Guest article
Reprint rights


Linking al Qaeda and Iraq

Stephen Hayes's
rush to closure


© 2003 WTM Enterprises. All rights reserved.


Stephen Hayes, a staff writer for the Weekly Standard, created a furor with his article "Case Closed" (November 24), which reports on raw intelligence data contained in a top-secret U.S. government memorandum. The intelligence material in the memo is organized into 50 numbered points and comes from several domestic intelligence and security agencies. Along with raw data, the memo includes expert commentary and analysis.

The Hayes article specifically addresses alleged Iraqi dealings with al Qaeda from 1990 through 2003 and the question of how the secret government memo affects our understanding of the relationship.

For the most part Hayes's presentation of the memo is professional and helpful. His limited analysis, while often one-sided and incomplete, is also interesting at times. However, the article's conclusion represents a radical and unexpected departure from the measured tone and cautious treatment of the memo that pervades the entire article. The Hayes conclusion — as it shall be referred to henceforth — is the subject of this essay.

First, please take note of what Hayes does not conclude. He does not conclude that the available evidence supports, let alone that it proves, Iraqi complicity in any of the known al Qaeda terrorist operations against the United States, including the attacks of 9/11. It is precisely the lack of hard evidence for Iraqi involvement in such widely known attacks that forces Hayes to propose official Iraqi involvement with al Qaeda in some vague effort to "plot against Americans." Thus, Hayes concludes: "There can no longer be any serious argument about whether Saddam Hussein's Iraq worked with Osama bin Laden and al Qaeda to plot against Americans."

A few preliminary remarks are advisable before we forge ahead with the main purpose of this essay. The Hayes conclusion is strong and unequivocal, both in its essence and implications — perhaps too strong, no matter the evidence in its favor.

One is left with the impression that Hayes thinks he is arguing to his conclusion deductively from self-evident premises or is drawing inferences from scientific data under controlled experimental conditions. If that were the case such certitude might be warranted. Instead, Hayes is analyzing data from the field of government intelligence — a world of intrigue, hidden agendas, double agents, and deliberate misinformation — in short, a world of unconscionable skullduggery. Moreover, he asserts his conclusion without explicitly clarifying its evidential support or elucidating its underlying logic, further suggesting a misplaced level of certitude. Hayes also arrives at his conclusion on the sole basis of information gleaned from interrogations and individual testimony; he presents no documentary or physical evidence in support of his conclusion, which again suggests an inappropriate certitude.

Intellectual caution, a rational skepticism, and a circumspect approach are in order, combined with a willingness to constantly revise one's conclusions in light of new discoveries. Rarely is a student of the human sciences justified in proclaiming certainty, especially at such an early juncture in our collective knowledge.

My interest is in evaluating the Hayes conclusion exclusively in light of the information and commentary provided in "Case Closed." Furthermore, since many have read "Case Closed" as a sort of retrospective justification for the Iraq war, my focus can be further narrowed to assessing the Hayes conclusion against intelligence referring to events prior to September 2002, for by that time it was clear the United States was preparing for war.

It is not always clear which data Hayes had in mind when drawing his conclusion. Thus, our first task is to identify information adduceable in support of the Hayes conclusion. Given the unequivocal nature of the Hayes conclusion and the limited purpose of this essay, we are searching for unambiguous and detailed evidence that directly implicates Saddam and al Qaeda as being in league against U.S. interests, prior to September 2002.

Next, we must subject that data to critical analysis with the goal of validating or invalidating it. For the Hayes conclusion to stand, it must be shown that the potentially supporting information is more likely true than false, or that the source in question is essentially reliable. Conversely, to cast doubt on the Hayes conclusion it must be shown that the information under scrutiny is probably false or the source unreliable.

In conducting our inquiry we shall adopt several non-controversial principles as aids to evaluating source reliability. Unnamed sources, unadorned sources, and sources without the proper credentials will downgrade source reliability. Named sources and sources identified as having "good" credentials will enhance source reliability. Uncorroborated testimony will be considered less reliable than independently verified source reporting.

