Getting Smart: Three steps toward a more intelligent intelligence community.

After spending three months loudly berating the intelligence community about 9/11, Congress has quietly decided to postpone any reforms until September. The most pressing and least settled issues on the national- security agenda will therefore be debated -- and perhaps decided -- on the eve of midterm elections. In that politically charged climate, the temptation for lawmakers to grandstand on intelligence will be great -- probably greater than at any time since the 1970s, when Democratic senator Frank Church of Idaho, looking toward a presidential bid, excoriated and eviscerated the CIA.

This time, however, the political dynamic may redound to the public good. The knock against our spies now is not that they're "rogue elephants," but rather that they're desk mice. Given this new atmosphere, it's expected that some of the worst of the Church restraints -- constitutionally needless limits on surveillance and covert action -- will be loosened.

Yet undoing Church will not, of itself, cure our intelligence dysfunction. Our spy system is, in fact, plagued by three longstanding problems, which the Church measures exacerbated but did not create. These interlocking disorders should be addressed, over the short, medium, and long term, by a three-step plan.

First, in the near term, the U.S. must rehabilitate its counterintelligence -- a vulnerability that has been virtually ignored during the reform debate. "CI," as it's known in the spy trade, disrupts the intelligence work of other nations and encourages allied spy agencies to share their secrets with us. Contrary to public perceptions, successful spying hinges less on Le Carre-esque betrayals than on cultivating international cooperation; during World War II, according to one government report, some 85 percent of our strategic intelligence came from the British. Yet liaison of this kind is always difficult work, the perennial cause -- as former CIA counterintelligence officer William R. Johnson has written -- of "peptic ulcers, high blood pressure, anxiety, depression, paranoid behavior, and bad dreams in CI executives."

Only a reputation for good CI on our end will make other services confident that they can share information with us without putting their sources at risk. During most of the Cold War, the CIA's foreign liaison was managed by the counterintelligence staff, under James Jesus Angleton; and though Angleton was later demonized for his relentless mole-hunts, friendly services liked the fact that he was always hunting. On his watch, liaison -- with the British and Israelis in particular -- was unprecedentedly close.

In 1973, alas, Angleton's suspicions about the sincerity of Soviet detente provoked the wrath of CIA director William Colby. In a power play designed to depose Angleton, Colby decentralized, and effectively destroyed, CI. The counterintelligence staff once vetted sources and controlled access to files; it was now to serve only an "advisory" function. Case officers would be allowed to assess the bona fides of their own agents; and to ensure that each case officer knew "enough" to assess his recruitments, the CIA's files would be decompartmentalized. For the past three decades, in other words, the spy world's equivalent of virus-protection software has been turned off.

The results have been catastrophic. Though no spies are known to have penetrated the CIA on Angleton's watch, our major spy agencies have since been riddled with moles: Hanssen, Montes, Ames. And as agencies rush to hire translators and analysts of foreign background, our vulnerability to espionage will only increase.

There are already signs of penetration among handlers of counterterrorist wiretaps. Sibel Edmonds, a former translator in the FBI's Washington field office, has raised suspicions about a coworker's connections to Middle Eastern persons whose intercepted conversations the Bureau was translating. "Investigations are being compromised," Edmonds wrote the FBI inspector general in March. The Bureau has confirmed to the Washington Post that Edmonds's coworker had "unreported contacts" with a foreign official targeted by the taps.

Clearly, until we fix our CI system, foreign spy services will have good reason to hold out on us. And that will have disastrous consequences, because we are relying increasingly on such liaison for access to cultures in which Americans can't discreetly operate. As former CIA director Robert Gates has put it: "You need a guy walking into Tripoli or Pyongyang who doesn't look like he just left Iowa."

Wholesale reform in CI must be undertaken now -- and it can be: Recentralization, a return to the Angleton model, could be mandated by an executive decision. The draft text of just such an order was prepared by Reagan staffers in March 1981; Bush staffers should use it as a model.

Such a measure would not bear full fruit, however, without broader bureaucratic reform. "Counterintelligence," Angleton himself once said, "is only as good as relations between the CIA and FBI." This is true of counterterrorism also, and it points to the next necessary reform: insisting on more effective teamwork between the CIA and the FBI.

This problem is far older than the recent bungles. In 1948, Republican presidential nominee Thomas Dewey made FBI-CIA rivalry an issue in his campaign. Since then there have been no fewer than twelve White House initiatives to defuse the interagency conflict. All have failed, because they've been based on the bad idea of melding law enforcement and intelligence.

