Richard Kerr served in the CIA from 1960 to 1992, including three years as deputy director for intelligence (1986-'89) as deputy director (1989-'92) and a few months as acting director in 1991. In 2003, at the suggestion of Donald Rumsfeld, a group was put together to review the intelligence on Iraq's weapons of mass destruction program, and Kerr was asked to head it. "The secretary of defense actually wrote a letter to the director of central intelligence and said, 'It would be useful to have a study ... of what intelligence said up to the point of the war and then use the ground truth to kind of test those assumptions and those judgments,'" Kerr recalls. His group ultimately published four reports (two of which remain classified): The first looked at pre-war intelligence on Iraq; the second evaluated the raw intelligence that went into the infamous National Intelligence Estimate; the third assessed the strengths and weaknesses of intelligence analysis; and the fourth suggested improvements. Here, Kerr discusses his findings; his thoughts on the proper role and the future of the CIA; and his impressions of Dick Cheney and George Tenet. This is the edited transcript of an interview conducted on Jan. 25, 2006.

• Some highlights from this interview
• His review of the intelligence on Iraq's WMD
• Was too much political pressure placed on the intelligence community?
• The impact on the CIA of 9/11 and the invasion of Iraq
• His interactions with Dick Cheney

Describe for me the Central Intelligence Agency you walked into [when you first started]?

It was a very mature organization, I would say, in terms of the age of the people who were there. During World War II, many of them had been in OSS [Office of Strategic Services]. It was the group that came out of the Ivy League, the Yalies. Well educated. ... One of my mentors had been a journalist; one ... had been a professor of literature at Princeton. ...

So with that came a great deal of experience behind you: A lot of support, a lot of criticism, very severe organization in terms of writing skills and precision about how you wrote, how you presented things in the analytic world. And a lot of discipline about that. ...

And the DO [Directorate of Operations]? What were they like?

DO -- curmudgeonly group. Again, a lot of field people with a lot of experience, people who worked very independent, chiefs of station who had better access than ambassadors in the country they worked, because they delivered something real. As opposed to criticizing the government for its policies, they could actually deliver real things to a country. ...

And the idea of this notion that seems to persist even today, that the CIA was a left-leaning, liberal agency? Where does that come from?

It may come out of the very group of people that I'm talking about in the sense that they were people out of New England, out of the more liberal colleges, perhaps. I didn't sense that in my experience. I didn't sense the organization as a liberal organization particularly. ...

The people that come into the organization have to go through the polygraph, so you narrow the population down to a considerable degree in that very process. It is no surprise to you that later in the 1970s, the people that probably had the easiest time getting through were Mormons, because of their lifestyle. Some of the people who had a harder time getting through were people who came out of the Berkeleys, and some of the areas where there was a more liberal attitude. ...

[There was criticism that the CIA missed some major events, such as the fall of the Berlin Wall; India's testing a nuclear weapon.] ... What was the effect as you look across the span of those years and other events, like the Church committee hearings, things like Iran-Contra? How did the agency change? ...

Maybe this is a personal reaction, but there's less impact than one might expect because you have a job to do. ... These issues that you raise, while they're important in terms of the background and they're very important in terms of what happens in the front office of the director or the deputy director, the rest of the organization tends to go on and do its job. It has tasks to do, it has real things, real problems.

You go through the same period you're talking about, I'll take exception, for instance, to your thing that we missed the wall. We may have missed the actual coming down of the wall, but what we did is beautifully handle the Velvet Revolution and the transition from East Europe communist-dominated countries to what amounted to independent countries. ...

[It's said that Cheney was unhappy with the CIA's intelligence on the USSR.]

No, I take exception to that. ... I had a lot of contact with Cheney when he was a congressman on the oversight committee, when he was secretary of defense and I was the deputy director. My view of him was he was a good, thoughtful critic, and that's exactly, from my perspective, what you wanted: You wanted someone who was interested, and he was very interested. He would take time out when he was on the House Permanent Select Committee [on Intelligence] to come to the agency and spend a Saturday talking to analysts about a subject or problem. Not very many congressmen do that.

Why would he do that?

Because he was interested in it. He was interested, in this case, in the Soviet Union, arms control. ... I don't think he was critical of intelligence in a fundamental way. He was questioning of intelligence, asking and probing. And from an intelligence officer's point of view, the most impressive thing you can have is consumers that are interested in what you do, question it and test the system. Because then what you can say to that is somebody's paying attention to what we do. The worst thing that can happen to an intelligence officer, in my view, to be ignored. ...

