The Unknown

The C.I.A. and the Pentagon take another look at Al Qaeda and Iraq.

By Jeffrey Goldberg , The New Yorker, February 10, 2003

In April of 1998, President Clinton sent his United Nations Ambassador, Bill Richardson, to South Asia. Richardson's stops included New Delhi, Islamabad, and, most unusually, Kabul, where he held the first (and, as it turned out, the last) Cabinet-level negotiations between the United States and the Taliban leaders of Afghanistan. Richardson, who is now the governor of New Mexico, is an effective diplomat. (He returned to international diplomacy briefly last month, when he met with two North Korean envoys in Santa Fe.) He is irreverent, and he is not timid, and his trip might have been a diplomatic success if it had not been an intelligence failure.

During the stop in New Delhi, Richardson met with officials of the new Hindu-nationalist government of the Bharatiya Janata Party. In one encounter, Richardson asked the defense minister, George Fernandes, if his country planned to explode any of its nuclear weapons. The Indians had not tested their bomb since 1974, but in early 1998 the newspapers in New Delhi—and in Islamabad, the Pakistani capital—were filled with speculation about the new government's intentions. The B.J.P. had stated in its election platform that it would "not be dictated to by anybody in matters of security and in the exercise of the nuclear option."

Fernandes, a self-described pacifist, told Richardson that India had no intention of exploding a nuclear device. Then he changed the subject to the situation in Burma. In other meetings, Richardson was given the same soothing message, and the mission to India was so relaxed that the Assistant Secretary of State, Karl Inderfurth, who was managing the trip, spent part of one day trying to set up a cricket demonstration for Richardson, a former minor-league baseball player. The demonstration was interrupted only once, so that Richardson could receive a six-minute intelligence briefing from a New Delhi-based C.I.A. officer.

I accompanied Richardson on the trip, and he allowed me to follow him into many of his meetings, except for C.I.A. briefings. But it is clear that no one from the C.I.A. told Richardson that the Indians were about to explode five nuclear devices in the Rajasthan desert, which is what they did less than a month after the delegation left South Asia. Not long ago, one of Richardson's former top aides, Calvin Mitchell, told me, "Even after we returned from the region, we received no intelligence that the Indians had lied to us."

Richardson was equally ill-informed in Afghanistan. In a single day, we visited Kabul and Sheberghan, a town in the north held by anti-Taliban rebels, and flew back to Islamabad at sundown. It was a strange day; at one stop, a senior National Security Council official fell into a sewage ditch, and the NBC correspondent Andrea Mitchell was nearly trampled by a posse of Uzbek horsemen. Nothing was stranger, however, than the meeting with the Taliban.

"We have a whole range of issues we're going to bring up with the Taliban leadership," Inderfurth had told me the day before the trip. Osama bin Laden, the Taliban's most famous guest, "is just one," he said.

The American delegation was met at the Kabul airport by Taliban gunmen in pickup trucks, who drove us to the Presidential Palace. The freewheeling Richardson decided to include me in the delegation, telling me to identify myself as a "note-taker," should anyone ask. An honor guard of Pashtun fighters greeted us and led us through a series of musty corridors to a small room with gray walls. The room was undecorated, except for a bookcase holding the collected works of Washington Irving. We had a long wait before the Taliban delegation arrived. It was led by Mohammed Rabbani, the deputy to Mullah Omar, the Taliban leader (who rarely left Kandahar, and who in any case refused to meet non-Muslims). The Taliban men were ignorant of diplomatic niceties, and Richardson's icebreaking small talk was met by incomprehension. But Richardson gamely moved through the issues. He expressed the Clinton Administration's concern that the Taliban was shielding a terrorist. Rabbani, who sweated profusely throughout the meeting, responded, "He is our guest here. He is under our control."

Richardson persisted; so did the Taliban. Richardson consulted his State Department and N.S.C. advisers; we waited to see how far he would push the matter. He dropped it, and continued with the agenda, which included a discussion of the possibility of running an oil pipeline across parts of Taliban territory. We were then led to a banquet hall, where we were served rice and pigeon as gunmen circled the table.

