I've tilted at windmills, and the windmills won.

Jonathan Alter, an editor of The Washington Monthly from 1981 to 1983, is a senior writer at Newsweek.

Six years ago, I was offered an unusual job at Newsweek that was at least partly the result of the efforts of the editor of this magazine. While I'm grateful to Charlie Peters for assisting my career advancement, the job he helped arrange also points up some practical failures associated with his view of how to get the truth out of the bureaucracy.

One of Charlie's all-time heroes in this regard is a woman named Lorena Hickock, a journalist and close friend of Eleanor Roosevelt. Hickock traveled across the country during the 1930s and delivered candid reports to the Roosevelts and Harry Hopkins about the way New Deal programs were really working. She deserves Charlie's canonization, for these are models of the kind of journalism that The Washington Monthly seeks to publish. Instead of merely reporting the pseudo-important spin of highlevel officials, she went out and assessed how their programs were affecting reat people. By circumventing the hidebound chain of command, FDR was able to learn what was actually happening in his own administration.

Hickock's success reinforced the lessons of Charlie's own experience at the Peace Corps, where he supervised a crew of independent evaluators who burrowed into programs in an effort to report the truth to the top. In 1978, Charlie convinced James Fallows, then Jimmy Carter's chief speechwriter, to seek a similar arrangement with the Carter White House. Because of opposition from high-level aides, the proposal for a corps of independent evaluators was scotched. Charlie later wrote that perhaps Carter's administration might have been saved had someone like Fallows reported in advance the truth about the preparations for the Iran hostage rescue mission of 1980, which ended in humiliation.

In 1982, Charlie got a chance to try his idea on a much smaller scale. Newsweek hired a new editor, William Broyles, who was a fan of the Monthly. Partly at Charlie's urging, he soon hired me for a variation on the Hickock role: to report directly to him on what was wrong with Newsweek, and to suggest ways to fix it. Ironically, this was a journalistic variation on a role the Monthly had traditionally detested-that of the management consultant. But bearing the Hickok and Fallows models in mind, I plunged ahead.

From the start, the idea failed. As anyone who has ever worked in a large organization knows, a person in this role is immediately labeled a spy. Every time I walked into a work area, the collected Newsweek writers (many of whom I now count as close friends) would immediately stop talking. A memo I wrote to Broyles outlining suggestions was intercepted and copied for the enjoyment of snickering colleagues. More important, my suggestions for improvements betrayed a lack of understanding of the difference between practical and impractical change. For instance, like some others at Newsweek, I thought then and still do that the separation between reporting and writing at newsmagazines usually made little sense. That distinction, in the years since, has been thankfully eroded. But I should have known that it could not be done overnight. Institutions don't turn on dimes. I was operating in a Washington Monthly, pie-in-the-sky manner, assuming that once the right answer was found, the argument over change would be settled, when, in fact, it was only beginning. Given my newness, my feel for the corporate culture was naturally weak, which made me a relatively poor source of intelligence about the magazine. After a few months, I left that job and became a regular Newsweek staff writer, which I remain today.

My experience with Broyles did not discredit Charlie's vision of how to wring the truth from a bureaucracy, but it did modify it. Yes, the boss must constantly circumvent the chain of command in order to find out what is going on in his organization. One of the most dispiriting days of the 1988 campaign for me was whenI interviewed Michael Duk-akis on this subject and he disagreed. (Dukakis believes that one should hire good people and question them closely, but obey the chain.) Organization charts really are the enemy of truth and the ally of those who lack confidence in their own leadership abilities. They also often bear little resemblance to the way power actually works in an institution. But Charlie's remedy for the problem was only half right. While gimlet-eyed outsiders can provide a fresh analysis, they do not substitute for informed intelligence from within. Outsiders undertaking evaluation alone will usually be neutralized.

FDR, who had extensive experience in the bureaucracy as a former assistant secretary of the Navy, understood this. He used Lorena Hickock, But he also reached down around his Cabinet secretaries and personally quizzed underlings (like Sumner Welles). Outsiders and insiders-the pincer effect. Ironically, even at the Peace Corps, Charlie operated from a base of insider knowledge about the organization that put the findings of his heralded group of outside evaluators in the right perspective.

One of the conclusions I've drawn from covering the media is that the outside-inside pincer approach should apply to journalism as well as to life inside a bureaucracy. The best coverage of the White House and presidential campaigns employs this technique. One reporter covers the story from inside, cultivating sources; the other, taking the outsider's perspective, worries much less about offending big-shot officials. One of the reasons The Washington Post uncovered Watergate, and the rest of the White House press corps missed it, was that the Post employed a variation of this approach. Almost by accident, the paper had an outside team in police beat reporters Bob Woodward and Carl Bernstein, who cracked the story in part because they had no reason to fear angering the Nixon administration officials. But they got more help than Charlie recognizes from regular Post reporters covering the government. Neither approach would have worked alone.

Bureaucracies are all alike, but they are also all different, if you know what I mean. And truth is not always decisive. Even had he won the assignment as all-purpose evaluator, Fallows could not have saved the Iran hostage rescue mission from disaster. The military would never have let him learn enough about the operation to predict convincingly that it would fail, and even if he did somehow learn it, the president would have likely sided with the experts. Only a skeptic from within the bureaucracy could have much chance of making a difference.

The Monthly prides itself in reporing on how things really work. But too often it sees the world as it wants it to be, rather than as it really is. That is a fair price to pay for an otherwise commendable form of idealism, but it isn't always relevant to daily life. I learned that lesson the hard way.