Time magazine's chief of correspondents testified today that he has never learned the names of the anonymous sources the magazine relied on for the paragraph in a 1983 article that prompted former Israeli defense minister Ariel Sharon's $50 million libel suit against the magazine.

Richard L. Duncan, supervisor of Time's 90 correspondents in the United States and abroad, said that before the report was published he depended on assurances from Time's Jerusalem bureau chief that the information about Sharon's role in the massacre of Palestinian refugees in Beirut in 1982 was reliable. Once Sharon filed the lawsuit, Duncan said, he thought it would be unwise for him to learn the names of the sources, although he did know general information about the positions they held.

". . . Some reporters go to jail for not revealing sources," Duncan testified, a tight smile flashing briefly across his face. "I guess the shortest answer would be I thought the fewest of us who might be liable to go to jail should be liable."

In fact, New York state has a law shielding reporters' confidential sources and Time Israeli correspondent David Halevy has invoked it frequently during his testimony in the trial.

Under intensive questioning by Sharon's lead lawyer and by U.S. District Court Judge Abraham D. Sofaer, Halevy acknowledged that he had no source for a key detail in the article but had relayed the information to Time on the basis of his "analysis."

Speaking in a soft voice, Duncan testified that he had felt no need to question Halevy about his sources after the Time article was denounced as "bald lies and slander" in the Knesset, Israeli's parliament, by then-Prime Minister Menachem Begin.

Duncan said this was because he thought Begin was reacting to erroneous interpretations of the article in the Israeli press and because Begin's language "seemed to be consistent with the political forum and the kind of vocabulary which upon occasion was used in the Knesset, so that it had somewhat less import than it might have had in another country at another time."

Sofaer again told the jury today, "This is not a 'nice-guy' case. People are trying to prove nasty things about each other here."

The judge's words came as he gave Sharon's lawyers broad latitude to try to prove that Duncan had been misleading and disingenuous in the sworn deposition he made before the trial.

Duncan testified today, "I may be dumb but I'm not disingenuous."

That response came as Sharon's lawyer hammered away at Duncan for failing to disclose in his deposition that he had put Halevy on probation in 1980 after an internal Time investigation failed to substantiate a story in which Halevy had reported that Begin had been ill.

In notifying Halevy about his probation, Duncan wrote that he expected "a more obvious effort on your part to ensure that what you report to Time . . . is printable, reliable information, reflecting not just informed speculation, but the most likely true situation."

Duncan said he had forgotten the letter when he was asked during the deposition about Halevy's performance. Asked today how many correspondents he had put on probation in his six years as their chief, Duncan replied that Halevy had been the only one.

During the time Duncan's deposition was being taken, Time's lawyers were battling with Sharon's attorneys in court to keep Halevy's personnel file out of the case. When Time later lost that battle, Sharon's lawyers gained possession of the letter and a memorandum Duncan wrote his superiors about the story on Begin's health.

In the memorandum, Duncan said, ". . . I believe our story is wrong. Inescapably that conclusion then requires a further decision as to whether Halevy was either (a) inexcusably shoddy in his reporting or (b) intentionally misled us, or (c) was the victim of an incredibly well orchestrated disinformation plot."

Duncan testified today that he leans toward believing the last explanation. He also acknowledged that Time had published a "partial retraction" of that story.

When asked as part of his deposition if a correction had been made, Duncan had responded, "We do not publish retractions per se. That is not our style. We usually invite someone who feels injured to write a letter, in which case we will publish that letter ordinarily and may or may not respond to the letter."