"Intelligence and law enforcement," British spy Ian Fleming wrote in a June 1941 memo, "simply do not mix." The two functions proceed from opposite and irreconcilable imperatives. Intelligence takes pains to protect sources; law enforcement uses sources to convict criminals in public. Secret intelligence goes on offense to stifle nascent threats; law-enforcement agents react to crimes that have already occurred. In Britain, these dual purposes are effectively detangled: Domestic intelligence (MI5) is separate from criminal investigation (Scotland Yard). But in the U.S. the functions have been smashed together, and each has been turned to purposes for which it was not designed. The bad results could fill a book.

The Bush administration is continuing to enforce this unnatural union, and demanding of cops that they be spies. The Homeland Security Plan states unequivocally: "The Department of Justice, and the FBI, will remain the lead law enforcement agencies for preventing terrorist attacks." Attorney general John Ashcroft has given the FBI a "new mission of preventing future terrorist activities." But a new mission does not, in itself, provide the new tools and methods necessary to fulfill the mission. To have obtained and interpreted the leads that could have helped prevent 9/11 -- to have searched Zacarias Moussaoui's laptop computer, and connected its contents with clues from other sources -- the FBI would have had to be a radically different outfit. It would have had to be a domestic spy agency, like Britain's MI5.

Terrorism is secret warfare, and it must be countered by a corps of secret warriors. Accordingly, the FBI's secret intelligence work should be severed from its public law-enforcement functions. Instead of being absorbed by the new Homeland Security agency, all internal-intelligence officers should be spun off into a new agency with spy powers but no law- enforcement mandate. The result would be a British-style system, in which criminal investigation would be handled by the national police (FBI), foreign intelligence would be managed by a foreign-spying unit (CIA), and internal security would be done by a domestic-espionage agency (the new unit). Sen. Bob Graham, chairman of the Senate Intelligence Committee, has already suggested splitting law enforcement and domestic intelligence into two agencies; so have others, including Newt Gingrich. The Bush administration should get on board.

Creation of a domestic spy unit, however, would be only a medium-term fix. The third reform we should undertake -- the most long-term measure of the three -- is junking the distinction between domestic and foreign intelligence. As money, technology, information, and people cross borders with increasing ease, attempts to distinguish foreign- from domestic-based threats make progressively less sense. The 9/11 hijackers lived in Florida but were supported from Germany, and met contacts in the Czech Republic and Spain. Was their conspiracy, then, a domestic or a foreign phenomenon? The impossibility of answering that question coherently shows the bankruptcy of the paradigm.

We have to revise our thinking about what national security is, how it is threatened, and how we must organize our defense. (Here again, the British are ahead of us: Conservative leader Iain Duncan Smith has criticized prime minister Tony Blair for attempting to distinguish between domestic and foreign terror threats.) In the U.S., there's an urgent need for the CIA to engage in domestic intelligence. The agency needs a clear mandate, for instance, to cooperate with and train urban police: Though terrorism cannot be treated primarily as a criminal matter, it's vital that local law-enforcement officers be trained to contribute any intelligence byproduct of their beat work. The CIA also needs authority to stay in "hot pursuit" when overseas suspects cross into the U.S. The National Security Act should be amended to let the CIA work within our borders -- in specific cases, against defined threats, under presidential findings, akin to those that provide authority and oversight for covert operations overseas.

We also have to ask: If the old, geographically based model of intelligence work is obsolete, what should replace it? Should we eventually combine the CIA and our domestic spy agency into a single unit, with one chief -- and thus no need to share information, no turf battles, no bureaucratic rivalry? Gregory F. Treverton, former vice chairman of the National Intelligence Council and now a senior policy analyst at RAND, is already proposing something along these lines. President Bush should, at a minimum, appoint a commission to examine the issues involved in de- provincializing spy work, and have it report to him before the end of his term.

There will be objections to the idea of domestic spying; protest will become more and more vocal as the rules of engagement within our borders come to mirror those in the wider world. The entire enterprise of domestic spying, it will be argued, runs counter to the American idea. Wasn't the CIA, for this very reason, denied all "internal security" functions? Shouldn't we fear erasure of the bright boundary between the Machiavellian methods we sanction overseas, and the Madisonian scruples that must constrain surveillance at home?

To treat these questions fully would require a Lockean treatise on secret government. We do not have one. But we do have history, and our Constitution, and neither requires us to draw the lines we do. Neither our Founders nor our 19th-century presidents ran covert actions cartographically; nor does the experience of European democracies suggest that internal spying breeds dictatorship; nor has the Supreme Court ever ruled unconstitutional the domestic surveillance of suspected foreign agents; nor does the Constitution anywhere mandate that intelligence jurisdictions conform to the shapes of oceans, rather than to the judgments of men.

Politics does compel consideration of the popular will, and firm action will not be uncontroversial. But this autumn may well offer an historic opportunity for rational, responsible change. If Congress and the president choose not to seize it -- and if the innocent are once more butchered en masse -- the pressures for less appropriate severities, for detention camps and mass deportations, will be far more difficult to deny.