You're talking about the period of the late '80s, I suppose?

I'm talking about when he was on the House Permanent Select Committee. And then he became, of course, secretary of defense. When he was secretary of defense, I was the deputy director of central intelligence. I used to meet with him, usually once a week for lunch, with [Director of Central Intelligence] Bill Webster, and we would talk about issues between CIA and Defense. In my view, he presented the least bureaucratic approach to organizational issues ... of anybody that I dealt with in Defense.

It's so fascinating, because the myth is that Cheney is this guy who ... will want to manipulate and push the Central Intelligence Agency?

You said two different words there: manipulate and push. Now, I never found in any situation that I dealt with of his trying to push you to do things you didn't believe, try to argue you into saying something that you didn't want to say. What you did find was the questioning of, "How do you know this? How confident are you? What about this source? Why did you choose to think about this particular information in one way and this in another? How did you make that choice, and how did you integrate those two ideas?"

So it was not an issue from my perspective of trying to manipulate, it was an issue of trying to push you to be as precise and clear, trying to get you to think about problems in a way perhaps you hadn't thought about.

I talked to Richard Clarke yesterday, and ... we're talking about the discovery [after the Gulf War] that there was a much more advanced nuclear program in Iraq than anybody had expected -- the way Clarke tells the story, than the CIA had articulated. Cheney at that moment says, "Oh, my God. These guys have not told me what I need to know. We did not know how close Saddam Hussein was to the nuclear weapon." And that at that moment, Dick Clarke says, Cheney walks away from the CIA and says, "Whoa, I can't trust these guys."

There's no question that when we went in and found that the Iraqis were working on several different approaches to try to get fissionable material, including centrifuges and a lot of other tech, we did not know that prior to the time we actually went in and saw it. ... But I would question that that was a breaking moment. We certainly did not know as much about the program as we should have, or as we could have, perhaps, and we might have liked to have. But we did know an awful lot about the Iraqis. ...

There are some general things about policy-makers and the CIA that are important to understand. In our reporting, for instance, on Iraq, we tried to lay this out. [Policy-makers] are much more comfortable with our judgments on technical areas because they're not technical for the most part. And they are relying on an organization that collects secrets and technical information, has analytic people who are working problems like nuclear weapons and other kinds of weapons. ...

The CIA said that there were nuclear weapons programs, [biological weapons] programs and [chemical weapons] programs under way. ... We also talked about a whole variety of other things: We talked about what might happen after a war began; how the army might react; what the stay-behind problem would be; ... the problems of guerilla tactics afterwards; the problems of bringing together the Sunnis and the Shi'a.

For the most part, our customers, I think, paid far less attention to us in those areas, even though we had considerable expertise, because they have their own views of this kind of softer area of the future. They have their own contacts, they can come to their own judgments, and they're less reliant on a CIA to give them an idea about what's going to happen once U.S. forces enter and you're faced with a whole new situation. ... I think the CIA actually did some things, wrote some things, provided some analysis, that would have been very valuable and should have been capitalized on and had more questions asked about them. ...

But if you're a policy-maker -- or a customer, as you call them -- and you want to go to war, or you think you should, you might even go so far as to ... get your own information unfiltered by intelligence.

I wouldn't put it in those terms that you just did. What I would say is if you're not at all certain about CIA's analysis, and you believe that there is information out there that is not being utilized the way you think it should be utilized, information that's not being interpreted the way you think, you might well create your own assets, your own analysts to deal with that.

But I wouldn't argue that's to create information; that's to look at it with different eyes from a different perspective. And it's not surprising that Defense, for instance, created its own internal organization. It's not surprising that INR [Bureau of Intelligence and Research] at State had a different view. They serve a different set of customers; they're driven, in my view, more by their customers' needs and desires than CIA is.

The one advantage CIA has that it reports to one person: That's the president of the United States, through the national security adviser. And my experience is that I've been able to do things over my career that you could not do in Defense or you could not do in State Department in terms of analysis and judgments about hard problems, telling things where they did not match neatly with policy. ...

I was never told to change a document, to tone it down, to change the yeses to no in 30 years, and the jobs I had, most of those jobs dealt with hard problems. ...

So when the vice president of the United States writes on a piece of paper, "This is so much better than the crap I get from the CIA," what does he mean by that?