Calvin Mitchell said, "We certainly didn't know much about Osama at the time. We didn't know the extent of his network or that he was bankrolling the entire Taliban."

When I reached Richardson recently at the governor's mansion in Santa Fe, he recalled his post-mission frustration. "When a foreign leader wants to deceive you, even the best intelligence is not going to prove in a foolproof way that the leader is deceiving you," he said. "But we need to have a better way of sensing the deception of foreign leaders."

Shortly after the failure to predict India's nuclear tests, George Tenet, the C.I.A. director, asked a retired Navy admiral, David Jeremiah, to conduct an investigation. At the time, Senator Daniel Patrick Moynihan—who had long emphasized a need to improve intelligence collection and analysis, as well as the oversight of the more than thirty-billion-dollar national intelligence apparatus—said, "The question is: Why don't we learn to read? What's the State Department for? The political leadership in India as much as said they were going to begin testing. There's a tendency at the State Department to say, 'Gee, the C.I.A. never told us.' "

Jeremiah found that the United States had an insufficient number of satellites focussed on India; that the intelligence community's photograph analysts were overworked and undertrained; and that the C.I.A. had too few spies on the ground. But underscoring all this, Jeremiah said, was a particularly American sort of assumption: both intelligence analysts and policymakers assumed that the Indians would not test their nuclear weapons because Americans would not, in similar circumstances, test nuclear weapons. In the world of intelligence, this is known as mirror-imaging: the projection of American values and behavior onto America's enemies and rivals. "I suppose my bottom line is that both the intelligence and the policy communities had an underlying mind-set going into these tests that the B.J.P. would behave as we behave," Jeremiah said at a press conference held to announce his findings.

America's early assessment of bin Laden was similarly flawed. In the American mind, of course, the bin Laden of April, 1998, was not the bin Laden of September, 2001. But his intentions were no secret. Two months before the Richardson meeting, bin Laden had issued a fatwa, a religious ruling, in which he called on Muslims to kill Americans—civilians and military. Yet, among the group of Americans travelling with Richardson five years ago, the fatwa was a passing source of black humor; the threat seemed too outlandish to be taken seriously.

In the foreword to Roberta Wohlstetter's classic 1962 study, "Pearl Harbor: Warning and Decision," the national-security expert Thomas Schelling wrote that America's ability to be surprised by the actions of its enemies is the result of a "poverty of expectations." He went on, "There is a tendency in our planning to confuse the unfamiliar with the improbable. The contingency we have not considered seriously looks strange; what looks strange is thought improbable; what is improbable need not be considered seriously."

Wohlstetter's work revealed that Pearl Harbor was not much of a surprise at all. It showed that the American government's fatal mistake was not a failure to pick up signals—overheard conversations, decoded cables, unusual ship movements—but a failure to separate out signals from noise, to understand which signals were meaningful, and to imagine that the Japanese might do something as irrational as attacking the headquarters of the U.S. Pacific fleet. In other words, the Americans heard the signals but didn't listen to them.

One day earlier this winter, I visited Fort Meade, outside Baltimore, which is the home of the National Security Agency, the country's main signals-intelligence group. The director of the N.S.A. is a cerebral, well-respected Air Force general named Michael Hayden, and I spoke to him about the challenges of signal collection. We sat in his office, a large room with a view of the N.S.A.'s obsessively guarded complex of black buildings. The office had been scrubbed of classified material in anticipation of my visit.

"Our noise-to-signal ratio is twenty to one, that one being something useful," Hayden told me. "Not necessarily tactically useful, just remotely useful. But even this is misleading, because it's twenty to one after we've done all sorts of things to make it humanly intelligible. You have to collect, process, translate, move it down the funnel, transform it from noise into a signal, before you know if it's useful."

