January 27, 2012 / by emptywheel


Latif and the Misattribution Problem: “All Arabs Look the Same”

I’m going to have at least two more posts on Adnan Farhan Abd Al Latif (for background see this post), both in response to this post from Mark Denbeaux and Ben Wittes’s response to it.

In this post, I want to demonstrate a possible mistaken assumption many of us have been making as we try to read through the redactions, which is that the only source of potential irregularity in the document at the heart of Latif’s habeas case may have arisen out of translation problems.

Denbeaux describes what he believes the circumstances of the report at the heart of the Latif case to be.

To illuminate how the presumption works, the majority utilizes a hypothetical that does not properly apply to Latif’s case. The hypothetical depicts a government intelligence officer taking the statement of a third party informant. The majority would have us presume that the officer accurately wrote down what the third party informant said, though not presuming the informant’s statement was itself true. This seems to make sense until you apply it to the facts of Latif. A fair and thorough reading of the opinion suggests that the document and information being redacted is a report from an interrogation of Latif that contains opponent-party admissions. The interrogation likely involved an interrogator, a translator, and Latif. Thus, the third party informant in the majority’s hypothetical is Latif himself.

But that’s not exactly what Janice Rogers Brown wrote in the majority opinion in this case. In addition to requiring us to presume that the officer accurately wrote down what the third party informant said, she also wants us to presume that the government officer accurately identified the source.

The confusion stems from the fact that intelligence reports involve two distinct actors-the non-government source and the government official who summarizes ( or transcribes) the source’s statement. The presumption of regularity pertains only to the second: it presumes the government official accurately identified the source and accurately summarized his statement, but it implies nothing about the truth of the underlying non-government source’s statement. [my emphasis]

That’s an important distinction because there are hints that misattribution might be a significant issue in this case as well.

First, the government reply to the Circuit Court tries to refute just that possibility along with mistranslation. “Those similarities – which square with the external evidence about [redacted] make it highly unlikely the report resulted from a mistranslation or misattribution” (PDF 8)

Then, even going back to Latif’s CSRT in 2004–at which the allegations he fought at Kabul, the same allegations at issue here, were presented–he insisted he was not the person referred to in the unclassified summary of charges against him.

I told you I wasn’t the person they were referring to. I never went to the places that you said I did. I am not the person this case is based on.

Also remember that the government is not relying on a discrete, self-contained report on Latif alone. Rather, it has presented just fragments of a larger report, as David Tatel noted in his dissent.

The Report’s heavy redactions–portions of only [redacted] out of [redacted] pages are unredacted–make evaluating its reliability more difficult. The unredacted portions nowhere reveal whether the same person [one and a half lines redacted] or whether someone else performed each of these tasks. And because all the other [redacted] in the Report are redacted, the district court was unable to evaluate the accuracy of [redacted] by inquiring into the accuracy of the Report’s [redacted].

That’s important because several of the intelligence reports reporting on detainees Pakistan turned over to the US in December 2001 are group reports (I’ve determined this by searching on the report name among Gitmo files). TD-314/00684-02, which I suspect is the report in question, includes reports from at least 9 detainees. TD-314/00685-02 (obviously, a closely related report) refers to at least 7 detainees. Another, TD-314/00845-02, catalogs the transfer of at least 8 detainees (a number of whom are also mentioned in TD-314/00684-02)  from Pakistani to American custody. And IIR 7 739 3396 02 lists 84 detainees purportedly captured with Ibn Sheikh al-Libi. That is, even if I’m incorrect in my supposition that TD-314/00684-02 is the report in question, chances are quite good that the report deals with multiple detainees in the same report and the redactions Tatel describes serve to hide the other detainee stories told in the same report.

Furthermore, the internal distinctions between detainees in these reports do not appear to be clear cut. Often, references to these documents also include a paragraph letter (as in footnote 4 of Abdul Rahman Owaid Mohammad al Juaid‘s Gitmo file, referring to paragraph L of TD-314/00684-02 ). If it is so difficult to find information relating to a particular detainee in these reports, it increases the chances the reporting on them might get muddled in the process.

With that in mind, consider how Latif’s story compares with two of the other detainees included in TD-314/00684-02.

Judge Henry Kennedy’s opinion granting Latif’s habeas petition, for example, describes how the government pointed to Latif’s intake form describing that he went to Kabul “just to look around” as evidence of his changing story.

