Why intelligence gathering stumbles

Once more, in the debate over the security of Americans in Beirut, attention is being focused on this nation's capacity to collect and use intelligence. Why did we not know in advance of the plans to bomb the embassy or the marines? Or did we? Did someone in the government have the information and not pass it along? Was it passed along and not used? Did high officials ignore it?

Americans expect our intelligence services to know in advance of threats to our interests and people. They are, in fact, putting their trust in a chancy and imperfect process. Individual mistakes in the gathering and evaluation of intelligence have undoubtedly been made, yet the real problems may lie elsewhere: in the sheer volume of information, the preference for tried and true sources, and the politics of evaluation.

At any given time, a huge volume of information is flowing into Washington from around the world through several channels: the CIA, the State Department, Defense agencies, the National Security Agency, and, where involved, the FBI. While every effort is made to alert those in crisis areas to relevant intelligence, information relating to Beirut, for example, may come from far away. Only a portion of the total information may be available to officials on the spot in time for them to act.

Information coming in to Washington will be organized and considered only in part. Much of it will consist of rumors, reports from unevaluated sources, intercepted communications waiting to be processed, and data from complicated electronic and photographic equipment requiring processing, interpretation, and collateral confirmation. In times of stress, the flow will be supplemented by snippets from those who want to appear informed, want to make mischief, are looking for money, or merely think they know something. Nevertheless, each item must be looked at and evaluated.

Items that are cryptic or unclear when first received may have more meaning in hindsight, often leading to claims that US officials were informed in advance. The question arises: Given the lives and interests at stake, why not act on every report of possible attack? To do so would keep personnel and families in a constant state of tension, seriously hampering the working of our embassies and, over time, reducing the credibility of the intelligence process.

The intelligence analyst in Washington tends to put the greatest credence on known, traditional sources. Today, in areas of upheaval such as Beirut, sources of the past may be of little value. Even five years ago, US officials were still in touch with Palestinian and Lebanese officials they had long known who were willing to talk to Americans and were close to most of the sources of violence. Information was exchanged and the security of Americans often assured. Most of these traditional sources have lost touch or departed. The task of penetrating the shadowy sources of today's terrorism carries more risk and is less certain of results, not only for US intelligence officers, but for their contacts as well. Evaluating the result is even more difficult.

Even in cases of reliable information coming, for example, through intercepted communications, the meaning is not always clear. Conversations can be cryptic, using ''double talk'' and subject to more than one interpretation. Information from numerous sources may give conflicting signals: The intelligence community must determine the true meaning. Each element may then tend to support its own sources and its own view of events. Valuable time is lost in getting information to the policymaker through bureaucratic and sometimes political battles over assessments.

When information is finally winnowed and conclusions reached, senior policymakers may not immediately accept its validity and act. They may prefer other sources fitting their own view of an issue. Further, the actions dictated by the information may be diplomatically, militarily, financially, and politically out of the question.

In the amount of information received by the US, it is possible to make a case that almost any event was foreseen by someone. That, however, is not the important question. What is meaningful is whether accurate information was sifted, assessed, received in time, and believed by those with the capacity to act and whether there existed, for them, feasible courses of action.

The judgment of individuals may occasionally be rightfully blamed for ''intelligence failures.'' More often, however, the answers lie in an inevitably complicated system in which chance may play as great a role as facts, talent, and judgment.