I asked Hayden whether he thought Pearl Harbor or September 11th had been the greater surprise. "Pearl Harbor was, essentially, not a surprise," he said. "It was that one could not divine the meaningful signals from the thousands that were out there." He thought about the question a little longer and added, "I'm going to say, and I might change my mind, perhaps it was more a failure of imagination this time than last. We failed to see how absolute their"—Al Qaeda's—"world view is. A signals-intelligence agency gets inside the head of an adversary, if you're doing your job at all. You get to know the inside of a target. But I don't think we properly appreciated how capable and how different, how evil, that mind-set is."

Hayden also suggested that September 11th was the greater surprise, because the United States was, in effect, already at war with bin Laden. "Al Qaeda had attacked us before," he pointed out, "and we had a broad effort against the group." He noted that, after the 1998 bombings of two American embassies in East Africa, Tenet had told the intelligence community that he was "declaring war" on Al Qaeda. Nevertheless, Hayden said, America was surprised.

I asked Donald Rumsfeld, the Secretary of Defense, the same question when, in late January, I met with him in his office in the Pentagon. Rumsfeld was moving troops to the Persian Gulf that day, but in our discussion he focussed on the role of surprise in intelligence failure. He wasn't interested in assigning blame for the failure to predict or stop the September 11th attacks; in fact, he objected when I used the word "failure," preferring, as Tenet does, the word "defeat." He said, "When you hear people criticizing the agency, I think it's important to reverse it and say, 'People ought to really critique how professional and how substantive the users of intelligence are in contributing critical feedback.' " In other words, the blame for "defeats" in intelligence can be ascribed as much to executive-branch policymakers and the intelligence committees of Congress as it can to Middle East analysts in the cubicles of the C.I.A. Echoing Moynihan's argument, Rumsfeld also said that policymakers mistakenly assume that information must be secret in order to have value. "There's something about me, I suppose, and others possibly, where we read intel and we begin to think that this is the sum total of what we know about a subject and not really go in and probe the open sources, which are rich in many cases," he said.

In the late nineteen-nineties, Rumsfeld, who was then working in private industry (he had already had one tour as Secretary of Defense, under President Ford), chaired a commission set up by Congress to examine the ballistic-missile programs of America's rivals and enemies. The commission, whose members included Rumsfeld's current deputy, Paul Wolfowitz, concluded that American intelligence agencies did not possess the analytic depth or the right methods of analysis to accurately assess the threat. "Intelligence assessments and estimates must be grounded in the facts," the commission concluded. "But to be useful, they cannot be limited to reporting only what is known about a particular program. Yet, in a large number of cases examined, Commissioners found analysts unwilling to make estimates that extended beyond the hard evidence they had in hand, which effectively precluded developing and testing alternative hypotheses about the actual foreign programs taking place."

Rumsfeld is especially drawn to Schelling's theory of surprise; he believes that surprise is often the by-product of analytical timidity. "The poverty of expectations—the failure of imagination—I found this just so interesting," Rumsfeld said. "We tend to hear what we expect to hear, whether it's bad or good. Human nature is that way. Unless something is jarring, you tend to stay on your track and get it reinforced rather than recalibrated. If I as a policymaker fail to make a conscious decision that you want to go around three hundred and sixty degrees and test things, you're likely to stay in a rut. And we've seen our country do that."

Rumsfeld believes that one long-held belief among Middle East analysts is overdue for reconsideration: the idea that doctrinal differences prevent Sunni and Shiite Muslims, and religious and secular Muslims, from pursuing common projects in anti-American terrorism. This is a subject of great relevance today, because the Bush Administration contends that Baghdad is a sponsor of Al Qaeda; critics of the Administration's foreign policy argue that bin Laden and Saddam Hussein are natural enemies. "The argument is that Al Qaeda has got a religious motivation, somehow or other, and the Iraqi regime is considered to be a secular regime," Rumsfeld said. "The answer to that is, so what? The Iraqi regime will use anything it can to its advantage. Why wouldn't they use any implement at hand?"

Rumsfeld's work on the ballistic-missile commission convinced him that intelligence analysts were not asking themselves the full range of questions on any given subject—including what they didn't know. Rumsfeld gave me a copy of some aphorisms he had collected during the process of assessing the ballistic-missile threat. "Some of these are humorous," he said, not quite accurately. One was "There are knowns, known unknowns, and unknown unknowns." (The saying is attributed, naturally, to "Unknown.") "I think this construct is just powerful," Rumsfeld said. "The unknown unknowns, we do not even know we don't know them."

