The State Confesses

This chapter requires no inference: the state already confessed. And the confession closes a loop drawn in the very first chapter.

In 2009, the Senate Select Committee on Intelligence began investigating the CIA’s torture program after learning the agency had destroyed videotapes of its own interrogations. Over roughly six years, the committee reviewed more than 6.3 million pages of the CIA’s internal cables, emails, and records, and released a five-hundred-twenty-four-page executive summary in December 2014.¹ What makes it the keystone case is who designed the program, and where they learned how. The contract psychologists James Mitchell and Bruce Jessen, whose actual background was in SERE training, the military program that teaches American personnel to resist torture, designed a menu of ten enhanced interrogation techniques.² Per the CIA’s own records, Mitchell and his colleagues consulted the work of Martin Seligman, the psychologist whose dogs, examined in the first chapter, taught science what learned helplessness actually is.³ This is not an analyst drawing a retrospective comparison. It is a documented, named process: a laboratory finding about uncontrollable shock, knowingly and by name, applied by a state to break human beings. The firm Mitchell and Jessen ran for this purpose was ultimately paid eighty-one million dollars — meaning the people recommending the program’s continued use had a direct financial stake in its expansion, a conflict of interest the Senate’s own report documents explicitly.⁴

Hold that fact against the marriage this book opened with. The mechanism Mitchell and Jessen looked up in a psychology journal and sold to a government for eighty-one million dollars is the same mechanism Richard Challen ran on his wife for more than thirty years without ever knowing it had a name, a citation, or a laboratory behind it. These two borrowings are not the same kind of link: the CIA’s is homology — a documented citation, Seligman’s learned-helplessness research invoked by name as the program’s scientific basis — while Challen’s is analogy, the same move reinvented by instinct, with no connection to the lab at all. The documented state-to-lab link vouches for itself, not for the husband’s “sameness,” which rests on resemblance alone. The method did not change across that distance — the reach did, and the record did. What Richard Challen left his wife to reconstruct alone, years later and against a murder conviction, the CIA wrote down in its own cables as it happened. Only one of those two records was ever meant to be read; both, in the end, proved the same thing.

The report’s central finding is not that the program was severe. It is that the CIA misrepresented its severity to the White House, Congress, the Department of Justice, and the public — every time it described the program to any of them. Abu Zubaydah was waterboarded at least eighty-three times; Khalid Sheikh Mohammed at least a hundred eighty-three times in a single month.⁵ Detainees underwent “rectal rehydration,” described in the CIA’s own cables not as a medical necessity but as a means of asserting total control.⁶ Sleep deprivation ran as long as a hundred eighty hours, often standing, producing hallucinations.⁷ Gul Rahman died of hypothermia at a CIA black site in Afghanistan after being left shackled and partially nude on a cold concrete floor; no one was disciplined.⁸ At least twenty-six of a hundred nineteen known detainees in the program were later determined to have been wrongfully held entirely — meaning the program tortured people it never had grounds to suspect of anything.⁹ And the report directly contradicts the CIA’s repeated public claim that these techniques produced unique intelligence unobtainable any other way, tracing its touted successes to sources that predated or ran independent of the program.¹⁰ This matters beyond morality. Every belief system examined here survives by being unfalsifiable from the inside. Here, the justification itself was checked against the agency’s own evidence and found false — institutionally, not merely doubted.

No criminal charges were ever filed. The Department of Justice, under two administrations, declined to prosecute, including a 2012 decision specifically declining charges related to Gul Rahman’s death.¹¹ The accountability mechanism that actually functioned was the Senate report itself — exposure and documentation, not punishment. That is the realistic ceiling of accountability at this scale.