Kind of have to talk to him. He may have had a bad day. He may have been disappointed in some things that he was given or not given. But I would have a hard time believing that that was a general condition; I think that was probably more in response to a specific problem. I've had all kinds of, over the years, policy-makers saying, "This doesn't tell me anything I didn't know. I don't agree with this at all. I think you're wrong." ...

I think there's a kind of a trend in administrations, and I characterize it this way: When a new administration comes in and has relatively little experience with intelligence, it looks on it as a great font of knowledge. It'll tell you everything. You can just go to it and it has all this information, and all you need to do is ask the right questions. After a while, ... they find that intelligence doesn't know the answers to all the problems, because there are no answers to some of the problems. ... So they tend to lose a little bit of enthusiasm for it. They get discouraged, and they go to other sources. ... So it's a cycle, and it's a cycle based on being disappointed that CIA can't answer questions that are unanswerable. ...

Tell me about George Tenet.

He was a staffer for the Senate Select Committee, senior staffer. Thoughtful and kind of probing because that's their job. More critical, perhaps, of intelligence generally, but someone you could deal with, talk to, get sensible answers from and give responses to him. ... A little tension across the line as you'd expect from any oversight committee and its organization. But good. I think a positive way.

And a smart guy, an ambitious guy? What kind of a guy?

Relaxed. I'd say ambitious but not ambitious in a negative way, just someone who's interested, who then moved to be the chief of intelligence in the White House on the NSC [National Security Council]. And actually, I didn't have a lot of experience with him when he was in that particular job.

[Tenet personally briefed President Bush each morning. What's the value of "face time" with the president?]

... I would say the opportunity to have some exchange with principal players -- the president and vice president, secretary of state, secretary of defense -- to be able to see how they react, to see how do they deal with the information, ... this is knowing your customer, trying to understand how your customer reacts to what kind of products are best. ... So that's part of it, and part of it is hearing back from them things that say, "I talked to the prime minister and he didn't say that, he said something else." ... So you get information back that then you can kind of layer into your own analysis and expand on and go back to analysts and say, "This is what the response was." ...

Now, you don't change everything just because of that, because the intelligence job is to give information that the intelligence community believes is important, as well as information that your customer believes is important. So you continue to do things. In some cases, you'll give them information even though they didn't really want it. In fact, maybe that's the key. Maybe if they say, "This runs counter to the way I think about the problem," then maybe you should spend more time on that problem, look at what they're saying, look at what you're saying, and see and kind of test your own conclusions and, if you're convinced of them, push them back on them. ...

And I do believe that you can get too close and become, if you will, partisan to the policy. You can be captured by the being close to the seat of power, and I think you have to be very careful ... that you don't get to the point where you nod and say yes, [but] you really mean, "No, I don't agree with you." I think you have to be independent enough and distant enough that you can say, "Mr. President, that's not the way we see it." Or, "Mr. Vice President or Mr. Secretary, that's not our view of it." And then follow up with why that's not our view. You can't become the buddy, I mean the kind of personal friend. I think you have to retain a distance.

And I believe the written word is ultimately the one you want to pass with some elaboration to it because the written word you've had time to look at the qualifiers, to look at how you phrased it, to run it by a variety of experts, to think about it more. When you're doing it verbally, I think you can be too quick to judgment, perhaps too cute in a comment, and those carry powerful weight. A little aside can be very important. So I think the personal contact is great, [but] I think you have to be very careful with it.

In your analysis of what happened in the run-up to Iraq, was George Tenet too close to the president in a personal way?

I don't know. I think perhaps, but what he told the president was what he believed and what the organization believed. But perhaps the words surrounding it were not well enough qualified. I don't know that, I wasn't there.

"Slam dunk"?

George is given, you know, to exuberance and enthusiasm, and he's an open, friendly -- that can lead you to say things that perhaps you, in hindsight, wish you wouldn't have said. ...

What did Tenet bring to the CIA?

Tenet brought enthusiasm and focus, I think, and put some new people into top levels of the organization. Spent a lot of time on the Directorate of Operations, probably far more than on the Directorate of Intelligence. That became his primary focus, and ultimately, the primary focus was terrorism itself.