During the commission's hearings, Rumsfeld went on, intelligence analysts would brief the commission members. "They'd say, 'This is a fact.' And we'd say, 'Well, when did you learn it?' 'On this day, X.' 'And when did it start?' 'Back here, several years back.' And, of course, it's embarrassing. When you get some pieces of information, the implication is, you know, that you've done a good job. But the real question is: When did it actually start and when did we find out about it?"

In the case of the missile programs of two countries he would not name, he said, "There were instances in which we didn't know something until two, four, six, eight, twelve, and, in one case, thirteen years after it happened. If we didn't know this for five years, that means that there may very well be things that started five years ago that we don't know about at all."

Rumsfeld said that the ideas contained in the commission's report are spreading through the fourteen organizations that make up the intelligence community (these range from the Defense Intelligence Agency to Coast Guard intelligence). "You find not infrequently now that there will be a section, and it will have a fairly typical analysis, and then it will be followed by a section labelled 'What we don't know.' "

There have been frequent reports of tension between the Defense Department and the C.I.A., particularly on the question of the links between Saddam and Al Qaeda. But the two men who lead these bureaucracies have kind words for each other.

"Rumsfeld should get a hell of a lot of credit for challenging the conventional wisdom, for challenging the bureaucracy," Tenet told me not long ago in his office at C.I.A. headquarters, in Langley, Virginia. Tenet, who does not conform to the patrician mold of the C.I.A.'s early directors—he is Greek-American and a proud son of Queens—seemed tired on the gray morning we met. His office is long and narrow; a torn American flag rescued from the ruins of the World Trade Center hung on the wall over his shoulder.

When Tenet and Rumsfeld talk about intelligence theory, it is hard to see major differences. Of his own agency Tenet said, "We spend a great deal of time encouraging analysts to get out of their own skins, to try to think the way the enemy thinks." He also said, "We're emphasizing the point, as the saying goes, that intelligence work is often not about evidence but about the absence of evidence." (Aides to both Tenet and Rumsfeld claim that their man devised this formula.)

Tenet, who became the C.I.A. director in 1997 (he was a Clinton appointee), gained the trust of President Bush early in the Administration, and he has survived thanks to the C.I.A.'s work during the war in Afghanistan; to his desire to rebuild the C.I.A.'s clandestine service, which had fallen into disrepair during the Clinton years; and to the fact that, while the C.I.A. committed major blunders in the days leading up to September 11th, the F.B.I.'s mistakes were catastrophic. In addition, he is able to communicate effectively on Capitol Hill.

In his office, Tenet listed questions that he thinks should be asked about the C.I.A.'s performance. "Are we making steady progress in penetration? Yes. Do we have success in technical intelligence-gathering? Yes. Did we stop the attacks? No." Tenet went on, "The Israelis probably have as ironclad a structure to deter terrorism as anybody in the world, but they continue to lose people. The way they think about this is that they are in a constant state of war and there are wins and losses in a war. This changes their mind-set." He added, "Failure means you're not paying attention. You knew about the target but you didn't do anything about it. Defeat is what happens in a war against an agile enemy who will find ways to beat you."

Tenet doesn't think that America has experienced the worst of Al Qaeda terrorism. "My worry is that, whatever we learn, it will never be enough. There will be another attack. They will take advantage of seams in our security. If you're looking for infallibility in the intelligence system, you're going to be constantly surprised and pained. You're dealing with an enemy that has studied your homeland extremely well."

As in the days before Pearl Harbor, there was no lack of meaningful signals in the weeks before September 11th. In fact, Tenet spent the summer of 2001 hectoring lawmakers and Administration officials about the danger he saw on the horizon. "George just badgered me," Porter Goss, the chairman of the House Permanent Select Committee on Intelligence, told me. Goss, a Florida Republican who is a former C.I.A. officer and an ally of Tenet's, said that he had agreed to be badgered. "I understood what was happening," Goss said. "But we failed anyway. We knew what was coming, but we could not sing the song in a key that anyone would hear."