The origin of SERE training is the cleanest proof that coercive technique is separable from the intent behind it. Developed after the Korean War specifically in response to American prisoners’ false confessions under Chinese and Soviet interrogation, SERE schools run simulated captivity under medical supervision — mock camps, stress positions, sleep deprivation, simulated waterboarding — with the explicit goal of stress inoculation, so a captured service member doesn’t experience these techniques for the first time in a real captor’s hands.¹² Mitchell and Jessen were themselves senior Air Force SERE psychologists — their careers spent on the resistance side of this exact apparatus.¹³ A separate 2008 Senate Armed Services Committee investigation found that personnel from the agency overseeing SERE training briefed interrogators at Guantanamo in 2002, and that senior administration lawyers sought out SERE-derived methods specifically because they had already been used on American service members and could therefore be argued not to constitute torture.¹⁴ No new technique had to be invented. The entire offensive program was built by taking an existing defensive-training inventory and pointing it in the opposite direction — the same sleep-deprivation protocol, the same simulated drowning, functioning as resistance training or torture depending entirely on who was authorized to use it on whom. Two separate congressional investigations, approaching from different angles, converged independently on this same pipeline.

In Vietnam, the same deniability ran not through a charismatic leader but through an org chart. The Phoenix Program, formalized in 1967, targeted the Viet Cong’s shadow political infrastructure rather than its combat units; provincial coordination committees built dossiers from informant networks and tasked paramilitary units, often recruited from former Viet Cong defectors, with capturing, turning, or killing the people named.¹⁵ What makes Phoenix distinct is its quota system: American advisors set monthly “neutralization” targets per province, and the system didn’t merely permit abuse — it incentivized local officials to manufacture targets to satisfy a number set by someone else. Personal enemies and debtors were denounced as enemy infrastructure to settle scores; village chiefs used the resulting blacklists to eliminate rivals.¹⁶ This is diffusion of responsibility built directly into an org chart: the American advisor can say he only set a number, never pulled a trigger; the Vietnamese official can say he was only following the quota. Official figures counted roughly eighty-two thousand people “neutralized” between 1968 and 1972, more than twenty-six thousand of them killed; the Church Committee’s later estimate ran higher still, with genuine dispute over what fraction were actually enemy infrastructure versus non-combatants caught by quota pressure — a range, not a settled number.¹⁷ William Colby, who ran the program for much of its existence and later became CIA Director, testified in 1971 that Phoenix was “an essential part of a very legitimate war program… seriously distorted on occasion” — moral justification and diffusion of responsibility, offered by the architect of the very quota structure that made the distortion structurally inevitable.¹⁸

After World War II, NATO and the CIA established secret “stay-behind” armies across Western Europe, originally meant to conduct resistance in the event of a Soviet invasion. Italy’s branch, codenamed Gladio, was confirmed to exist by Prime Minister Giulio Andreotti in 1990.¹⁹ During Italy’s “Years of Lead,” a series of bombings — including the 1980 Bologna massacre, the deadliest terrorist attack in Italian history — were initially blamed on left-wing militants, but judicial investigations eventually implicated neo-fascist groups with documented ties to Italian intelligence and the Gladio network, producing real convictions.²⁰ The state’s own apparatus never needed to plant a single bomb with its own personnel: it cultivated proxies who would, then let public fear justify expanded security-state power while discrediting the political left electorally. What makes this case unique is a participant’s own on-record confession: Vincenzo Vinciguerra, convicted for a 1972 bombing later traced to Gladio, testified plainly that the attacks were not the work of a madman but a coordinated state strategy designed to push the public toward demanding more security and authority — the clearest first-person admission of cui bono reasoning anywhere in this literature, stated not by an analyst reconstructing motive after the fact, but by the man who carried it out.²¹

