David H. Petraeus, former director of the CIA. To paraphrase Howard Baker: “What could the director be expected to know, and when could he be expected to know it?”
With the election behind us and David Petraeus having testified in closed House and Senate hearings, we may hope for a more measured and less emotional examination of the events in Benghazi, Libya, on Sept. 11, 2012. In a previous post, I looked at some of the background behind the issue of post security. In this post, rather than rehearsing the controversies that have raged in recent weeks, I would like to highlight some of the common pitfalls that we encounter when we try to judge large, unexpected events during and shortly after the fact. The controversies that so often arise may reflect a combination of the nature of such events, the ambiguities of the intelligence assessment process, and the pressures of politics.
In the case of Benghazi, the full story is still not known—and it may never be—yet many people have made up their minds that they already “know” what happened. More fundamentally, people assume that “someone,” especially the CIA, must have known what was happening at the time and is not telling. The fact is, however, analysts are still piecing together the evidence. They certainly did not know the full story as it was happening, and it takes a considerable amount of time after the fact to reconstruct the most likely version of what probably happened.
The most important thing to remember is that confusion is common in sudden, large, unexpected events, whether they be surprise attacks or natural disasters. The first reports from the scene are often incorrect because the early evidence is fragmentary, contradictory, and sometimes erroneous. Afterward, when we already have a good idea of what happened, we can look at the same reports and say, “Obviously, these clues, and not those, were the ones they should have been focusing on.” At the time, however, that judgment is not so easily made. Moreover, early analysis will be influenced by the existing theories or preconceived notions of the analysts; sometimes those theories will be correct, or close enough not to matter, but in other cases they will be out of date or even completely wrong. Either way, that initial analysis will then set the framework for the next wave of analysis and evaluation. If the initial assessment is off, the accumulation of information will eventually bring analysts around to the right direction, but that correction may be delayed to the extent that the incoming flow of information continues to be slow, partial, and contradictory. Information of that sort can be subconsciously dismissed or interpreted to fit into existing frameworks for a considerable period of time.
Analysis is also subject to political cross-pressures. Analysts fear they will be held accountable for mistakes, so their preference is to wait, to take the time necessary to be sure. Because they want to avoid outright errors, and because the situation is by its very nature vague, confusing, and imprecise, their reports will also tend to be hedged, vague, and imprecise—much to the vexation of political leaders. Political leaders and intelligence analysts often have different understandings of the nature of intelligence. Political leaders want to receive precisely the opposite of what analysts want to give. They want details, and they want them to be clear, definitive, unambiguous, and correct. They want the entire story immediately, and they do not want to hear that the story has evolved since they first heard it. Leaders have invested enormous quantities of tax dollars into intelligence budgets, and they are responsible for making consequential decisions based on intelligence; when the crunch time comes, they do not want to hear that the benefits they expected when they made those investments are not really possible. Adding further to the tension between the two groups is the natural preference of intelligence professionals to keep as much information secret as possible versus the temptation of politicians to use that information to bolster public positions and arguments.
The election season only adds to the political pressures already at work. Politicians tend to believe members of their own party and suspect those of the other party. If the administration is not forthcoming with details, the opposition will accuse it of a cover-up. On the other hand, if it is forthcoming, the opposition may accuse it of playing politics with national security by divulging classified information. Factions within an administration (or bureaucrats who fear they will be blamed for something) may release selective bits of information either to reinforce or to undermine a particular version of events. Each side will jump to the conclusion that the clues that support its interpretation are obviously the correct and important ones and then proceed from that assumption.
At the same time, people engaged in political arguments over controversial events often simplify and distort those facts that are known (or are believed to be known), making each position simpler, but also more internally consistent and more different from the contending argument than the facts may warrant. Thus, in the Benghazi case, the two interpretations become caricatured as either a totally spontaneous event inspired by outrage over an Internet video or a terrorist attack planned long in advance by a known al-Qa’ida-affiliated organization that was untouched by emotion and uninfluenced by the video. The variants are treated as mutually exclusive, although there is no reason that parts of each cannot be true and it is unlikely that either is fully correct as is. Much of this political positioning occurs at the psychological level, before the conscious angling for advantage even begins.
In the case of the Benghazi attack, the selective release of three e-mail messages reinforced the confidence of those arguing that intelligence officials knew the complete truth all along. Yet those messages are vulnerable to the same errors as any early reports of such events. Just what did they say (keeping in mind that they constitute only a small piece of the communications that were flying around that day)?