We shall always give precedence to the expert analysis provided by the U.S. intelligence community against the raw intelligence contained in the article. We shall make a modest effort to reconcile apparent contradictions when we have good reason to prefer one line of evidence to another line of discordant evidence. Finally, the more detailed the testimony of a given source, the better — ill-defined information will serve to diminish source reliability. We will always employ good old common sense.


A careful reading of the article reveals two potential sources of direct support for the Hayes conclusion. Both sources provide references to al Qaeda, Iraq, sabotage or attack, and American interests in the sense required by the Hayes conclusion.

The remaining data, while worrisome if true, nevertheless fails to provide convincing support for the Hayes conclusion. Indeed, this entire corpus of information is open to a variety of coherent and plausible interpretations, depending on one's presuppositions and preferred reading of the data. Even if assumed true, only supposition and conjecture can bridge the evidential gap separating this body of data from the Hayes conclusion.

The first support for the Hayes conclusion comes from intelligence dealing with 1998. Hayes presents this information in numbered points 15-18, with subsequent expert analysis of the data. The potentially incriminating point (17) reads:

Iraq sent an intelligence officer to Afghanistan to seek closer ties to bin Laden and the Taliban in late 1998. The source reported that the Iraqi regime was trying to broaden its cooperation with al Qaeda. Iraq was looking to recruit Muslim "elements" to sabotage U.S. and U.K. interests.

If true, this source has just provided damning evidence for Iraqi complicity in a vile, if abortive, plot against American interests. The question, then, is how reliable is the source? The apparently damaging information is easily undermined on grounds provided in the article itself. It is important to note that nothing ever resulted from the effort to recruit Muslim "elements." At best we are considering an attempted attack, not an actual attack. The information is uncorroborated. The source is unnamed. There is no independent verification available. All of that does not bode well for the source's reliability.

The reliability of the source is never established on independent grounds or even asserted in the article itself. For example, the memo often refers to sources as "well-placed" or as "regular and reliable" or as presenting "the most credible information." In this instance we simply get a "source" and are asked to take him at his word. The claim made by the "source" is vague, suggesting he lacked access to the relevant information. No specific details are provided by the source. The most important basis for skepticism, however, rests on the analysis of this and related points provided in the memo itself. The presumably authoritative analysis reads:

Reporting ... from different sources, corroborate each other and provide confirmation of meetings between al Qaeda operatives and Iraqi intelligence in Afghanistan and Pakistan. None of the reports have information on operational details or the purpose of such meetings. The covert nature of the relationship would indicate strict compartmentation [sic] of operations.

Please read that expert analysis carefully in light of Hayes's unequivocal claim for official Iraqi complicity in attacking American interests and my own brief and admittedly amateurish analysis above. The expert analysis clearly concludes on an unsupportive note for the thesis that Iraq and al Qaeda actively conspired against American interests: None of the reports have information on operational details or the purpose of such meetings.

That expert commentary seemingly demolishes the reliability of the source in question. If he were reliable, why not highlight and expand upon his alarming claims in the analysis section of the memo? Instead, the experts tell us the operational details are shrouded in mystery, including supposed efforts to recruit Muslim "elements" to sabotage American interests. That is the studied conclusion of the American intelligence community.

The second supposed evidential support for the Hayes conclusion derives from intelligence reporting for the period from July 1999 to 2003. Commenting on that intelligence, Hayes writes: "Intelligence reports about the nature of the relationship between Iraq and al Qaeda from mid 1999 through 2003 are conflicting." Thus, Hayes admits there may be room for rational doubt concerning the evidence from the mid 1990s to 2003. Why is that the case?

It turns out that the final bit of information purported to implicate Iraq in a conspiracy against the United States presupposes an active relationship between al Qaeda and Iraq beyond July 1999. Interestingly, a close reading of the article indicates one is within his epistemic rights to basically dismiss much if not all of the post-July 1999 reporting, on the basis of testimony provided by a senior Iraqi intelligence officer in U.S. custody, Khalil Ibrahim Abdallah. Abdallah insists that Saddam terminated all links with bin Laden in July 1999, calling into question all post-July 1999 intelligence reporting that suggests official contact between Iraqi and al Qaeda agents. Thus, to take Abdallah's testimony seriously is to cast doubt on allegations of Iraqi and al Qaeda efforts after July 1999 to plot against American interests. Is there good reason to take Abdallah's reporting seriously?

