In 1970, Christopher Pyle disclosed in public writing that the U.S. Army was running a domestic intelligence program aimed at anti-war and civil-rights activists. His disclosure began a series of public-accountability leaks, the most famous of which was Daniel Ellsberg's leak of the Pentagon Papers. These disclosures, the FBI's notorious abuses in COINTELPRO, and the Watergate leaks taught Americans that the national-security system can take profoundly dangerous constitutional turns. They formed the foundation of the political will that led to the passage of the Foreign Intelligence Surveillance Act in 1978. And although Ellsberg and his collaborator Anthony Russo were prosecuted, their cases were dismissed.

This series of events that helped cement the role of unauthorized public disclosure as a systemic check on the predictable cycles of error in the national-security system. The leakers of the 1970s became heroes who exposed systemic failure, and the nation did not punish those who helped it correct the excesses.

For a quarter of a century, leaks continued apace, though not the kind of public-accountability disclosures that typified the early 1970s. These were the normal grist for the mill for the national press: gossip and backstabbing, trial balloons and glorified insider war stories—leaks that, as David Pozen showed, existed long before the 1970s and have continued without noticeable increase or decrease ever since. Only one leak, the clearly improper disclosure of satellite images of a Soviet aircraft carrier to Jane's, was ever prosecuted, and the norm of not prosecuting for leaks to the press was so strong that Senator Daniel Patrick Moynihan persuaded President Bill Clinton to pardon that one offender.

It was only in 2002, when Jesselyn Radack disclosed that the prosecution of “American Taliban” John Walker Lindh had involved several violations of Lindh's constitutional rights, that the new wave of whistleblowing and public-accountability leaks reemerged. Thomas Tamm and Russ Tice each disclosed the Bush administration's warrantless-wiretap program to The New York Times; AT&T employee Mark Klein disclosed the company’s complicity in illegal wiretapping; and William Binney, Thomas Drake, and others challenged internally—and in Drake's case disclosed publicly—early aspects of NSA dragnet surveillance. Chelsea Manning disclosed a major document cache to WikiLeaks, driven by what she viewed as American forces’ callous disregard for civilian casualties and silent complicity in Iraqi government torture. Most critically, Edward Snowden's disclosures led to the introduction of dozens of bills in Congress, a judicial opinion, and two executive-branch independent reviews that demanded extensive reforms to surveillance programs.

As with the Vietnam-era leaks, the wave of accountability leaks in the past dozen years was the result of the national-security establishment overstepping its constitutional boundaries in a state of emergency. Reacting to the shock of 9/11, the national-security system responded as best it could. Some of the responses were measured and necessary. Others were abusive mistakes followed by aggressive cover-ups. The army ordered an inquiry into allegations of torture by military personnel, then forced out the general who dutifully reported the "sadistic, blatant and wanton" "systematic and illegal abuse" he found. The CIA embraced torture with a vengeance, lied about it to the administration and Congress, and went so far as to spy on its Senate overseers in its attempt to cover its tracks. The NSA built a surveillance apparatus that overstepped its statutory and constitutional authority, and in the process of normalizing that apparatus dragooned parts of the judiciary and Congress into providing a veneer of legitimacy to its actions. Viewed in this light, the war that President Obama and his attorney general have waged on whistleblowers can no longer be reasonably seen as simply a response to increased leaks in general. Instead, it reflects an increasingly embattled national-security system pushing deeper into barely defensible or indefensible constitutional dead ends and ramping up aggressive criminal prosecutions to defend itself from the ultimate failsafe for stopping abusive systems: individual conscience.

One thing is clear. Without the men and women of conscience who have come out over the past 12 years and disclosed aspects of the abuses, the system would have kept on grinding.

The single most important lesson of Snowden's disclosures is that even well-designed and well-intentioned systems of checks and balances become corroded and subverted over time. No matter how perfect a set of rules or an institutional or organizational system was when it was created, it cannot remain so in the face of time, change, and the pressures of new emergencies. The FISA system that emerged from the 1970s worked reasonably well in the latter stages of the Cold War, when it was grappling with the kinds of threats and the technological environment for which it was designed. But as Binney put it, everything changed after 9/11. The urgency of preventing another attack swept away the rules. Partly in response to heroic efforts of some insiders to change policies, and mostly due to the public exposure of the warrantless wiretapping program, elements of the judicial-oversight system were reintroduced starting in 2006, and Congress passed the FISA Amendments Act in 2008. But the practices that had developed had become so complex and diverse that the new oversight system was completely inadequate to contain them.

The failures of the FISA system, both old and new, are not unique. All large systems suffer from these kinds of failures as they age, as new conditions challenge old practices, and as the rationale for processes once cherished is lost in the humdrum of bureaucratic routine. Democrats might prefer to focus on General Motors or Lehman Brothers. Republicans will emphasize the Obamacare website or the Department of Health and Human Services. And almost everyone points to the FEMA Katrina response. Unless one believes that the national-security system has a magical exemption from the dynamics that plague any of these other large organizations, one should expect large and substantial mistakes from the complex interaction between the Department of Defense, the CIA, NSA, and FBI and their various contractors, from Boeing and Lockheed Martin to Halliburton and Booz Allen Hamilton.