I did a study for George a couple years ago, the year before he left, on the TTIC, Terrorist Threat Integration Center. ... In that process, I went to some of the evening meetings that George had on terrorism. He had one every evening, where he got the principal players together from around the community -- FBI, Homeland Security; both the operators and the analysts and the counterterrorism people -- and just kind of reviewed what was going on, who was doing what, what actions were being taken. ...

Every evening?

Yes, every evening, I believe, at 5:00. This was not just a "Tell me what the information is"; this is "What are we doing about this?" ... A very operational, directive kind of session, at least an hour. I went to three or four of them because I wanted to get a sense for how he was doing it. One of the things that worried me a great deal was when George left, it wasn't obvious to me who was going to perform that function at that level. ...

Did he [focus on] counterterrorism pre-9/11, or is that a thing that is born out of the planes hitting the buildings and all that follows?

I think it was before 9/11 because ... I was asked to come in for a review of the counterterrorism program with particular emphasis on bin Laden, and that was before 9/11. ... And in fact, George called me down after we did this to talk about it. So I would say he was engaged considerably before 9/11. ...

And given what you know about that world, [how do you think Tenet handled Rumsfeld and the vice president], the personal politics of it all? ... It seems to me just from outside that he was in the fight of his life with all these characters around this war?

... All directors have some concerns and problems with the secretary of defense. All of them have some concerns and problems with the secretary of state. The natural allies are the national security adviser and the head of CIA. ... Defense is a huge elephant on the scene, and it commands major resources, and it has an awful lot of clout. State Department, there's always struggle between chiefs of station and ambassadors, ... who's doing this and who's interfering with whom. And so there's always some tension across that line, but there's always tension with the secretary of defense. ...

Where are you when 9/11 happens? How do you hear about it, and what do you believe about this agency you love and what will happen to it?

Well, I was at home, North Carolina. I saw it happening on television, like most of the rest of the country. I mean, it's pretty clear right at the beginning that this was going to have major implications for intelligence. ...

I guess the one thing that worried me early on was it was quite clear it was going to be damaging to the organization, and very few people, in my view, understand what intelligence can and cannot do. There are problems it cannot solve, no matter how hard it works, how many people it has to collect and how strong its analytic effort is. There are some problems that are just not easily knowable.

Terrorism is a good example of an area where you need enormous amounts of small tidbits of information, and you're trying to link those together in a mosaic that is very hard to come by. This idea of [connecting the] dots always bothered me because it's as if the dots are out there and it's just a matter of recognizing. The problem is the dots: Looking at the dots without any of the connecting tissue, looking at the dots without understanding the context is kind of meaningless. Information is the key, and quite simply we did not have the information that would have allowed us, in my view, to have done much with this in advance of this problem.

... The vice president of the United States says on television on Sept. 16, "America needs to understand we're going to go to the dark side, we have to go to the dark side and do some things that may not be obvious." What does he mean by that, from your point of view?

I would say that is beginning to think about pre-emptive actions: How do you take an action before an event, not wait until it's happened and then go after the people and try to put them in jail. And I think terrorism by its very nature requires actions before the terrorist strike and requires you to do things in anticipation. And that's the dark side, because I think that means that you ... not only have to find individuals and kill them before they kill you, but you have to be aggressively going out to try to not only collect information, but to run operations aimed at disrupting it. So I think that's a different world than we had been in, in the Cold War and most of CIA's existence. ...

... The agency itself, just purely from the agency's perspective, was ready to go, ready to roll, right around those first few days.

It was, although it was suffering from the problem of not having provided the intelligence that allowed authorities to deal with it before the event. So it kind of is in two tracks: One is it's got to do something immediately because there's more [threats] out there; and it also had to deal with this problem of, how did it miss this, and what could be done to change that calculus? ...

[At some point before the war you are asked to review the intelligence. What happened? Did you get a phone call?]

No, it's fairly straightforward in that the secretary of defense actually wrote a letter to the director of central intelligence and said, "It would be useful to have a study, to have a review, of what intelligence said up to the point of the war and then use the ground truth to kind of test those assumptions and those judgments."

Why would he do that?

Good question, whether it was aimed at kind of showing that intelligence wasn't that good or aimed at genuinely trying to understand intelligence. And I wouldn't assume that it wasn't the latter. I mean, Rumsfeld was interested in intelligence. He had done a very interesting study on the ballistic missile estimate that had been done, so he was interested in kind of the truth, intelligence judgment versus the reality. ...