The C.I.A., of course, did not know the time or the place of the anticipated attack, and so could offer no tactical intelligence. It made at least one critical error, when it learned, long before September 11th, that two of the hijackers, who had attended a crucial Al Qaeda planning meeting in Malaysia, were travelling to America. Despite this knowledge, the C.I.A. did not place the men's names on terrorist watch lists until August of 2001.

Tenet, while admitting that his agency made mistakes, said that excessive focus on such mistakes obscures larger truths. "People can fault individual decisions," he said. "But it's vignette-driven. 'Why didn't you watch-list these particular guys?' But with these vignettes you're going to miss the bigger picture—that they're going to come in using different techniques next time." Many intelligence officials assume that watch lists have only limited value; terrorists, they say, will travel to post-9/11 America using false documents.

The desire to identify the people responsible for intelligence failure is natural, but it may be irrelevant, according to Thomas Schelling, who teaches at the University of Maryland. "I tend to think there's too much interest in finding blame for September 11th," Schelling said. "Surprise has two very different meanings. One is 'I didn't expect it,' and the other is 'I couldn't anticipate everything.' It seems to me that if smart people had somehow made a list of one hundred potential Al Qaeda targets, and then from that figured out that they might hijack airplanes and use them in attacks, they would still have a hard time telling the officials of airports what to be on the lookout for, or what these men would need to hijack a plane. I tend to think that what Al Qaeda did was beautifully conceived but not terribly difficult to do. Once they had the concept, the rest was easy."

On the question of failure of imagination, Tenet seemed to disagree with Rumsfeld. "We don't have an absence of imagination," he said. "What we have is threat fatigue. We're inundated by this stuff. We have to guard against numbness as we go through intense periods of threat reporting."

Top C.I.A. officials told me that analysts in the agency's Counterterrorist Center had imagined the airborne suicide attack as a tactic, but had also imagined dozens of other ways in which terrorists could strike American targets. I asked one official why his analysts could not match a target to a technique—why they couldn't guess that the World Trade Center, which had been the target of one terrorist attack, in 1993, would be the target of another, from the air. "We had reports over the last six or eight years that Al Qaeda people are interested in aviation," this person told me. "You have concrete knowledge that the World Trade Center is a target. What keeps those two pieces of information from joining? Well, if we took every credible tactic we hear about and applied it to every credible target, the alarm would be sounded every day."

But Tenet said that his analysts have been encouraged to extend themselves, to lower the threshold for what is credible. In intelligence, he said, "very few snippets of evidence can take you to a judicial conclusion. Nothing is crystal clear." He's also pushing his analysts to think in different ways. "We're moving people away from linear thinking," he said, and added, "It takes years and years to walk back the risk aversion in a bureaucracy, but we're doing it."

In the ideological taxonomy of the Bush Administration, the C.I.A., because it has long downplayed the theory of an Iraq-Al Qaeda connection, has been regarded as being on the side of the doves. The hawks have accused the C.I.A. of politicizing the intelligence process by dismissing information that would substantiate the connection—and in that way strengthen the Administration's case against Iraq.

A key moment in this argument took place one weekday last August, when a small group of Defense Department officials drove from the Pentagon to the headquarters of the C.I.A., for what they expected to be a tension-filled meeting with the agency's top analysts. Leading the Pentagon team was Douglas Feith, the Under-Secretary of Defense for policy, who is considered to be an Iraq hawk in the style of his superiors, Wolfowitz and Rumsfeld. Feith brought with him a Defense Intelligence Agency analyst, Tina Shelton, and a Naval intelligence reservist, Christopher Carney. The Defense Department had asked Shelton and Carney to reëxamine evidence collected by the C.I.A. about the relationship between terrorist networks and their state sponsors, including Iraq and Al Qaeda, and to re-analyze the data in the manner suggested by Rumsfeld's ballistic-missile-threat commission; that is, to build a hypothesis, and then see if the data supported the hypothesis, rather than the reverse. "If you take thirty movie reviewers and show them the same movie," Feith told me, "they will understand its meaning in thirty different ways, and they will even understand the plot in different ways, and I'm not talking about watching 'Rashomon.' "