In contrast to much of the reporting in Hayes's article, Abdallah is named and is identified as once having served in a sensitive position with access to the relevant facts; and he has little incentive to lie. He is also potentially available for future interviews by independent researchers. Finally, his testimony deals in specific dates and provides some details open to independent verification. A strong case can be made for Abdallah's credentials and reliability.

It would not be an overstatement to point out that the Hayes conclusion stands or falls with Abdallah's testimony. Does Hayes refute Abdallah? Unfortunately, Hayes never really engages or evaluates the veracity and implications of Abdallah's testimony. Instead, he simply dismisses that testimony, on dubious grounds, when he writes, "The bulk of the reporting on the relationship contradicts this claim." Hayes's reason for dismissing Abdallah is open to attack on at least two fronts. We will argue that the available data supports the reliability of the Abdallah reporting against the post-July 1999 reporting that is supposed to contradict it. We will also argue that Hayes may have confused an apparent contradiction with an actual contradiction.

While it is often intellectually defensible to prefer one body of testimony — data A — to another line of apparently contradictory testimony — data B — given the larger number of reports in favor of data A, that is not obviously the case when (1) we have good and independent reasons to prefer data B to data A; (2) the relative number of reports in favor of data A is underwhelming; and (3) the veracity of data A is open to serious challenges on independent grounds. Regarding the Abdallah testimony, Hayes never considers those objections, let alone deal with them adequately. A sound argument can be made that all three conditions are satisfied in the case of Abdallah and the supposed data that contradicts his testimony. We have already argued for his basic reliability as a witness; a careful reading of the Hayes article reveals the number of apparent claims against Abdallah's testimony to be unimpressive; and, as we shall see below, good reasons exist for rejecting much of the post-July 1999 intelligence reporting.

Thus, Hayes seems to beg the question, for what is at issue is the reliability of Abdallah's testimony compared with the few and allegedly contradictory testimonies so vital to the Hayes conclusion. Hayes fails to address the issue and ultimately fails to provide an intellectually satisfying reason for preferring the apparently contradictory data supportive of his conclusion over the testimony of Abdallah — reporting that directly clashes with his conclusion. With a wave of the hand Hayes dismisses seemingly reliable data detrimental to his theory in favor of unsubstantiated and inconclusive reporting, conveniently substituting assertion for argument. That move is crucial to the Hayes conclusion, making his rather cavalier dismissal of Abdallah's testimony seem self-serving.

Even if we grant the mere identification of a contradiction supported by "the bulk of the reporting" as a sufficient and rational basis for adopting a line of data congenial to one's conclusion, the question remains: is there an actual contradiction? Hayes never really argues for a contradiction; he simply asserts the existence of one. Actually, it is not clear whether Abdallah's testimony is contradicted by later (post-1999) reporting. All of the additional 1999 reporting is easily reconciled with the information provided by Abdallah. Indeed, with some intellectual effort almost all of the later (2000-2003) information can be harmonized with Abdallah's reporting as well. That includes the "Shakir" information and Iraqi intelligence's alleged meetings in the spring of 2001 with 9/11 ringleader Mohamed Atta. There is one glaring exception — information that does not seem so easy to harmonize — which we shall analyze later.

Space and time constraints require that it be left to the reader's imagination to reconcile all of the apparently contradictory information with Abdallah's claim that Saddam terminated Iraqi involvement with al Qaeda in July 1999. Suffice it to say, numerous possibilities exist — some more plausible than others. If we keep in mind important distinctions and note the use of ambiguous words and phrases we can reconcile all of the 1999 reporting with Abdallah's testimony. Here are a few tips for a possible harmonization regarding all of the additional (2000-2003) data: read Abdallah's testimony as "official" and then think in terms of "double agents," "rogue agents," and "unofficial exchanges" for some of the apparently contradictory information.

In sum, it is not clear that the mere existence of a contradiction combined with relative reporting "bulk" always provides a sufficient basis for preferring one line of data to another line of data, particularly when the "bulk" in question is actually unimpressive in terms of numbers of reports. That observation is underscored when good reasons can be adduced in support of the rejected line of data — in this case the testimony of Abdallah. Furthermore, Hayes never really proves that a contradiction exists. Thus, the real possibility of a successful attempt at comprehensive harmonization or even a limited reconciliation further undermines the reason given by Hayes for summarily rejecting Abdallah's information — potentially vindicating Abdallah's reporting and providing yet further solid ground for taking the implications of his testimony seriously.