… they point to the intake form completed when Latif came into United States custody, which reports that Latif” went to [Pakistan] for treatment of an ear problem,” and that Latif “was in Kabul,” where he “claims he went just to look around about 4-5 months ago” and where he “stayed [for] 3-4 months.”

The reference is to the intake form (PDF 34). In a subsequent post, I’ll show that the government has not, itself, given this document the presumption of regularity, though that hasn’t stopped the government from pointing to the “just look around” explanation as proof that Latif’s story has been inconsistent. But I also find it striking that the story–placing a Yemeni detainee in Kabul “just to look around”–is similar to one attributed to another of the detainees whose Gitmo files depends on TD-314/00684-02, Sharaf Ahmad Muhammad Masud (though the timing is different). Masud’s story is attributed to TD-314/00684-02 and 3 subsequent reports (so Masud reportedly kept telling for some months after he arrived in Gitmo). Masud appears to have been caught by the Pakistanis in the same group as Latif was, and according to their two Gitmo files, he was formally turned over to US custody in the same place, Kohat, Pakistan, on the same day, on December 30, 2001, Latif was (though other documents say Latif was turned over on December 31).

Or consider the claims about Latif’s admission of ties to the Taliban. Significant parts of that material is redacted from the public filings in the case. But his Gitmo file claims that TD-314/00684-02 recorded him admitting that, after he traveled from Karachi to Quetta to Kandahar to Kabul,

Detainee admitted receiving weapons training from the Taliban and then fighting in support of the Taliban on the front lines.


Detainee admitted after training he was sent to the front lines north of Kabul. Detainee remained there until the Taliban retreated and Kabul fell to the Northern Alliance.

That closely resembles the story attributed to al Juaid, who is alleged to have explained that,

From Karachi, detainee traveled to Quetta, PK and Kandahar, AF.6 Detainee then continued to Kabul. In approximately August 2001, detainee was assigned to the rear lines of the Northern front opposite the Northem Alliance (NA). Detainee remained in this position until the Taliban withdrew under heavy coalition bombardment on approximately 20 November 2001.

Now these stories–minimal training and a few months of fighting on the back lines in Kabul–are sourced exclusively to TD-314/00684-02 for both Latif and al Juaid. With no citation, both reports claim the man was part of Osama bin Laden’s 55th Arab Brigade; both reports assess (with no citation) the man fought briefly at Tora Bora, and traveled from there, with Ibn Sheikh al-Libi, to the Pakistani border. Both were rounded up, reported to be involved in the bus revolt, and turned over at Kohat. Al Juaid was turned over to the Americans in Kohat one day after Latif was, on December 31 (though other documents say Latif was turned over to the US on the same day–on December 31). The stories tying them to the Taliban are almost identical.

Significantly, al Juaid was transferred to Saudi custody in 2007, even though there is, if anything, more corroborating evidence (though not much) of ties to al Qaeda than there is with Latif. Al Juaid was held for six years based on almost exactly the story as Latif, like him, sourced almost exclusively to this one document, but the government was willing to transfer him into Saudi custody years ago.

Now I suspect the government would argue the similarities in the stories told–only in TD-314/00684-02, of course–prove that there was a pattern of behavior here. But the opposite is just as likely: that for Arabs detainees who had ever been in Kabul, the story reported in this one report with obvious flaws was assumed to have been the same, training (even if there was neither time nor evidence in the detainees’ known chronology for such training), fighting at the back lines, and then departure after the US started bombing Kabul. With a side trip to Tora Bora, regardless of the lack of any witnesses corroborating such a side trip, included for good measure.

In other words, it seems that the detainees whose purported confessions are reported in TD-314/00684-02 may have gotten muddled in the process.

The redactions in all the filings in the case are too extensive to be sure misattribution is one of the potential problems with this report. But it sure seems like American and Pakistani interrogators were having a tough time keeping their Arab captives straight. So I think it possible, at least, that part of the problem with the report used to hold Latif is that it told of jumble of poorly distinguished stories.

Copyright © 2012 emptywheel. All rights reserved.
Originally Posted @ https://www.emptywheel.net/2012/01/27/latif-and-the-misattribution-problem-all-arabs-look-the-same/