The presentation was made in a small conference room, and as many as twenty C.I.A. executives and analysts crowded in, along with the director of the D.I.A., Vice-Admiral Lowell E. Jacoby, and Tenet himself. According to several people with knowledge of the meeting, Carney and Shelton told the C.I.A. officials that, based on their own reading of agency intelligence, it appeared likely that Saddam's relationship with Al Qaeda was serious and that it dated back to the terror group's early days in Sudan. Bin Laden had his headquarters in Khartoum in the early nineteen-nineties, before moving to Afghanistan, in 1996. "These people weren't hired to do alternative analysis," Feith claimed. "But once they read deeply into the material, which, by the way, was good C.I.A. material, they came up with some fresh connections and ideas and analysis." Feith went on, "When we fed this analysis back into the C.I.A., they were happy to receive it. Tenet understands, as Rumsfeld understands, that an extra set of eyes on intelligence material is a good thing."

The Defense team had expected resistance from C.I.A. officials, but, to the surprise of many in the room, Tenet was open to the Pentagon analysis. However, one top official familiar with Tenet's thinking told me that early last year, well before the August meeting, C.I.A. officials had asked the agency's Red Cell team, an internal think tank, to undertake "a different sort of analysis, 'go a little more hypothetical on the question, and see what you come up with.' They gave us a report, and it seemed pretty hypothetical. But then it stopped seeming so hypothetical."

There's nothing new about hypothesis-driven analysis. Angelo Codevilla, a Boston University professor of international relations and a former senior staff member of the Senate Intelligence Committee, pointed to the risks—in particular that "you put the monkey on the back of the policymaker. This is dangerous if you have policymakers who don't want to hear more than one opinion. George Bush, just like his father, doesn't like to be faced with choices." Codevilla also said that analysts could "stray too far from the data," adding that the real problem is that "the C.I.A. has not been gathering enough quality data." According to a senior Administration official, the C.I.A. itself is split on the question of a Baghdad-Al Qaeda connection: analysts in the agency's Near East-South Asia division discount the notion; the Counterterrorist Center supports it. The senior Administration official told me that Tenet tends to agree with the Counterterrorist Center.

When I saw Tenet, I asked if he now considered Saddam to be a primary sponsor of Al Qaeda. "Well, read my letter to Senator Graham," Tenet replied.

In October of 2002, when Bob Graham was the chairman of the Senate Intelligence Committee, Tenet wrote to him, explaining the C.I.A.'s understanding of the Iraq-Al Qaeda connection. It is a curious letter, which begins with a statement that "Baghdad for now appears to be drawing a line short of conducting terrorist attacks with conventional or CBW"—chemical and biological weapons—"against the United States." At the same time, Tenet said, Iraq has "provided training to Al Qaeda members in the areas of poisons and gases and making conventional bombs." Tenet added, "Credible information indicates that Iraq and Al Qaeda have discussed safe haven and reciprocal non-aggression," and he suggested that, even without an American attack on Iraq, "Baghdad's links to terrorists will increase."

The evolution of Tenet's beliefs has made those opposed to an invasion of Iraq uneasy. Senator Graham thinks that the C.I.A.'s "evolved" understanding of the Iraq-Al Qaeda connection is the result of pressure from Rumsfeld. "Maybe the C.I.A. has been coöpted in this whole thing," Graham told me. "I'm not personalizing it to George, but institutionally the C.I.A. is being challenged by a very aggressive Defense Department."

Others who have watched Tenet, however, say that he does not trim his opinions for political reasons. "I find him to be a straightforward person on analysis," Nancy Pelosi, the House Minority Leader, who until recently was the ranking Democrat on the intelligence committee, told me. Pelosi added that she considers Iran a greater terrorist threat than Iraq.

