human-rights lawyer Martín Almada — himself a survivor of torture under the regime of Alfredo Stroessner — filed a habeas data petition seeking his own police file. Acting on a search order from judge José Agustín Fernández, investigators found roughly three tons of documents stacked in a police station in Lambaré, a suburb of Asunción. The cache became known as the Archivos del Terror, the “Archives of Terror,” and UNESCO later inscribed it on the Memory of the World Register as world documentary heritage.

The files were not rhetoric; they were bureaucracy — arrest logs, interrogation transcripts, prisoner-transfer orders, death certificates and inter-agency cables, written by the regimes for their own internal use, never intended for public eyes. Researchers working from the archive have tallied entries corresponding to some 50,000 people recorded killed, 30,000 recorded “disappeared,” and 400,000 recorded detentions across the Condor states — figures drawn from the regimes' own record-keeping, not from outside estimation.

The paper trail did not stop at Paraguay's border. Beginning in the 1990s, the U.S. National Security Archive, a nongovernmental research group, used Freedom of Information Act requests to pry loose thousands of pages of State Department and CIA cables on Condor. Among them: a CIA cable from June 1976, the first known declassified U.S. document to name “Condor” outright, describing it as a “cooperative arrangement” to build a shared computerized data bank on leftist targets; and a State Department briefing memo from August 1976 informing Secretary of State Henry Kissinger that Condor members had organized to “find and kill terrorists” both at home and in Europe. None of this is characterization by critics — it is the language of the intelligence agencies' own internal paperwork, describing their own knowledge in real time.

The evidence was not an accusation from outside the regimes. It was the regimes' own filing cabinets, and eventually Washington's own cables, describing the same operation from both sides.

Condor's reach beyond South America is itself documented in a criminal court record, not a theory. On 21 September 1976, a remote-controlled car bomb killed former Chilean ambassador Orlando Letelier and his American colleague Ronni Moffitt at Sheridan Circle in Washington, D.C., blocks from Embassy Row. U.S. federal prosecutors indicted DINA chief Manuel Contreras and other Chilean intelligence officers; American expatriate Michael Townley, who built the bomb, pleaded guilty in a U.S. court and testified that he acted on DINA's orders, recruiting anti-Castro Cuban exiles to help. Chilean courts later convicted Contreras and fellow DINA officer Pedro Espinoza for ordering the killing. It remains the only state-sponsored assassination of a foreign diplomat ever carried out on American soil.

That documentary record eventually reached a courtroom verdict. After a trial lasting thirteen years, an Argentine federal court delivered its judgment on 27 May 2016, convicting former de facto president Reynaldo Bignone and fourteen other former military officers of crimes against humanity for their roles in Condor, sentencing several — including Uruguayan colonel Manuel Cordero Piacentini — to 25 years. It was the first time a court anywhere in the region had formally ruled, as a matter of law rather than historical argument, that the cross-border conspiracy among the Southern Cone dictatorships had existed and operated as charged.

What is genuinely still disputed

Condor's existence, its member states, its cross-border kidnappings and assassinations, and its reach onto U.S. soil are not in serious dispute among historians, courts or the declassified record itself. What remains a live, honest question — and where overclaiming becomes easy — is the precise degree of direct U.S. operational involvement, as distinct from U.S. knowledge and, in places, acquiescence.

The documentary record is genuinely damning on the question of what Washington knew and when. Kissinger's August 1976 briefing shows the State Department was aware, in advance, that Condor intended assassination operations in Europe. Cables from CIA Western Hemisphere division chief Ray Warren show officials debating whether to warn French and Portuguese counterparts about specific planned killings on their soil — and in at least the French case, taking some preemptive step to alert local authorities. State Department officials drafted a formal démarche intended to warn the Condor governments against carrying out assassinations abroad. But Kissinger, briefed on the démarche in early August 1976, instructed that no further action be taken — and the warning was never delivered to Chile or Argentina before the Letelier assassination six weeks later. That is a documented decision not to intervene, made at the highest level of U.S. diplomacy, with consequences that are hard to look past.

What the released documents do not show is a CIA hand on Condor's own steering wheel — no cable ordering a specific Condor hit, no record of the agency supplying the target list, no proof that Condor's Phase III assassination teams operated under American command rather than under the Southern Cone services' own authority. Some declassified material does describe U.S. training and communications support provided to allied Latin American intelligence services in this period, which critics reasonably read as material assistance to the same networks that ran Condor; defenders of the historical record note this support predated and was broader than Condor itself, and that “enabled through inaction or general Cold War patronage” is not the same finding as “directed the operation.” Serious historians of the period continue to debate exactly where along that spectrum — from bystander, to enabler, to co-author — the U.S. role should honestly be placed, and further U.S. document declassification could still move that line. It should not be dismissed as unknowable, but it also should not be flattened into “the CIA ran Condor,” a claim the current public record does not establish.

Why a proven conspiracy still shapes belief

For years before Almada's files surfaced, exiles, human-rights lawyers and relatives of the “disappeared” across the Southern Cone insisted that their governments were hunting people down across borders in a coordinated way — and were routinely told, by the regimes themselves and by skeptical outsiders, that this was exile paranoia or Cold War propaganda. The Archives of Terror, and the cables that followed, did more than confirm isolated abuses; they validated the specific claim that the abuses were coordinated, run by an actual joint structure with a name, a founding meeting and a budget. That is a rare and consequential kind of vindication, and it is easy to see why it left a durable mark on how the region — and outside observers of it — evaluate official denials generally.

It is also easy to see why that validated distrust sometimes overshoots the documented case. Because the U.S. role in Condor is confirmed to include real knowledge, some support and a deliberate choice not to warn regimes off a specific assassination campaign, the temptation is to round that up to full U.S. authorship — the idea that Condor was, at bottom, a CIA creation running Latin American proxies rather than a Latin American creation Washington chose largely to tolerate. The honest record supports the first claim (extensive American complicity through inaction and partial support) far more solidly than the second (direct American command), and the difference matters: it is the difference between a documented failure of U.S. foreign policy and an undocumented claim about who actually built and ran the machine.

The vindication was real: exiles who said their governments were hunting them across borders together were telling the truth. The lesson is not that every further escalation of the claim is equally true — it is that this one was proven, and proof is exactly what makes the boundary of what was proven worth respecting.