Whistleblowing is a central pillar of the way American law deals with these dynamics of error, incompetence, and malfeasance in large organizations. From workplace-safety violations to Medicare and Medicaid fraud to anti-corruption campaigns around the world, we protect and reward those who follow both internal procedures and those who expose abuse to the public. Internal audit and review processes, while important, need the backstop of insiders with knowledge.

But national security is different. There are limited protections for internal whistleblowers, and none at all for those who go to the press. Defenders of that approach argue that the critical nature of national security justifies complete secrecy. But that very critical nature also means that mistakes can have devastating effects, while the secrecy that national-security organizations demand makes them more likely to get stuck in erroneous patterns.

Secrecy disables many of the mechanisms that other systems use to correct failure dynamics. In the public sector, informed and interested outsiders facilitate robust judicial, legislative, or executive oversight. In the private sector, both the stock market and regulators rely on public information and disclosure requirements to punish error, incompetence, and malfeasance. From defective products to poor business judgment, information flow is the critical ingredient of correction. The internal and external information silos that typify national security and the secrecy and mystique of the agencies all disable the standard mechanisms we use to counterbalance the error dynamics of other large organizations. The complexity and uncertainty of the threats the national-security system faces compound the difficulties so that even insiders—to say nothing of outsiders—struggle to evaluate whether an element of the system is working or has gone off track.

Because it is practically impossible for outsiders to check the national-security system, protecting insider whistleblowers is especially critical. The primary goal of 1970s FISA reform was to create a system of “insider-outsiders.” A few select members of Congress (the intelligence committees) and a few select judges (the FISA Court) would represent the normal public-accountability process. But the case of NSA dragnet surveillance showed that these watchdogs were hobbled by secrecy and resource imbalances that made it impossible for them to function as real backstops to abuse. The result was oversight theater: a spectacle of oversight used to lull critics and defuse public demands for reform. Only repeated acts of conscientious disclosure by insiders have brought us close to a public airing of the limits of surveillance.

Reform that tries to rebuild the well-behaved aspects of the system but ignores the critical circuit-breaker role played by unauthorized whistleblowing is set up for failure. Respecting leaks and leakers is no slight to the patriotism or professional bona fides of people in the national-security system, the congressional staff, or the judicial officers who form the insider-review system. It is an admission that it is practically impossible to build a purely self-contained review regime.

Conscience is not correlated with well-behaved systems. It doesn't come from having a particular position of trust, line of education, or role in oversight. It comes from the diverse ways we grow up and is shaped by our experiences and triggered by different pressures. Conscience can move anyone, from the most loyal technical insider like Binney to the retired telephone-company employee like Klein, so it functions like a random inspection. This lack of predictability makes unauthorized public-accountability leaking so critical when all the authorized systems feed back into each other to compound and legitimate a failure.

Improving legal mechanisms for well-behaved whistleblowing is therefore important but insufficient. A less subservient version of the current system for national-security employees, expanded to include contractors, could help raise and solve waste or poor management. It might even be useful in raising narrow legal and civil-rights abuse concerns, as long as they involve practices that are not seen as mission critical. But they are likely useless for raising and addressing basic strategic decisions, such as whether to create a surveillance system capable of capturing everything, and building it with so much redundancy that it is effectively outside of any review. Decisions of that sort are too deeply ingrained in institutional practice and mission pursuit to be susceptible to internal whistleblowing, and those who try are likely to be suppressed, just as Binney, Drake, and their co-workers were. Only piercing the echo chamber can lead to meaningful reform in such cases, so it’s here, where the risks of error and distortion are greatest, that unauthorized disclosure is most important. We saw it with the Pentagon Papers in 1970, and we saw it again with Snowden.

This is why immunity for Edward Snowden is so essential. In principle, there should be a public-accountability defense in criminal law, similar to self-defense and defense of others. I have written about why we should create such a defense and how to design it. But Congress should also introduce a simple direct intervention: adding retroactive immunity for Snowden to the NSA reform bill currently under consideration on Capitol Hill. Retroactive immunity would simply mirror immunity granted in the 2008 FISA Amendments Act to telecommunications companies that violated the law by collaborating with the illegal surveillance, and which the White House has sought to extend to other firms that handed over private data in the new reform bill.

Retroactive immunity would build constitutional culture rather than a permanent legal solution. Our (fuzzy) memories of the 1970s teach us, collectively, that unauthorized national-security leakers who expose substantial wrongdoing were heroes, and that respect, not a prison term, was their due. That is the lesson that immunity for Snowden would reinforce. It will not make leaking a low-risk activity, nor will it erase the dread of repercussions like Manning’s 35-year prison sentence. But immunity will be a strong statement to insiders that if the system has gone badly enough off track, and if public disclosure can lead to genuine benefits, then a conscientious individual can do the right thing. Even if the leak is illegal, the public will support bona fide whistleblowers who expose significant abuses, and the whistleblowers will not be forced to spend their lives in prison or exile while those whose misdeeds they exposed profit on the speaking circuit.