Tenet's thinking on the subject was deliberate, according to several agency sources. Information gleaned from the interrogations of high-level Al Qaeda prisoners pushed Tenet to rethink the opinion, advanced by C.I.A. officials such as Paul Pillar, the National Intelligence Officer for the Middle East, that ideological differences between the secular Saddam and Islamic radicals, such as Al Qaeda, made it unlikely that these two enemies of America would form an alliance. Clearly, the Rumsfeld view, which maintains that the commonly held hatred of the United States trumps ideology and theology, is ascendant, at the C.I.A. as well as at the Pentagon. Pillar himself, in a faxed comment, conceded that, "despite major differences, tactical coöperation is possible," but added that "the contingency that would be most likely to motivate Saddam to develop a relationship with radical Islamists that would be deeper than limited tactical cooperation would be a belief that he was about to lose power"—such as in a United States-led attack on Iraq.

According to several intelligence officials I spoke to, the relationship between bin Laden and Saddam's regime was brokered in the early nineteen-nineties by the then de-facto leader of Sudan, the pan-Islamist radical Hassan al-Tourabi. Tourabi, sources say, persuaded the ostensibly secular Saddam to add to the Iraqi flag the words "Allahu Akbar," as a concession to Muslim radicals.

In interviews with senior officials, the following picture emerged: American intelligence believes that Al Qaeda and Saddam reached a non-aggression agreement in 1993, and that the relationship deepened further in the mid-nineteen-nineties, when an Al Qaeda operative—a native-born Iraqi who goes by the name Abu Abdullah al-Iraqi—was dispatched by bin Laden to ask the Iraqis for help in poison-gas training. Al-Iraqi's mission was successful, and an unknown number of trainers from an Iraqi secret-police organization called Unit 999 were dispatched to camps in Afghanistan to instruct Al Qaeda terrorists. (Training in hijacking techniques was also provided to foreign Islamist radicals inside Iraq, according to two Iraqi defectors quoted in a report in the Times in November of 2001.) Another Al Qaeda operative, the Iraqi-born Mamdouh Salim, who goes by the name Abu Hajer al-Iraqi, also served as a liaison in the mid-nineteen-nineties to Iraqi intelligence. Salim, according to a recent book, "The Age of Sacred Terror," by the former N.S.C. officials Daniel Benjamin and Steven Simon, was bin Laden's chief procurer of weapons of mass destruction, and was involved in the early nineties in chemical-weapons development in Sudan. Salim was arrested in Germany in 1998 and was extradited to the United States. He is awaiting trial in New York on charges related to the 1998 East Africa embassy bombings; he was convicted last April of stabbing a Manhattan prison guard in the eye with a sharpened comb.

Intelligence officials told me that the agency also takes seriously reports that an Iraqi known as Abu Wa'el, whose real name is Saadoun Mahmoud Abdulatif al-Ani, is the liaison of Saddam's intelligence service to a radical Muslim group called Ansar al-Islam, which controls a small enclave in northern Iraq; the group is believed by American and Kurdish intelligence officials to be affiliated with Al Qaeda. I learned of another possible connection early last year, while I was interviewing Al Qaeda operatives in a Kurdish prison in Sulaimaniya. There, a man whom Kurdish intelligence officials identified as a captured Iraqi agent told me that in 1992 he served as a bodyguard to Ayman al-Zawahiri, bin Laden's deputy, when Zawahiri secretly visited Baghdad.

Ansar al-Islam was created on September 1, 2001, when two Kurdish radical groups merged forces. According to Barham Salih, the Prime Minister of the Patriotic Union of Kurdistan, the group seized a chain of villages in the mountainous region outside the city of Halabja, and made a safe haven for Al Qaeda fighters. "Our intelligence information confirmed that the group was declared on September 1st at the behest of bin Laden and Al Qaeda," Prime Minister Salih told me last week, in a telephone conversation from Davos, Switzerland. "It was meant to be an alternative base of operations, since they were apparently anticipating that Afghanistan was going to become a denied area to them."