Putting aside the question of formal contradiction, there exist four independent and apparently contradictory lines of data requiring further investigation.

First and second. The first line deals with a certain Mr. Shakir and his alleged involvement with some of the 9/11 highjackers during 2000. The second deals with 9/11 ringleader Mohamed Atta and his alleged meetings with Iraqi intelligence officials in Prague in 2001. Even if we grant the contradictory nature of this information, no extensive analysis is needed to discount this data: Hayes does a fine job undermining both examples.

Hayes never definitively identifies a conspiracy between Iraq and al Qaeda for 9/11. But the alleged activities of Shakir and Atta are intimately linked with 9/11. By not fingering an Iraqi connection for 9/11, Hayes tacitly acknowledges the lack of hard evidence for official Iraqi involvement with Shakir and Mohamed Atta — unless another conspiracy was afoot for which absolutely no evidence exists.

And Hayes explicitly acknowledges the tenuous character of the evidence connecting Shakir, al Qaeda, and Iraqi intelligence when he writes, "Was Shakir an Iraqi agent? Does he provide a connection between Saddam Hussein and September 11? We don't know. We may someday find out." Those words speak for themselves.

What about Mohamed Atta and his alleged connection with Iraqi intelligence? First, it should be noted that even if proven to have occurred, an Atta meeting with Iraqi intelligence in Prague is not self-interpreting. As alluded to above, a variety of plausible explanations are possible, some quite prosaic.

In any event, once again Hayes does our work for us, presenting just enough conflicting evidence from the CIA, FBI, and "some of the most hawkish Bush administration officials" to cast serious doubt on the alleged Iraqi-Atta connection. An actual money trail connecting Atta and Iraqi intelligence would have provided real evidence for an Atta connection. Alas, Mr Hayes acknowledges the non-existence of such hard evidence. Moreover, Hayes refuses to say, definitively, that Atta actively conspired with Iraqi intelligence. Something is holding him back. Perhaps it is the lack of hard evidence and the multitudinous interpretive possibilities suggested by the data.

Third. The third example at odds with the testimony of Abdallah concerns an Abu Musab Zarqawi, an alleged associate of al Qaeda. This data, whatever its evidential merit, deals with events after September 2002; it is not relevant to our discussion, given the stated purpose of this paper. If proven true, the data suggests an agreement between Iraq and al Qaeda to defend Baghdad when a U.S. invasion appeared imminent. That provides a less-than-sturdy justification for the United States to initiate the war. We can safely ignore this data for the moment.

Fourth. Now we come to our final piece of evidence, evidence that directly contradicts the statement of Abdallah and, more ominously, points to an official Iraqi plot to attack American targets in Saudi Arabia. This data is the "glaring exception" referred to above. Before considering this evidence, let's survey the road we've traveled thus far.

Abdallah, a high-ranking Iraqi intelligence officer in U.S. captivity, stated that Saddam severed all official relations with al Qaeda in July of 1999. We have found no solid reason to doubt that and all that it implies. With some imagination all additional 1999 data and perhaps the bulk of the additional intelligence reporting is readily harmonized with Abdallah's testimony — calling into question the reason for rejecting Abdallah's testimony in the first place.

The evidence for an official Shakir connection has proven insubstantial, and the Atta evidence inconclusive at best. It seems the testimony of Abdallah has survived a number of tests and enjoys considerable reliability. The burden of proof in challenging Abdallah is heavy.

Does the final piece of data we shall consider rise to that challenge, implicating Saddam in a post-1999 conspiracy against U.S. interests?

The data in question is from point 27 and reads:

According to sensitive CIA reporting ... the Saudi National Guard went on a Kingdom-wide state of alert in late Dec. 2000 after learning Saddam agreed to assist al Qaeda in attacking U.S./U.K. interests in Saudi Arabia. [Ellipsis in original.]