I got three other people that I knew who were retired and had experience, one in terrorism, one in military and one in the Middle East, senior analytic people out of the analyst side. We pulled together all the intelligence for the period roughly a year before the war began ... and reviewed all of the finished intelligence, all of the product up to the point of the war and divided it into a whole set of various categories. Because, as you might expect, the product was not divided into those categories to start off with, but you had: weapons of mass destruction; oil; stability -- what would happen post-Saddam; allies in surrounding countries' reaction; internal policing. You know, the whole set of religious discussions, the religious analysis of the Shi'a and the Kurds and the Sunni.

And so we divided it up into what amounted to 12 major categories of analysis and then went through each of those and provided an assessment of what we said, how competent we thought we were in saying it, how complete we were, how good we thought the analysis was, how rich we thought the information was. And that was the first report we did. ...

The second report was specifically aimed at the National Intelligence Estimate. In that report, we looked at the raw data that supported that estimate as well as the judgments. The ... first report, we did not go back and look at the raw information, we just looked at the finished product and made judgments on it. But on the National Intelligence Estimate, on that particular report, we looked at the supporting data. ...

We did a third report which was, having done the first two reports, what do they tell you about the strengths and weaknesses of intelligence analysis, the quality of the product, the kind of review? In other words ... is this a one-off problem or does this reflect some systemic problems in intelligence analysis? And then we did one other report ... which kind of said, given those three reports, here's some things that might be thought about [in] trying to improve the system. So that's the array of reporting. ...

The last two reports are unclassified so you can read those. The first two are classified. But we had a lot of information. And I don't think anybody has looked at the product and looked at the raw reporting on WMD at the level that we did.

[What did you find out about the yellowcake uranium allegation?]

See, yellowcake is all kind of in the noise from my perspective. It was one piece of information that was an input to the National Intelligence Estimate, not particularly regarded as a critical piece, far more critical pieces in there. ... There are thousands, literally thousands, of reports dealing with weapons of mass destruction and that's one of them. ... Looking at it from the analyst's point of view, this is one piece in a big story that fundamentally would not change the story hardly one bit.

There is a persuasive amount of information that leads you to the conclusion that there are weapons of mass destruction. There is very good information -- some of which turns out to be wrong, fabricated or not accurate -- on biological weapons which is quite persuasive. There's a lot on chemicals which is very persuasive. ...

The nuclear program was always seen as a program that was out in the distance. There was disagreement between State Department and CIA on how aggressively they were pursuing it. State believed less so, but they believed there was a program. CIA believed it was moving along apace. No one felt it was imminent, going to happen tomorrow. And yellowcake was not going to make it happen tomorrow.

There are other pieces of information that allow you to say there is a program ongoing. The information was quite persuasive, although a lot of it rested on judgments and information that came out of the first Gulf War and when we had inspectors and military on the ground. A lot of ... our analysis was extrapolation from that earlier period. And a lot of the pieces of information logically fit into that extrapolation and they got a regular flow of information that allowed them to say, "Yeah, they're doing this and that. They're trying to acquire this and that."

There are some serious flaws in both the collection and analysis. A lot of the collection during the two years leading up to the war of specific "who was doing what, where it was happening, what stage they were at on development of weapons," is very meager. Very thin. And a lot of that kind of analysis drew on earlier information that said, "Well, if they had had centrifuges, if they could do this 10 years ago, they probably have advanced here. We have little tidbits of information suggesting they had advanced." So a lot of that was based on not highly current or specific information, but kind of general information that programs were continuing. ...

The CIA believed that these weapons programs were continuing, and there's no question the administration, therefore, believed that as well. I don't think there was any effort, certainly no effort on the part of CIA, to hype that problem. What they did not do were some very important kinds of analytic testing. In the summer before I began the study, I was actually called in with another group [of outsiders] to look at the weapons of mass destruction program in Iraq. ... I felt the information was ... thin in terms of specifics that would allow you to provide high confidence about what you knew about the weapons program.

Thin in what sense?

It was not detailed enough. It was not specific enough in terms of who was doing what and where and what the status was. It was just kind of too thin.

This was the summer of 2002?

That would be summer of 2002. And I think that was the nature of the problem, that the analysts felt they had this continuum of information. They kept putting pieces on that continuum saying, "Aha, here's another piece." And if you go looking for information to prove a particular point or to establish that analysis, you often can find it, and it will reinforce your judgment. I mean, this is a real problem for analysis and for analysts.

