In the fall of 1954 the director of the Bureau of Prisons (BOP), James V. Bennett, asked Warren E. Burger, the head of the Justice Department’s Civil Division (and future chief justice), for guidance on a number of questions relating to the medical treatment of prisoners. Burger referred one of the questions to the Justice Department’s Office of Legal Counsel (OLC), whose interpretations of law are binding across the executive branch. The question concerned the use of “truth serums”—drugs that at the time were believed to compel people to answer questions truthfully against their will. Bennett wanted to know whether the Constitution permitted BOP personnel to administer these drugs to prisoners without their consent.

In his letter to Burger, Bennett maintained that BOP personnel sometimes needed ways of getting truthful information from prisoners—for example, to determine whether they needed particular kinds of treatment or whether prison rules had been broken. The BOP had used sodium amytal in the past, he said, to find out whether an inmate was faking illness to avoid having to participate in a prison’s “rehabilitation program.” In one case at Terre Haute, a “suspected malingerer” was given an injection of sodium amytal and then interviewed by a psychiatrist. Through this procedure, “malingering was established.” The problem was that “the inmate was very agitated because he had not given his consent and threatened to sue the doctor for malpractice.”

The OLC addressed Bennett’s question in a six-page legal opinion signed by J. Lee Rankin, the assistant attorney general in charge of the office. (Two years later, Rankin became the solicitor general.) Rankin wasn’t unsympathetic to Bennett’s request. He noted that “the treatment and supervision of prisoners is a difficult and exacting process” and that the necessity of controlling the prison population presents a “continuous challenge.” Prison officials, he wrote, are “assumed to have broad authority in the administration of accepted medical treatment and procedures to prisoners committed to their custody.” But he observed that the authority of prison officials is not without bounds. He quoted Justice Robert Jackson’s concurring opinion in Skinner v. Oklahoma (1942), in which the Supreme Court invalidated a statute that authorized the compelled sterilization of “habitual offenders”:

There are limits to the extent to which a legislatively represented majority may conduct biological experiments at the expense of the dignity and personality and natural powers of a minority—even those who have been guilty of what the majority defines as crimes.

Rankin’s opinion was made public only recently. It was one of several hundred OLC opinions that the Justice Department turned over last year to settle a Freedom of Information Act lawsuit brought by the Knight First Amendment Institute at Columbia University, which I direct. But during the seven decades that passed between the time Rankin wrote the opinion and the time it was released, the question of what measures government agents may lawfully take to compel prisoners to talk arose again.

Soon after the September 11 attacks, the CIA asked the OLC what interrogation methods it could use against the suspected al-Qaeda leader Zayn al-Abidin Muhammad Husayn—known as “Abu Zubaydah”—whom the agency had imprisoned at a black site in Thailand. FBI agents had been interrogating Abu Zubaydah and believed they had established a rapport with him. The CIA was convinced, however, that the use of what it called “enhanced interrogation techniques” could compel him to share more information about the attacks and other al-Qaeda plans.

On August 1, 2002, the OLC issued a legal opinion drafted by John Yoo (now a law professor at Berkeley) and signed by Jay Bybee (now a federal judge) that endorsed a collection of sadistic methods ranging from “walling” and “stress positions” to “sleep deprivation” and “the waterboard”—methods the CIA ended up using not only on Abu Zubaydah but on other prisoners as well.

The OLC document authorizing the torture of Abu Zubaydah was styled as a legal opinion, but it was an embarrassment to the genre, twisting both the facts and the law to evade statutory prohibitions against the brutal treatment of prisoners. Another opinion issued by the OLC that same day argued that the prohibition against torture applied only to an extremely narrow range of practices—those causing the kind of pain associated with “organ failure, impairment of bodily function, or even death”—and added that the president had constitutional authority to override even that prohibition.

In testimony to Congress, one prominent legal scholar called the opinions “an ethical train wreck”: “I believe it is impossible that lawyers of such great talent and intelligence could have written these memos in the good-faith belief that they accurately state the law.” A later investigation by the Justice Department’s Office of Professional Responsibility found that Yoo and Bybee had committed professional misconduct—“intentional” professional misconduct, in the case of Yoo—by failing “to exercise independent legal judgment and render thorough, objective and candid legal advice.”

There was an era in which the OLC was sometimes described as the “constitutional conscience” of the executive branch. It’s difficult to say whether it deserved this label, because so much of its work is secret. The Freedom of Information Act requires executive agencies to publish their interpretations of the law, redacting them only as necessary to protect privacy, national security, or other specific, identified interests. In practice the OLC publishes its opinions selectively, justifying this mainly with the dubious argument that they are “deliberative” in character even when they reach conclusions that bind other federal agencies. The result is that we have access to only a part of the OLC’s voluminous oeuvre—and we don’t know what we don’t know. Most of the OLC’s old opinions have never been published, though many of them still shape government policy today. And while the OLC has written fewer opinions in recent years, many of these are secret, too. The Abu Zubaydah opinion was published only after the ACLU spent years litigating for its release.

