For the first time in its 50 years, the U.N. decided last spring to determine causes of success and failure of its increasingly complex peacekeeping experiments. It established in the Department of Peace-Keeping Operations a "Lessons-Learned Mechanism," according to Under-Secretary General Kofi Annan, "to provide a greater capacity for in-depth study and analysis of experience, the conclusions of which can be applied to on-going as well as future operations." The Mechanism convened two high-level meetings in June and in mid-September, to decide first what went wrong in Somalia.

The meetings cost hundreds of thousands of dollars and were funded by several foundations and governments. Although the Mechanism gathered the critical family of officials who planned and executed operations, as well as area cognoscenti, their testimony was not recorded because only broad conclusions were to be reported.

This is why the Mechanism is not succeeding: Dysfunction in the field was replicated after the fact. Hard lessons cannot be learned politely, yet the same diplomatic approach that failed to respond to the social and political environment in Somalia failed to yield lessons at the meetings.

Unlike previous missions, as in Namibia, where luck was presented as flawless success, and Cambodia, where laurels were snatched for defeat, without face-saving elections it was more difficult to declare victory in Somalia. Apologists argue there are no more walking skeletons, but concede they may return. The "withdrawal" in March was declared a "masterpiece" in military terms by senior U.N. officials, yet the charge remains that the powers great and small "abandoned" Somalia.

The Mechanism's September meeting recognized the essentially political source of the symptoms of famine and violence in Somalia. There were several passionate pleas for the development of an operational concept for responding to anarchical conditions, which diplomatic peacekeeping and military force could not address. U.S. Adm. Jonathan T. Howe, the special representative of the Secretary-General during the U.N.'s combat phase in 1993, called for clear political authority in the field and unity of civilian command.

It was further suggested that a U.N. ombudsman directly responsible to the Security Council could provide to local and international communities greater accountability for executive decision-making on the ground. Also, a log should be kept of geographic specialists and other kinds of experts that can be consulted on call.

These limited proposals were marginalized. If knowledge is power, in the U.N. knowledge is dangerous and officials secretive, which is organizationally suicidal. By far the majority of "lessons" were banal conclusions that have been known for decades, such as the need for clearer mandates and more resources. Despite calls for frank discussions, a diplomatic environment prevented cool analysis; what is acceptable to say may not be useful to know. Self-criticism in sessions was transformed formally into self-justification as a whole.

This is not the way to build an institutional memory. Generals, statesmen and even national bureaucracies learn from concrete accounts, not disembodied truisms. Diplomatic conventions are useful for adopting negotiated documents, not capturing the context of factual events.

The Lessons-Learned Mechanism needs to write official histories that are honest - warts and all - if its mandate is to be fulfilled. The ideal would be to attach an observation team to U.N. missions from beginning to end, with comprehensive exposure to dimensions of each operation, to record flaws and strengths faithfully. This raises issues of the relative balance of academic freedom and access to decision-making imperatives, as well as the degree of openness of the final account. However, these proved surmountable technical questions during the second world war when the U.S., Canadian, British and Soviet armies were learning how to write their histories, including about sensitive topics like intelligence.

Without rigorous commitment to accuracy, the Mechanism will be no more effective than other U.N. methods of internal information gathering. Security Council missions to operational areas are highly politicized. The Joint Inspection Unit writes only non-controversial reports. And the Office of the Internal Oversight Services established at the end of 1994 failed its test of independence when it whitewashed allegations of U.N. mismanagement in Western Sahara.

If the U.N. cannot face the truth, it cannot do the job. Major General Imtiaz Shaheen, the first force commander in Somalia, has called on the organization to "discard its secretiveness." Otherwise, there will be little learned from the deaths of 136 peacekeepers and untold thousands of Somalis.