September 11 attacks

The September 11 attacks became the decisive test of the U.S. intelligence system’s ability to turn scattered information into warning. The Congressional Joint Inquiry concluded that agencies had failed to grasp the individual and collective significance of information relevant to the plot, and had missed opportunities to disrupt it. Yet the book treats the attacks not as proof that decentralization itself was mistaken, but as evidence that decentralized agencies lacked effective ways to share and aggregate what they knew.

Background: A fragmented intelligence system

After the Second World War, the National Security Act of 1947 created the Central Intelligence Agency with an ambition of unity in command and intelligence. In practice, that centralization never took hold. The intelligence community grew into a collection of agencies with overlapping missions, including the CIA, the National Security Agency, the National Imagery and Mapping Agency, the National Reconnaissance Office, the Defense Intelligence Agency, and the intelligence arms of the military services. Although the CIA director was nominally responsible for the community, he exercised little supervision; much of the funding came from the Department of Defense. The FBI, which handled domestic law enforcement, stood almost entirely outside the community, despite the obvious relevance of information about foreign terrorists inside the United States.

This structure had a logic as well as a cost. Separate agencies could specialize, develop expertise, and interpret information from different perspectives. Information dismissed as noise by one group of analysts might reveal a clue or relationship to another. More broadly, knowledge is often local and difficult to transmit: people closest to a problem may possess experience that cannot readily be condensed into a report. The weakness was not simply that agencies were decentralized. It was that their autonomy was not matched by a reliable means of bringing information and judgments together.

The attacks and the missed warning

Before September 11, the system’s shortcomings had not produced an acknowledged reckoning, even though the intelligence community had failed to anticipate the 1993 World Trade Center bombing and the 1998 bombings of the U.S. embassy in Kenya and the USS Cole in Yemen. The attacks of September 11 made the failure seem undeniable. The Congressional Joint Inquiry found that intelligence agencies had not capitalized on information that appeared relevant, and that missed opportunities had allowed the plot to proceed. In the book’s formulation, the episode recalled Pearl Harbor: a catastrophic surprise that made the prior warning record newly difficult to ignore.

The attacks involved nineteen men, a fact that sharpens the question of what intelligence agencies could reasonably have foreseen. A retrospective inquiry can identify details that look significant once the outcome is known; that does not establish that analysts should have recognized their importance beforehand. Drawing on Roberta Wohlstetter’s account of the Pearl Harbor warning failure, the book stresses the difficulty of selecting meaningful signals from a mass of confusing information. Strategic surprise may be intrinsically hard to prevent, and the scale of an attack is not the only measure of how difficult it is to predict.

That caution does not end the argument. The intelligence community had failed to anticipate four major terrorist attacks between 1993 and 2001. Even if predicting the exact attacks on the World Trade Center and the Pentagon was impossible, a different organization might have made it more likely that analysts would recognize the combined significance of the information already available—and form a reasonable, concrete estimate of the danger. The central issue is therefore not whether every clue should have yielded a precise prediction, but whether the system could have made dispersed clues collectively intelligible.

The reform dispute: central control or aggregation

Congress’s response emphasized information sharing. Agencies had produced localized snapshots instead of a coherent picture of threats, and the right information had not reached the people who needed it. Senator Richard Shelby, in a supplement to the Senate Intelligence Committee report, criticized the FBI’s decentralized organization for leaving information in largely independent fiefdoms. His proposed remedy revived the idea behind the postwar intelligence reforms: create an agency above the disputatious bureaucracies, able to direct intelligence as a whole. In this account, decentralization had led the country astray, and centralization promised to correct it.

The book challenges the assumption that placing a small group at the center would by itself make the intelligence community wiser. Central authority might make it easier to supervise information, but it could also discard the advantages of decentralized work: specialized expertise, local knowledge, and independent judgment. A handful of decision-makers, however intelligent, would not necessarily outperform the broader community. The missing element was a mechanism to aggregate both information and the judgments of the people who held it. The proposed distinction is consequential: sharing is necessary, but appointing a central body to decide what matters is not the only way to achieve it.

Even relatively mechanical steps could have improved the flow of information while leaving agencies autonomous. Linking their computer databases, for example, would have made relevant records easier to find. Yet two years after the attacks, the government still lacked a unified watch list drawing on data from all parts of the intelligence community. The episode thus exposes a gap between the scale of the failure and the simplicity of at least some possible remedies: better connection did not require every specialist agency to be dissolved into one hierarchy.

Significance: making local knowledge collective

The larger problem raised by the attacks is how to preserve decentralized knowledge while making it useful across a whole system. Decentralization encourages people to specialize and respond to problems close at hand, but it offers no guarantee that discoveries in one part will reach another. Information can remain stranded, losing value precisely because no process carries it beyond its local setting. A capable intelligence system would need both sides of the arrangement: the independence that generates varied knowledge and a means of combining that knowledge into a collective picture.

This is why the book distinguishes centralization from aggregation. Shelby’s proposal addressed the absence of organization by putting an authority in charge; the alternative is to connect agencies and combine their judgments without assuming that one office must perform all interpretation. The September 11 attacks give that distinction its urgency. They reveal the danger of isolated institutional snapshots, but they do not establish that a single centralized intelligence authority would have read the evidence correctly. The enduring lesson is narrower and more demanding: decentralization can help a system know more, but only if it also knows how to make what its separate parts learn collectively usable.