Like the previously alleged attack, this one never occurred; we are dealing with a purported plot to attack U.S. interests. Why couldn't Saddam ever pull off these alleged attacks?

Yet again we are dealing with no clearly identified source in a position of knowledge, a troubling lack of detail, and zero independent corroboration. All of those legitimate concerns are made more acute when we remember that Saudi Arabia is a U.S. ally. In light of that, there should be solid evidence for the alleged conspiracy available to U.S. intelligence. Where is that evidence? This data raises more questions than it answers.

The total lack of independent verification for the information is perplexing. It should be easy to provide independent and objective evidence for the alleged Saudi mobilization of December 2000, especially given Washington's close relations with the Saudis. The lack of corroboration is curious indeed.

The information also seems to beg the question. After all, how did the Saudis learn that Saddam had agreed to assist al Qaeda in attacking U.S. interests? That is the crucial question that demands a comprehensive answer. Was their information and analysis reliable? Naked assertion is no substitute for rational argument and direct evidence.

We should not overlook the conspicuous information gap in the above reporting. How much information is camouflaged by that ellipsis: according to sensitive CIA reporting ... the Saudi National Guard....? What is the nature of the information? Would inclusion of the relevant information strengthen, weaken, or prove neutral regarding this line of data? — a good question with no obvious answer.

What are we to make of the absence of expert analysis regarding this alarming information? Why do the experts always fail to even suggest that an Iraqi conspiracy was in the works against the United States? If the intelligence community had good information and really believed in the data — especially given its troubling character — doesn't it stand to reason that they would at least provide some analysis, if not draw some conclusions? Their silence betrays a lack of trust in the data, perhaps for some of the reasons detailed above.

There is a final point worth noting with respect to the purported Saudi alert. December 2000 was less than a year before 9/11. Why would al Qaeda attack the United States in Saudi Arabia, potentially jeopardizing its massive and deadly 9/11 terrorist attacks? On a related note, what kind of operation is al Qaeda? Can it really plan, organize, and implement multiple terrorist operations on a global scale — including some of unprecedented evil and magnitude — simultaneously? That is not obviously the case. Once again a legitimate question finds no answer, not even an attempted answer.

We have before us a line of data full of holes that raises more questions than it answers, a line of data that conflicts with the emphatic words of a seemingly reliable source, Abdallah. In light of the above analysis and when placed in the context of Abdallah's weighty testimony, a reasonable man cannot be faulted for his skepticism regarding the data and its claim that Saddam was operating in union with al Qaeda, plotting against the United States in December 2000.


We set for ourselves a well-defined goal: the identification and critical analysis of direct evidence for Iraqi complicity with al Qaeda against U.S. interests, prior to September 2002. We have limited ourselves to the intelligence reporting, commentary and analysis, and background information provided in Stephen Hayes's article "Case Closed."

Our efforts produced two putative sources of incriminating evidence in favor of conspiracy. That intelligence data was evaluated and graded according to a set of explicit principles and common sense. We reasoned that for the Hayes conclusion to be proven true, the data or sources in question should be considered more likely reliable than unreliable. We also reasoned that for the Hayes conclusion to be proven false, the opposite condition should obtain. Actually, the certitude of the Hayes conclusion probably calls for a more stringent standard of proof. But we shall ignore that observation in drawing any conclusions.

The intelligence reporting from 1998 suffers from numerous rational challenges — from a lack of independent verification, to the problem of an unidentified source, to an unacceptable vagueness. More important, the expert analysis and commentary internal to the memo fails to substantiate the data in question and actually directly undermines the information with an explicit acknowledgement that specific operational details are lacking — the opposite of what is contained in the reporting under consideration. Skepticism seems to be in order regarding this particular data.

The intelligence reporting from December 2000, concerning Saudi Arabia, also evinces numerous problems. Once again, the conspicuous lack of independent verification is problematic, as is the lack of expert analysis — a curious omission given the disquieting nature of the reporting. The motivation and capability to pull off such an attack is less than obvious, especially when it is recalled that 9/11 was less than a year off. By his own words — and lack of words — Hayes renders the Atta connection debatable and the Shakir connection inconclusive. The information provided by Abdallah appears basically reliable and has survived a number of challenges. While the Abdallah testimony lacks independent support, we can find no solid reason to doubt it. But it does seem reasonable, following this round of analysis, to doubt Hayes's assessments.