Salih also said that a month before the September 11th attacks a senior Al Qaeda operative called Abdulrahman al-Shami was dispatched from Afghanistan to the Kurdish mountain town of Biyara, to organize the Ansar al-Islam enclave. Shami was killed in November, 2001, in a battle with the pro-American forces of the Patriotic Union of Kurdistan.

The Ansar al-Islam enclave, according to Salih and American intelligence officials, soon became the base of operations of an Al Qaeda subgroup called Jund al-Shams, or Soldiers of the Levant, which operates mainly in Jordan and Syria. Jund al-Shams is controlled by a man named Mussa'ab al-Zarqawi, a Jordanian of Palestinian extraction. Zarqawi is believed by European intelligence agencies to be Al Qaeda's main specialist in chemical and biological terrorism. Zarqawi is also believed to be behind the assassination, on October 28th, of an American A.I.D. official in Jordan, and also two unsuccessful assassination attempts: last February 20th, Ali Bourjaq, a Jordanian secret-police official, escaped injury when a bomb detonated near his home; and on April 2nd gunmen opened fire on Prime Minister Salih's home in Sulaimaniya. Salih was unhurt, but five of his bodyguards were killed; two bystanders were killed in the Bourjaq assassination attempt.

The Administration believes that Zarqawi made his way to Baghdad after the United States' invasion of Afghanistan, when he was wounded. According to American sources, Zarqawi was treated in a Baghdad hospital but disappeared from Baghdad shortly after the Jordanian government asked Iraq to extradite him. American intelligence officials believe that Zarqawi was also among an unknown number of Al Qaeda terrorists who have sought refuge in the Ansar al-Islam over the past seventeen months.

Recently, I asked two former C.I.A. directors, James Woolsey and Robert Gates, to talk about the problem of analyzing an incomplete set of evidence—the same challenge that stymied intelligence analysts in the days before December 7, 1941, and September 11, 2001.

Woolsey, who served as President Clinton's first C.I.A. director, said that it is now illogical to doubt the notion that Saddam collaborates with Islamist terrorism, and that he would provide chemical or biological weapons to Al Qaeda. "At Salman Pak"—a training camp near Baghdad—"we know there were Islamist terrorists training to hijack airplanes in groups of four or five with short knives," Woolsey told me. "I mean, hello? If we had seen after December 7, 1941, a fake American battleship in a lake in northern Italy, and a group of Asian pilots training there, would we have said, 'Well, you can't prove that they were Japanese'?"

Gates, who was C.I.A. director under George H. W. Bush, said that the evidence linking Saddam to Al Qaeda is not irrefutable, but he noted that ambiguous evidence is an occupational hazard in intelligence work. Gates suggested that the current debate over Iraq's ties to terrorism is reminiscent of a debate about the Soviet Union twenty years ago. Then, he said, "you had analysts in the C.I.A. who said, 'Absolutely not, it would be contrary to their interests to support unpredictable, uncontrollable groups.' There were other analysts who said, 'Baloney.' They had a lot of good history, and circumstantial reporting on their side, but they didn't have good evidence. Once the Soviet Union collapsed, and we got hold of the East German Stasi records, we learned, of course, that both the East Germans and the Soviets were supporting Baader-Meinhof and other terrorist groups."

Gates continued, "I have always argued, in light of my fairly detailed knowledge of the shortcomings of our intelligence capabilities, that the fact that we don't have reliable human intelligence that proves something conclusively is happening is no proof at all that nothing is happening. In these situations, the evidence will almost always be ambiguous. On capabilities, it's not ambiguous. Can Saddam produce these weapons of mass destruction? Yes."

The ambiguity, Gates said, has to do with "intentions," and he went on, "If the stakes and the consequences are small, you're going to want ninety-per-cent assurance. It's a risk calculus. On the other hand, if your worry is along the lines of what Rumsfeld is saying—another major attack on the U.S., possibly with biological or chemical weapons—and you look at the consequences of September 11th, then the equation of risk changes. You have to be prepared to go forward with a lot lower level of confidence in the evidence you have. A fifty-per-cent chance of such an attack happening is so terrible that it changes the calculation of risk." ♦