In fact, some of the things they did not do is they really didn't ask themselves a couple very important questions -- What information was new over the past couple years? What was really new? What added to our knowledge? -- and kind of parse that information out and look at that as a separate body of information and say, "Alright, if you'd looked at that information over a three-year period prior to the war, would you be convinced that there were nuclear weapons programs or weapon programs under way?" Or would you say, "Boy, this is awfully thin stuff?" I think the answer is you'd say, "This is pretty thin stuff," if you took it away from the history and the past. ...

The other side of this [is] one of the things they did not do is look at what the impact of 10 years of sanctions would have on Iraq. Whether or not they really could sustain this kind of extensive program.

They never asked that question?

Really did not deal with that as a single question. They did pieces of it.

One of the real problems, and this is a problem for intelligence, not just in Iraq, ... is it does a lot of analysis, and it does a lot of reporting on particular topics. But it sometimes doesn't connect those together in a neat way for a reader or for your own understanding. It does them as kind of single, stand-alone pieces. And the analysts may have quite a good view of that perspective of that, but it's very difficult when you look at it, and we found this particularly true when we read all the information. ...

That's not the way the customer reads them, and that's not the way they're read by others, by the general consumers of intelligence. They're read here, put in the burn bag or put in the file, and then three weeks later there's another subject, two weeks later there's another subject and then another one and another one. And there is no kind of continuity of thinking.

I think that's a major problem in Iraq, is that we did not step back and kind of isolate our information and assess it independently. And we did not write in a way that allowed even our consumers to get the best out of the information we had.

[Are you talking about the type of information that made it into Colin Powell's speech at the United Nations?]

A lot of it. But, you know, it's thousands of documents and hundreds and hundreds of reports distilled down to, what, 30 minutes, 15, 20 pages of double-spaced, with 50 or 60 people at various times probably trying to get their word in, to have some comment on it or to say, "No, this should be emphasized and not that." I mean, I don't know how the process in that specific case worked, but it is not an easy process. Somebody's got to draft it; somebody then has to review it. Analysts then have limited opportunity to check it.

It's like a presidential speech. You may submit some words, but your opportunities to review the words that actually are put into the speech are very, very limited if they exist at all. That's just the way the process works. So it quickly gets out of your hands. So I don't know how this one worked, but I know how others have worked. It's a difficult process at best. ...

So the things [Secretary Powell] takes based on, as he says it, the personal assurances of Director Tenet, that he stands up to this stuff, is that the kind of thing you would have expected Tenet to do?

Yes, I think he was persuaded, comfortably persuaded, by the intelligence and by the analysis that they were right.

So he was not holding his breath or gilding the lily?

I doubt it very much. I mean, you always in this situation begin to hold your breath because you're worried about what's going to happen, how it's going to be presented. Are they going to take some things out of context and then I'm going to be asked about them and am I going to be able to then back that up? So that's always kind of an iffy situation, but I think George was probably quite confident about the larger set of judgments.

I think the analysts were confident about the larger set of judgments. If you then would have asked them, "Now, give me the other side of this. Where are our shortcomings, what don't we know, where are the limitations of our collection?" If you'd asked the intelligence analysts, they would have given you a kind of fairly long dissertation on that. And maybe you would have said, "Well, wait a minute. Let me ... try to put those two together and see whether I make sense out of it." So you might have come away with a little different perspective.

But after all, what you had was a policy-maker trying to sell ... a policy action. Not surprising that you would emphasize those things which sold the argument. I mean, how many times do people in any kind of argument display the weakest part of their argument? You give it strength. So it's not surprising.

So we sit there, and we watch the secretary of state say, "There are these aluminum tubes to be used for centrifuges," being very declarative. Given what you know about what went into making that assessment--

I think the CIA's view was that that was the fact; that was their best judgment. I know there were others who believe they were part of some rocket or some kind of propulsion, but I think the agency believed they could be used for that. And we know they'd had centrifuges. We know they'd built them, and they were present in the first Gulf War -- there were hundreds of them. It's not hard to make that connection and say these may be for centrifuges. That's not a stretch to do that. ...

And when somebody like Carl Ford [disputes that]?