Three additional points bear mentioning.

• Hayes provides some illuminating information from 1993 showing the plausibility of an Iraqi-al Qaeda alliance, despite the bitter ideological differences between Saddam and bin Laden. This 1993 reporting is real evidence against the speculation of some that Saddam and bin Laden would never cooperate owing to their ideological enmities. To the contrary, the Hayes conclusion posits an active conspiracy at least through 2000. Was the bin Laden agreement with Saddam durable and trustworthy enough to last from 1993 through 2000, especially given their widely acknowledged distrust of one another? Hayes has done good work in this area, but more work needs to be done for the sake of his conclusion.

• It is safe to assume that a sophisticated global conspiracy requires a sufficient motive. While Hayes provides some insight into an Iraqi motive for "bullying" and "cajoling" fellow Arab states to break the embargo against Iraq, he fails to fully describe an adequate motive for official Iraqi cooperation with al Qaeda in attacking U.S. interests.

• Some meaningful historical and contemporary context would have been most helpful in evaluating the Hayes conclusion. For example, how does the intelligence reporting concerning Iraq and al Qaeda compare with similar reporting for other states and al Qaeda? Surely other states have authorized contacts with al Qaeda. How might the pattern, frequency, and quality of those hypothetical contacts compare with the Iraqi contacts? We would have been grateful had we been offered a meaningful standard for comparison.


Unless we adopt a naive reading of the Hayes article and simply dismiss the information provided by Abdallah, the Hayes conclusion, even a diluted version of it, cannot stand. Perhaps future discoveries will bolster the Hayes conclusion. Hayes does in fact promise future revelations in support of his conclusion. Unfortunately — and based solely on information presented in "Case Closed" — this tentative restatement of the Hayes conclusion seems appropriate:

Evidence points to Iraqi contacts with al Qaeda over much of the 1990s. Some of those contacts may have included limited training, perhaps in bomb-making, and offers of exile for some al Qaeda operatives — or at least consideration of such offers. The real purpose of the contacts remains murky. It is not clear whether official Iraqi contacts with al Qaeda persisted through July 1999; reliable but uncorroborated testimony suggests all official ties were severed in July 1999; however, some intelligence reporting suggests contacts through 2003, although that is difficult to substantiate. Perhaps unofficial contacts persisted through 2003. Fragmentary and unverified reporting suggests two failed Iraqi attempts to plot against U.S./U.K interests in alliance with al Qaeda. At present no specific conclusion is warranted regarding this claim.

Incidentally, that conclusion provides no morally legitimate basis for war — at least not from a Catholic or even traditional just-war view of things.

Perhaps if we had access to the secret government memo exploited in the article or, better yet, all of the relevant data from all sources — physical, documentary, radio-intercept, testimonial, and satellite — we could provide a coherent and damning picture of Iraqi involvement with al Qaeda. Hayes has not been forthcoming with the memo and enjoys a decided advantage against his critics — hardly a mark of serious scholarship aimed at truth. Then again, access to all the data may be the worst thing possible for the Hayes conclusion. Either way, our modified conclusion is the best we can offer Hayes, given his article and the memo excerpts thus provided.

December 18, 2003

Joseph Audie is a Ph.D. candidate in biophysics and a believer in the perennial philosophy, best propounded by the Dumb Ox.

Further reading.Several replies to the Hayes article are available on line:

"Case Decidedly Not Closed," Newsweek, November 20, posted at MSNBC.
"The dubious link between Iraq and al Qaeda" by Josh Marshall, The Hill, November 19.
"New Leak Smells of Neocon Desperation" by Jim Lobe, Antiwar.com, November 20.
"Together Again for the First Time: The perpetual hunt for an Osama-Saddam Entente" by Jesse Walker, Reason Online, November 19.
"The Case of the Misunderstood Memo" by Daniel Benjamin, Slate, December 9.

Hayes has also responded to some of his critics; see the Weekly Standard site.

© 2003 WTM Enterprises. All rights reserved.

Notice  to visitors who came straight to this document from off site: You are deep in The Last Ditch. You should check out our home page and table of contents.