On Jan. 14, the Supreme Court heard argument in a case involving a writer for The New Yorker magazine. On March 27, the court will hear a case involving reporters in the Twin Cities. I am minded to write about the two cases, not because they are of towering significance, but for a different reason.

One shortcoming of those of us in the writing business is painfully clear: We seldom write anything about the writing business. So, today: shop talk.The New Yorker case arose in 1982 when a free-lance writer, Janet Malcolm, began a series of interviews with Dr. Jeffrey M. Masson. In the field of Freudian psychiatry he is a renowned figure, or at least a famous and controversial figure. Under Malcolm's informed questioning, he talked freely of his turbulent tenure and his subsequent dismissal as projects director of the Sigmund Freud archives.

The interviews led to a two-part article for The New Yorker in 1983. Masson was not pleased. He sued for libel, charging that the quotations attributed to him made him appear "unscholarly, irresponsible, vain, lacking in honesty and moral integrity." He said some quotations were "fabricated."

As it transpired, "fabricated" was too strong a word. The writer was able to establish that if Masson had not used EXACTLY the words attributed to him in direct quotations, he had said things that were substantively very close. Both a trial court and the 9th U.S. Circuit Court ruled in favor of the magazine. The writer had "rationally interpreted" her subject's views.

The question before the high court is important to the public figures we interview, and it goes directly to the credibility of our profession. When we put a statement in direct quotation marks, must it be EXACTLY what was said? My own answer is yes. On any issue of critical substance, we ought not to alter a single word. Some of my colleagues disagree.

On to the second case, brought by Dan Cohen against the Minneapolis Star Tribune and the St. Paul Pioneer Press. The story goes back to October of 1982. Cohen, a well-known political figure in the Twin Cities, was a public relations adviser to Wheelock Whitney, Republican candidate for governor.

Four days before the election, Cohen got in touch with reporters for the two papers. In return for their promise to protect his confidence, he would leak damaging documents about the Democratic candidate for lieutenant governor. The reporters agreed not to identify Cohen as their source.

Then their editors overruled the two reporters. Over the reporters' anguished protest, the stories ran the next day with full attribution to Cohen. The Whitney campaign immediately fired him. Cohen sued the newspapers. A jury awarded him $200,000 in compensatory damages and $500,000 in punitive damages, but on appeal both awards were wiped out.

To promise anonymity to a confidential source, and then to violate that promise, is dishonorable. Reneging causes sources to dry up. In our business such promises routinely are given and kept. It is all but unthinkable for editors not to stand behind their reporters in such transactions.

But Cohen's leak was immediately before the election. There would not have been time for the Democratic candidate to give her side of the damaging revelations. Solely because of this time element, I would have killed the story altogether. If the competition had used the story without attribution, I would then have identified Cohen as the source. I expect many of my colleagues would have acted differently.

So this is shop talk. We ought to talk with you more often about such issues.