Inside East Germany, deniability required no proxy and no quota — only population-scale infrastructure. At the Stasi’s peak, the agency ran an estimated hundred seventy to a hundred ninety thousand active civilian informants against a population of sixteen million people, a ratio of roughly one informant for every ninety citizens — and, counting the occasional and one-time collaborators, denser still — among the most pervasive peacetime surveillance networks any modern state has built.²² These informants were recruited from neighbors, coworkers, doctors, and — extensively documented — spouses, parents, and children informing on each other. The dissident Vera Lengsfeld discovered, after German reunification granted citizens access to their own files, that her husband had been informing on her for years under a codename, including details from inside their own marriage.²³ This is the isolation mechanism the opening chapters described — implemented not as a tactic against one target but as the default condition an entire society lived inside. As the regime collapsed, Stasi officers ran out of time to shred the archive and tore documents by hand instead, leaving hundreds of millions of paper fragments. A German research institute’s algorithmic reconstruction project has spent years since piecing the torn record back together — deniability attempted through destruction, defeated decades later by forensic reconstruction rather than by any paperwork surviving intact.²⁴

A seven-hundred-page internal CIA compendium, declassified in 2007 and known as the “Family Jewels,” catalogs the agency’s own illegal acts in its own words: opening and photographing roughly a quarter million pieces of first-class mail, wiretapping journalists, break-ins at former employees’ homes, unwitting LSD testing on American citizens, and assassination plots against foreign leaders.²⁵ The LSD program shows deniability engineered not just into an operation but into its own erasure. Codenamed MKUltra and run by CIA chemist Sidney Gottlieb from 1953 to 1973, the program dosed subjects who had not consented and frequently did not know they had been drugged at all. At CIA-run safehouses under the cover name Operation Midnight Climax, paid sex workers lured men back to apartments wired for observation, so operatives could watch a stranger’s reaction to a drug he did not know he had been given.²⁶ In 1973, CIA Director Richard Helms ordered most of the program’s records destroyed, specifically ahead of an anticipated Senate investigation into the agency’s own conduct. What survives of MKUltra’s public record exists largely because a cache of financial documents had been misfiled in a budget office rather than destroyed with the rest. That surviving cache, not any functioning oversight mechanism, forced a dedicated 1977 Senate hearing confirming the program had existed at all — distinct from, and following, the Church Committee’s own 1975–76 investigation.²⁷ That earlier Church Committee separately documented the CIA’s secret relationships with roughly fifty American journalists and its financial influence over media outlets, in direct violation of its own charter — though the popular name for this, “Operation Mockingbird,” does not actually appear in the congressional record and is partly mythologized, even as the underlying fact of media assets and influence is real and documented.²⁸

Every prior case shows a state agency running coercion or laundering its own legality; this one shows the outside professional body whose entire function was to be an independent ethical check instead coordinating with the very agency it was supposed to check. Once it became public that psychologists had designed the CIA’s interrogation program, the American Psychological Association convened a task force in 2005 to determine whether its own ethics code permitted such participation. The task force concluded it could, using language permissive enough to provide ongoing cover.²⁹ After a journalist’s 2014 reporting raised collusion allegations, the APA commissioned its own independent investigation. The resulting 2015 Hoffman Report found that a majority of the task force’s voting members had undisclosed ties to the Department of Defense or the intelligence community, and that the APA’s own ethics director had privately coordinated with Pentagon officials to steer the task force toward a predetermined, permissive outcome — while publicly representing the process as independent — findings those named in the report have since disputed, in litigation the APA partly settled.³⁰ The APA’s governing council voted, following the report, to formally ban psychologist participation in national-security interrogations, and the ethics director was fired.³¹ This is the clearest documented instance of an institutional layer being captured not alongside state coercion but in direct service of it: the professional body’s ethics process functioned as the deniability layer itself, and saying yes let the program claim its psychologists operated within sanctioned professional ethics. It is also the rare case in this chapter where exposure produced an actual institutional consequence — a firing, a reversed policy — rather than documentation alone, and worth asking why: a professional association can dismiss an employee in a way a government has never managed to prosecute itself.

The lab science was imported knowingly. The legal cover was sought knowingly. The professional-ethics cover was manufactured knowingly. Three mechanisms, three institutions, one underlying logic — and every claim tracing back to the government’s own records, released, eventually, to the public it had been withheld from.