The first, sent at 4:05 p.m.,Washington time (10:05 p.m., Benghazi time), reported on the subject line: “U.S. Diplomatic Mission in Benghazi Under Attack (SBU)” “SBU” indicates that the message is sensitive but unclassified. The text reads: “The Regional Security Officer reports the diplomatic mission is under attack. Embassy Tripoli reports that approximately 20 armed people fired shots; explosions have been heard as well. Ambassador Stevens, who is currently in Benghazi, and four COM personnel are in the compound safe haven. The 17th of February militia is providing security support.” Now, this tells us quite a bit about the situation, except that the “20 armed people” appears to refer to the diversion at the front gate, not the much larger group later reported to have come in through the rear. “COM” usually means “chief of mission” and presumably refers here to members of the ambassador’s staff, but the State Department later listed only two people with Ambassador Stevens in the safe haven: Sean Smith, who died along with the ambassador, and one Diplomatic Security agent, who survived. Moreover, the two 17th of February militiamen who showed up for duty that day were, by their own admission, hiding on the roof through most of the event.
A second message, sent at 4:54 p.m. (10:54, Benghazi time), reported that “the firing at the U.S. Diplomatic Mission in Benghazi has stopped and the compound has been cleared.” So the incident appears to be over. Yet minutes later U.S. personnel were abandoning the compound under fire. In fact, at this time, only two Americans had been killed. The other two deaths occurred hours later.
The third message, sent at 6:07 p.m. (12:07, Benghazi time), has attracted the most attention. It came under the subject heading “Ansar al-Sharia Claims Responsibility for Benghazi Attack.” The text reads, “Embassy Tripoli reports the group claimed responsibility on Facebook and Twitter and has called for an attack on Embassy Tripoli.” Now some people have read this in accordance with the rule: “If a terrorist says it on Facebook, it must be true.” Unfortunately, many terrorist attacks are accompanied by multiple contradictory claims of responsibility. In this case, the CIA ultimately concluded that it probably was true, but this was based on further evidence, including communications intercepts. It is not accurate to say that this was “known” at 6:07 p.m. on September 11; it had to be confirmed. By the way, no one has been able to verify that Ansar al-Sharia’s Facebook page actually said that at the time; it said something quite different the next day.
In connection with this, the classified version of Ambassador Susan Rice’s talking points reportedly laid responsibility to both Ansar al-Sharia and Al-Qa’ida in the Islamic Maghreb (AQIM). The Office of the Director of National Intelligence deleted the specific names and replaced them with the term “extremists” in the unclassified version. This was reportedly done because the link was considered “tenuous” even at that time. A further reason could have been to conceal from the two groups how much we knew (or did not know, as the case may be). Other leaked reports, however, suggest that the classified talking points were wrong in part; AQIM was not directly involved and learned of the event only after the fact, when Ansar al-Sharia called them up to brag about it. The relationship between the two groups is still not fully understood. In any case, the intelligence community still does not consider the talking points to have been distorted; unnamed sources within the community have expressed surprise at the political reaction.
Regarding other aspects of the current “common knowledge,” it is worth mentioning that both Reuters and the New York Times have stood by their interviews with people originally described as unarmed demonstrators. It now appears, however, that they arrived at the scene after the attack was initiated rather than before, as observers, cheerleaders, or perhaps looters. Also, the CIA continues to stand by its assessment that the attack was inspired by the Internet video and the attack on the U.S. embassy in Egypt earlier that day. The agency apparently has evidence that the attackers watched the Egypt assault on TV; moreover, that is what Ansar al-Sharia told AQIM when they called to brag. It is also what the attackers told the unarmed bystanders at the time. So, if the attack was “preplanned,” the planning occurred that same day. In this connection, the CIA has dropped the description “spontaneous attack” in favor of “opportunistic attack.”
The main point, again, is that confusion reigns during such events. Intelligence professionals understand this, and they take it into account. They avoid making claims that may be proved false as the evidence evolves. Beyond that, as a general rule, they prefer to say as little as possible in public. Intelligence agencies will do their best to sort the facts out, but assessments take time. The next time something like this occurs and you are sure that the intelligence community knows the truth the moment it happens, keep in mind the assessments of two historical events that were considerably larger and more significant than this one. First, on June 6, 1944, the Allies landed in Normandy, opening the Second Front in Europe. The Germans had expected the attack, but they expected it to come at Calais. It was at least seven weeks, about the end of July, before the German high command was convinced that Normandy was not a diversion for the real landing coming at Calais. By that time there were 1–2 million Allied troops in France.
Second, on October 12, 1950, the CIA—with the concurrence of the State Department, the army, the navy, and the air force—sent a memo to President Truman. It assessed that, despite Chinese threats and despite the fact that the possibility could not be ruled out entirely, the agency deemed it unlikely that China would intervene in the Korean War. At the time the memo was sent, China was already sending troops in moderate-sized units into Korea. It would be about three more weeks, early November, before General MacArthur’s Far Eastern Command realized that it had been fighting the Chinese army.