... Carl Ford is not a scientist. Carl Ford doesn't know any more about aluminum tubes than I do, and so he's relying on somebody else. ... He's relying on some expertise somewhere, just like CIA was relying on its expertise somewhere. Not an unusual situation, to have the two principal National Labs, Lawrence Livermore and Los Alamos, disagree on a fundamental issue. And you end up having to take sides on which one do you believe? So you use your expertise to try to assess the two and come to the conclusion that differs. Very common. ...

Carl is a very good analyst, an old CIA analyst. ... If anything Carl should have taken stronger footnotes because it sounds like they believed more clearly what they were saying in the footnotes than the footnote reflected.

What did you find when you looked at Curveball, and should we have relied upon the assertions based on what Curveball had to say?

... Quite clearly in hindsight, [Curveball] should have been dealt with much more skeptically. ... There were a couple problems. One, we didn't have an alternative sourcing, and we didn't have a way to get to that source and talk to them directly. I don't think there was ever any direct communication, even though we tried. So it's one of those things, if you looked at it, if I showed it to you, was reporting to you today, you would say, "That's pretty sophisticated stuff." I mean, it's very hard to know how to deal with that if you have no corroboration. All you can do is say we wish we could have a greater contact with our sourcing, with our sources than we had. But, you know, I look at that and I can understand why they used that as part of their analysis.

And when I read James Risen's book, and it tells me that [Tyler Drumheller] was trying to get [then-Deputy CIA Director John] McLaughlin to say, "Wait a minute, this guy, unreliable." This is before the fact. What do I make of that?

I don't know. I know Risen's statement. I don't know what the real words ... to the analysts about the sources were. I know what they were after the fact. And that's one of the things we tried to do, is we didn't talk to analysts about what they wanted to say, or what they thought that they were saying. We used what they said as the basis for our judgment because I didn't want to get into this, "I didn't write it the way I thought it," or "It was slightly different."

All I know is I looked at the reporting, the reporting looked pretty good, the sourcing looked pretty good, and if you're an analyst, I would say it's not a bad judgment to have used it the way it was.

Did you ask or do you know whether Tenet knew it was a single source from foreign intelligence?

Oh, I'm sure he did. Yeah, it was pretty explicit. ... It's possible one could be confused and think there was more than one source involved, possible. But I thought it was pretty explicit.

You didn't ask Tenet the question?

No, no. As I said, we didn't ask anybody questions about what they thought about things after the fact. We didn't think that was a useful road to go down. ...

[Did you look at the intelligence from Ibn al-Sheikh al-Libi, who ran an Al Qaeda training camp, and who allegedly gave information under torture?]

The reporting was not used as a part of the CIA analysis. And we spent a fair amount of time looking at this subject because of the accusations that, in this area, there was a lot of pressure on the analysts to come up with a conclusion that in fact there was a connection between Al Qaeda and Saddam. And in fact, of course there is a connection. There were contacts between people, certainly between some of Saddam's intelligence people and the terrorists.

The information we saw, we could find no reason to believe that there was any joint operations or collaborations. Did they allow them transit or did they help them or did they know that somebody was in Iraq? The answer to that probably is yes. But was there a joint operational thing where they're cooperating? We didn't have any information that would support that argument.

[What about the allegation that Mohamed Atta met with Iraqi intelligence in Prague?]

There was always a question of whether that really happened. I think that's the analyst's judgment of that one. But again, I would go back and say everything that the analysts wrote about that was presented to the policy-maker said certainly Saddam is involved in terrorism. I mean, he harbored some major terrorists in the country for a number of years. He trained some of his own terrorists because they were very effective in killing people outside of Iraq, opposition leaders. Killing their own people inside the country. I mean, they ran a pretty effective program, and they did meet with Al Qaeda operatives. There's no evidence that we could find that showed there was any joint planning, joint activity, joint operational activity, which was kind of what we were looking for.

So did you watch Powell's speech on television?

I can't remember whether I did or not. I think I did, yes.

Remember what you thought when the evidence went by that he was laying out for the world?

No, I can't. I mean, I'll have to admit I don't remember. It didn't have a major impact on me. ...

And Tenet's sitting behind him? What did you think of that?

That's not surprising. I mean, I wouldn't draw all kinds of implications out of that other than we were trying to support a rather significant policy through the use of intelligence. We've done that before, never quite in this way, although yes, in many ways we've done that before. We've provided the intelligence, which was the basis for presentation to the U.N. or others. So I guess I wouldn't find it remarkable in that sense.

When you look at the evidence, you see that it's kind of thin but you could see th