Still, to read Rankin’s opinion is to begin to understand why some might once have associated the word “conscience” with an office that in recent years has been known principally for green-lighting torture, mass surveillance, and extrajudicial killings. One of the many defects of the OLC’s Abu Zubaydah opinion was its credulousness—it accepted unquestioningly the CIA’s assertions about the way the proposed interrogation methods would be used, what risks they would pose to the prisoner, and whether they would be effective. But Rankin doesn’t take as a given that sodium amytal and similar drugs actually work. He briefly reviews the available evidence. He quotes one authority: “The term ‘truth serum,’ like ‘lie-detector,’ is a misnomer. There is no specific which will cure or reveal deception—no drug which will infallibly induce truth.” Such drugs could at most be relied on to induce “a state of mind…whereby an individual’s capacity to…choose those [thoughts] to which he desires to give utterance is inhibited.” Rankin doesn’t reach a firm conclusion, but he writes, “The degree of reliability which can be ascribed to the result of the use of such drugs is open to serious question.”

Where Yoo and Bybee’s reasoning is sophistic, myopic, and rigged, Rankin’s is straightforward, considered, skeptical. At one point Rankin turns to the question of whether the involuntary administration of sodium amytal would invade a prisoner’s right of privacy. He writes:

Although the drug application would lack the physical discomfort of the rack and the screw, it nevertheless constitutes a method of third degree which could have lasting consequences beyond physical pain.

Other differences are striking as well. Yoo and Bybee appear to assume that the authority they extend to the CIA will be used exclusively in laboratory-like conditions and only against the one prisoner. Rankin is unwilling to make similar assumptions, and so he concerns himself with the “oppressive uses to which the practice” proposed by the BOP could be put: “If the technique were considered acceptable in detecting malingering, would not its use be even more inviting in solving prison crimes or infractions?” And, he asks, why would we stop there?

Aside from any dramatic quality, could it not be argued that continuous insight into the thought processes of all of the prisoners would permit the prison authorities to take greater and more effective precautionary measures against disorder and worse, and to better plan for rehabilitation?

Rankin writes that “these ends are pleasant in contemplation,” but he snaps quickly out of his reverie: “The price in terms of personal degradation is higher than the law allows.”

It hardly needs to be said that the question presented to Yoo and Bybee was very different from the one presented to Rankin, even if both questions concerned the permissibility of forcing prisoners to talk. The CIA is not the BOP, terrorism is not malingering, and a black site is something different from a federal prison. We don’t know how Yoo and Bybee would have answered the question that was put to Rankin, and we don’t know how Rankin would have answered the one that was put to Yoo and Bybee.

And yet it does seem fair to say, based on available evidence, that these lawyers had radically divergent understandings of their responsibilities. Yoo and Bybee appear to have thought of themselves chiefly as lawyers for the Bush administration, and so the question they put to themselves, when the CIA sought permission to use methods the US had once condemned as torture, was whether the law could be made to accommodate those methods even if they were, on their face, grotesque and even freakish and precisely what the laws in question were meant to forbid. (“You would like to put Abu Zubaydah in a cramped confinement box with an insect” is how Yoo and Bybee describe one of the CIA’s proposals—a proposal that they go on to endorse.)

Rankin seems to have understood his obligations very differently. His opinion is attentive to the interests of the agency he is advising, yet its final allegiance is not to the agency’s policy preferences but to deeper principles that structure our government and society. His analysis is an effort to discern what is sometimes referred to as the “best” view of the law. Rankin concludes with a quote from Justice Hugo Black’s opinion in Chambers v. Florida (1940), a case in which Thurgood Marshall persuaded the Supreme Court to overturn the murder convictions of four black men whose confessions the police had procured through coercion and intimidation:

We are not impressed by the argument that law enforcement methods such as those under review are necessary to uphold our laws. The Constitution proscribes such lawless means irrespective of the end.

Today the OLC instructs its lawyers to provide their clients with the “best understanding of what the law requires”—to provide “an accurate and honest appraisal of applicable law, even if that appraisal will constrain the Administration’s or an agency’s pursuit of desired practices or policy objectives.” On paper, at least, Rankin’s view of the OLC’s mandate has prevailed. Whether it has prevailed in reality is less clear. How OLC lawyers are actually reconciling their loyalty to the administration they serve with fidelity to the rule of law is, in present circumstances, mostly unknowable.

For at least two decades transparency and human rights advocates have been calling on the OLC to publish its final legal opinions as a matter of course—not just its historical opinions, but its contemporary ones as well—arguing that it is antidemocratic to deny the public access to opinions that shape government policy, delineate the boundaries of government power, and establish the scope of individual rights. After conferring with constitutional scholars and litigators as well as many former OLC lawyers, the American Constitution Society issued a statement two years ago urging the OLC to release, among other things, a comprehensive index of its opinions. Last year Senators Patrick Leahy and Tammy Duckworth introduced a bill that would subject the OLC to additional transparency requirements.

For now, though, what the public knows about the OLC’s work is still almost entirely up to the OLC itself. This is a shame. Recent history makes all too clear how profound the costs to our society can be when the OLC subordinates constitutional principle to political expedience. It should not be a mystery to the public how the OLC is interpreting the law, or whether it is doing the work we need it to do.